From c864faba0be36d702b4cea5f01dbee827e9d607a Mon Sep 17 00:00:00 2001
From: luckylionheart <uyscuti0513@gmail.com>
Date: Thu, 16 Oct 2025 13:53:20 -0400
Subject: [PATCH 1/2] WIP: fix #448

---
 index.js      | 4320 ++++++++++++++++++++++++++-----------------------
 test_nlp.json |  542 ++++---
 2 files changed, 2628 insertions(+), 2234 deletions(-)

diff --git a/index.js b/index.js
index 1f555ed..f85d004 100755
--- a/index.js
+++ b/index.js
@@ -13,2042 +13,2322 @@
  * Own matchAll used for the chemical formulae
  */
 function getMatchAll(inputstr, regexp) {
-	const matches = [];
-	inputstr.replace(regexp, function (...args) {
-		const arr = [].slice.call(args, 0),
-			extras = arr.splice(-2);
-		arr.index = extras[0];
-		arr.input = extras[1];
-		matches.push(arr);
-	});
-	return matches.length ? matches : null;
+  const matches = [];
+  inputstr.replace(regexp, function (...args) {
+    const arr = [].slice.call(args, 0),
+      extras = arr.splice(-2);
+    arr.index = extras[0];
+    arr.input = extras[1];
+    matches.push(arr);
+  });
+  return matches.length ? matches : null;
 }
 
 function capitalize(string) {
-	return string.charAt(0).toUpperCase() + string.slice(1);
+  return string.charAt(0).toUpperCase() + string.slice(1);
 }
 
 const OptimadeNLP = function () {
-	/*
-	 * Definitions
-	 */
-	const stop_words = [
-		'a',
-		'about',
-		'above',
-		'after',
-		'again',
-		'against',
-		'all',
-		'am',
-		'an',
-		'and',
-		'any',
-		'are',
-		"aren't",
-		'as',
-		'at',
-		'be',
-		'because',
-		'been',
-		'before',
-		'being',
-		'below',
-		'between',
-		'both',
-		'but',
-		'by',
-		"can't",
-		'cannot',
-		'could',
-		"couldn't",
-		'did',
-		"didn't",
-		'do',
-		'does',
-		"doesn't",
-		'doing',
-		"don't",
-		'down',
-		'during',
-		'each',
-		'few',
-		'for',
-		'from',
-		'further',
-		'had',
-		"hadn't",
-		'has',
-		"hasn't",
-		'have',
-		"haven't",
-		'having',
-		'he',
-		"he'd",
-		"he'll",
-		"he's",
-		'her',
-		'here',
-		"here's",
-		'hers',
-		'herself',
-		'him',
-		'himself',
-		'his',
-		'how',
-		"how's",
-		'i',
-		"i'd",
-		"i'll",
-		"i'm",
-		"i've",
-		'if',
-		'in',
-		'into',
-		'is',
-		"isn't",
-		'it',
-		"it's",
-		'its',
-		'itself',
-		"let's",
-		'me',
-		'more',
-		'most',
-		"mustn't",
-		'my',
-		'myself',
-		'no',
-		'nor',
-		'not',
-		'of',
-		'off',
-		'on',
-		'once',
-		'only',
-		'or',
-		'other',
-		'ought',
-		'our',
-		'ours',
-		'ourselves',
-		'out',
-		'over',
-		'own',
-		'same',
-		"shan't",
-		'she',
-		"she'd",
-		"she'll",
-		"she's",
-		'should',
-		"shouldn't",
-		'so',
-		'some',
-		'such',
-		'than',
-		'that',
-		"that's",
-		'the',
-		'their',
-		'theirs',
-		'them',
-		'themselves',
-		'then',
-		'there',
-		"there's",
-		'these',
-		'they',
-		"they'd",
-		"they'll",
-		"they're",
-		"they've",
-		'this',
-		'those',
-		'through',
-		'to',
-		'too',
-		'u',
-		'under',
-		'until',
-		'up',
-		'very',
-		'was',
-		"wasn't",
-		'we',
-		"we'd",
-		"we'll",
-		"we're",
-		"we've",
-		'were',
-		"weren't",
-		'what',
-		"what's",
-		'when',
-		"when's",
-		'where',
-		"where's",
-		'which',
-		'while',
-		'who',
-		"who's",
-		'whom',
-		'why',
-		"why's",
-		'with',
-		"won't",
-		'would',
-		"wouldn't",
-		'you',
-		"you'd",
-		"you'll",
-		"you're",
-		"you've",
-		'your',
-		'yours',
-		'yourself',
-		'yourselves',
-	]; /* exact */
-
-	const arity_keys = [
-		null,
-		'unary',
-		'binary',
-		'ternary',
-		'quaternary',
-		'quinary',
-		'multinary',
-		'multinary',
-		'multinary',
-		'multinary',
-		'multinary',
-	]; // NB null is for "0-ary"
-
-	const periodic_elements = [
-		'h',
-		'he',
-		'li',
-		'be',
-		'b',
-		'c',
-		'n',
-		'o',
-		'f',
-		'ne',
-		'na',
-		'mg',
-		'al',
-		'si',
-		'p',
-		's',
-		'cl',
-		'ar',
-		'k',
-		'ca',
-		'sc',
-		'ti',
-		'v',
-		'cr',
-		'mn',
-		'fe',
-		'co',
-		'ni',
-		'cu',
-		'zn',
-		'ga',
-		'ge',
-		'as',
-		'se',
-		'br',
-		'kr',
-		'rb',
-		'sr',
-		'y',
-		'zr',
-		'nb',
-		'mo',
-		'tc',
-		'ru',
-		'rh',
-		'pd',
-		'ag',
-		'cd',
-		'in',
-		'sn',
-		'sb',
-		'te',
-		'i',
-		'xe',
-		'cs',
-		'ba',
-		'la',
-		'ce',
-		'pr',
-		'nd',
-		'pm',
-		'sm',
-		'eu',
-		'gd',
-		'tb',
-		'dy',
-		'ho',
-		'er',
-		'tm',
-		'yb',
-		'lu',
-		'hf',
-		'ta',
-		'w',
-		're',
-		'os',
-		'ir',
-		'pt',
-		'au',
-		'hg',
-		'tl',
-		'pb',
-		'bi',
-		'po',
-		'at',
-		'rn',
-		'fr',
-		'ra',
-		'ac',
-		'th',
-		'pa',
-		'u',
-		'np',
-		'pu',
-		'am',
-		'cm',
-		'bk',
-		'cf',
-		'es',
-		'fm',
-		'md',
-		'no',
-		'lr',
-		'rf',
-		'db',
-		'sg',
-		'bh',
-		'hs',
-		'mt',
-		'ds',
-		'rg',
-		'cn',
-		'nh',
-		'fl',
-		'mc',
-		'lv',
-		'ts',
-		'og'
-	]; /* exact */
-
-	const periodic_elements_cased = periodic_elements.map(function (x) {
-		return capitalize(x);
-	});
-
-	const periodic_element_names = [
-		'hydrogen',
-		'helium',
-		'lithium',
-		'beryllium',
-		'boron',
-		'carbon',
-		'nitrogen',
-		'oxygen',
-		'fluorine',
-		'neon',
-		'sodium',
-		'magnesium',
-		'aluminium',
-		'silicon',
-		'phosphorus',
-		'sulfur',
-		'chlorine',
-		'argon',
-		'potassium',
-		'calcium',
-		'scandium',
-		'titanium',
-		'vanadium',
-		'chromium',
-		'manganese',
-		'iron',
-		'cobalt',
-		'nickel',
-		'copper',
-		'zinc',
-		'gallium',
-		'germanium',
-		'arsenic',
-		'selenium',
-		'bromine',
-		'krypton',
-		'rubidium',
-		'strontium',
-		'yttrium',
-		'zirconium',
-		'niobium',
-		'molybdenum',
-		'technetium',
-		'ruthenium',
-		'rhodium',
-		'palladium',
-		'silver',
-		'cadmium',
-		'indium',
-		'tin',
-		'antimony',
-		'tellurium',
-		'iodine',
-		'xenon',
-		'caesium',
-		'barium',
-		'lanthanum',
-		'cerium',
-		'praseodymium',
-		'neodymium',
-		'promethium',
-		'samarium',
-		'europium',
-		'gadolinium',
-		'terbium',
-		'dysprosium',
-		'holmium',
-		'erbium',
-		'thulium',
-		'ytterbium',
-		'lutetium',
-		'hafnium',
-		'tantalum',
-		'tungsten',
-		'rhenium',
-		'osmium',
-		'iridium',
-		'platinum',
-		'gold',
-		'mercury',
-		'thallium',
-		'lead',
-		'bismuth',
-		'polonium',
-		'astatine',
-		'radon',
-		'francium',
-		'radium',
-		'actinium',
-		'thorium',
-		'protactinium',
-		'uranium',
-		'neptunium',
-		'plutonium',
-		'americium',
-		'curium',
-		'berkelium',
-		'californium',
-		'einsteinium',
-		'fermium',
-		'mendelevium',
-		'nobelium',
-		'lawrencium',
-		'rutherfordium',
-		'dubnium',
-		'seaborgium',
-		'bohrium',
-		'hassium',
-		'meitnerium',
-		'darmstadium',
-		'roentgenium',
-		'copernicium',
-		'nihonium',
-		'flerovium',
-		'moscovium',
-		'livermorium',
-		'tennessine',
-		'oganesson'
-	]; /* fuzzy */
-
-	const lat_p2i = {
-		cubic: 1,
-		hexagonal: 2,
-		trigonal: 3,
-		tetragonal: 4,
-		orthorhombic: 5,
-		monoclinic: 6,
-		triclinic: 7,
-		rhombohedral: 3,
-		cub: 1,
-		hex: 2,
-		hexag: 2,
-		trig: 3,
-		tet: 4,
-		tetr: 4,
-		tetrag: 4,
-		orth: 5,
-		ortho: 5,
-		monocl: 6,
-		tric: 7,
-		tricl: 7,
-		rhom: 3,
-		rhomb: 3,
-	};
-	const lat_fgrs = Object.keys(lat_p2i);
-	const lat_i2p = {
-		1: 'cubic',
-		2: 'hexagonal',
-		3: 'trigonal',
-		4: 'tetragonal',
-		5: 'orthorhombic',
-		6: 'monoclinic',
-		7: 'triclinic',
-	};
-
-	const mpds_classes = [
-		'ab initio calculations',
-		'ab initio literature',
-		'actinoid',
-		'adamantane',
-		'aegirine',
-		'alkali',
-		'alkaline',
-		'allargentum',
-		'almandine',
-		'alum',
-		'alunogen',
-		'amide',
-		'analcime',
-		'anatase',
-		'anorpiment',
-		'anorthoclase',
-		'antiferroelectric',
-		'antiferromagnet',
-		'antiferromagnetic',
-		'arsenate',
-		'arsenide',
-		'ashcroftine',
-		'auricupride',
-		'aurocupride',
-		'azide',
-		'baileychlore',
-		'bariopyrochlore',
-		'baryte',
-		'beryl',
-		'beta-alumina',
-		'beta-boron',
-		'biguanide',
-		'binary',
-		'birefringent',
-		'borane',
-		'borate',
-		'borax',
-		'boride',
-		'borocarbide',
-		'borohydride',
-		'boronitride',
-		'botryogen',
-		'bromanilate',
-		'bromide',
-		'bromoimide',
-		'calomel',
-		'carbamate',
-		'carbide',
-		'carbonate',
-		'carbonyl',
-		'carboxylate',
-		'celestine',
-		'cell and atoms',
-		'cell-only',
-		'celsian',
-		'cesiokenopyrochlore',
-		'chalcogen',
-		'charge-density wave state',
-		'chevrel',
-		'chimney-ladder',
-		'chloranilate',
-		'chlorate',
-		'chloride',
-		'chloritoid',
-		'chlorosulfate',
-		'chromate',
-		'chrysoberyl',
-		'chrysotile',
-		'cinnabar',
-		'clathrate',
-		'clinochlore',
-		'clinoclase',
-		'clodronate',
-		'close-packed',
-		'cluster glass',
-		'colossal magnetoresistance',
-		'conductor',
-		'corundum',
-		'cosmochlor',
-		'croconate violet',
-		'croconate',
-		'cryptomelane',
-		'cuprate',
-		'cuspidine',
-		'cyamelurate',
-		'cyanamide',
-		'cyanamidonitrate',
-		'cyananilate',
-		'cyanide',
-		'cyanotetrazolate',
-		'cyanoureate',
-		'cyanurate',
-		'cyprine',
-		'davyne',
-		'deuteride',
-		'deuterium',
-		'devilline',
-		'diamagnetic',
-		'diamond',
-		'diarsenate',
-		'diaspore',
-		'diazanide',
-		'diazenide',
-		'dichromate',
-		'digermanate',
-		'diiodobromide',
-		'dinitramide',
-		'diopside',
-		'dioptase',
-		'dioxobromate',
-		'dioxoiodate',
-		'dioxosulfate',
-		'dioxothiosulfate',
-		'diphosphate',
-		'diphosphonate',
-		'dipolyhedral',
-		'diselenate',
-		'disilicate',
-		'disordered',
-		'disulfate',
-		'dithiocarbamate',
-		'dithiocarbonate',
-		'dithionate',
-		'dithiooxalate',
-		'dithiophosphate',
-		'dithiosquarate',
-		'divanadate',
-		'epidote',
-		'euchlorine',
-		'euclase',
-		'eudialyte',
-		'eulytine',
-		'fermi liquid',
-		'feroxihyte',
-		'feroxyhyte',
-		'ferrielectric',
-		'ferrimagnet',
-		'ferroelastic',
-		'ferroelectric',
-		'ferromagnet',
-		'ferromagnetic',
-		'fluor-schorl',
-		'fluoride',
-		'fluoroborate',
-		'frank-kasper',
-		'friauf-laves',
-		'fulleride',
-		'fulminate',
-		'galena',
-		'gamma-brass',
-		'garnet',
-		'giant magnetocaloric effect',
-		'gismondine',
-		'glaucodot',
-		'glaucophane',
-		'grossular',
-		'guanidinate',
-		'gypsum',
-		'hafnon',
-		'half metal',
-		'halogen',
-		'hard magnet',
-		'harmotome',
-		'haueyne',
-		'heavy fermion',
-		'hedyphane',
-		'helimagnet',
-		'helvine',
-		'hexasulfate',
-		'hexathionate',
-		'host-guest',
-		'humboldtine',
-		'hydrate',
-		'hydride',
-		'hydroxide',
-		'hypercinnabar',
-		'hypophosphate',
-		'ice',
-		'imide',
-		'iminate',
-		'intercalation',
-		'intermediate valence',
-		'intermetallic',
-		'iodate',
-		'iodide',
-		'ionic conductor',
-		'iridium',
-		'isoferroplatinum',
-		'isopolyhedral',
-		'isothermal section',
-		'kornerupine',
-		'kosmochlor',
-		'lanthanoid',
-		'lavendulan',
-		'levyne',
-		'lime',
-		'liquidus projection',
-		'litharge',
-		'lithiophosphate',
-		'luminescent',
-		'machine learning',
-		'machine-learning',
-		'magnesiochloritoid',
-		'magnetoelastic',
-		'magnetoelectric',
-		'manganate',
-		'massicot',
-		'mellitate',
-		'melonate',
-		'metacinnabar',
-		'metal',
-		'metalloid',
-		'metamagnet',
-		'metavoltine',
-		'mica',
-		'microcline',
-		'microline',
-		'mictomagnet',
-		'minium',
-		'molybdate',
-		'multiferroic',
-		'multinary',
-		'nasicon',
-		'natron',
-		'natrophosphate',
-		'natroxalate',
-		'negative thermal expansion',
-		'nepheline',
-		'nickeline',
-		'niobocarbide',
-		'niter',
-		'nitranilate',
-		'nitrate',
-		'nitratine',
-		'nitride',
-		'nitroformate',
-		'noble gas',
-		'non disordered',
-		'non-disordered',
-		'non-linear optics',
-		'nonaflate',
-		'nonmetal',
-		'nordenskioeldine',
-		'nosean',
-		'olivine',
-		'optically isotropic',
-		'organic',
-		'orpiment',
-		'orthoborate',
-		'orthoclase',
-		'orthogermanate',
-		'orthonitrate',
-		'orthophosphate',
-		'orthosilicate',
-		'oxalate',
-		'oxamate',
-		'oxide',
-		'oxoiodate',
-		'oxonitrate',
-		'oxotetrazolate',
-		'oxy-schorl',
-		'ozonide',
-		'palladogermanide',
-		'paracelsian',
-		'paramagnet',
-		'pararealgar',
-		'pauli paramagnet',
-		'peer review',
-		'peer reviewed',
-		'peer-review',
-		'peer-reviewed',
-		'perchlorate',
-		'perhydrate',
-		'periclase',
-		'periodate',
-		'permanganate',
-		'pernitride',
-		'peroxide',
-		'pertechnetate',
-		'phosphate',
-		'phosphide',
-		'phosphinate',
-		'phosphonate',
-		'photocatalyst',
-		'photovoltaic effect',
-		'photovoltaic',
-		'piezoelectric',
-		'pnictogen',
-		'polaron conductor',
-		'polycrase',
-		'potassic',
-		'prism',
-		'prussian blue',
-		'pseudorutile',
-		'pyrochlore',
-		'pyroelectric',
-		'pyrope',
-		'quadridavyne',
-		'quartz',
-		'quasicrystal',
-		'quaternary',
-		'quinary',
-		'radioactive',
-		'rare earth',
-		'realgar',
-		'refractory',
-		'relaxor',
-		'retzian',
-		'rhodarsenide',
-		'rhodizonate',
-		'rhomboclase',
-		'rocksalt',
-		'rubicline',
-		'ruddlesden-popper',
-		'rutheniridosmine',
-		'rutherfordine',
-		'rutile',
-		'salammoniac',
-		'sanidine',
-		'sapphirine',
-		'sarcopside',
-		'schorl',
-		'selenide',
-		'selenidel',
-		'semiconductor',
-		'semimetal',
-		'shape memory effect',
-		'shape memory',
-		'siderotil',
-		'silanide',
-		'silicate',
-		'silicide',
-		'sillen-aurivillius',
-		'skyrmion',
-		'solidus projection',
-		'solvus projection',
-		'spessartine',
-		'spin glass',
-		'spinel',
-		'spodumene',
-		'squarate',
-		'steenstrupine',
-		'stibarsen',
-		'subsolidus relations',
-		'sulfamate',
-		'sulfamide',
-		'sulfate',
-		'sulfide',
-		'sulfidel',
-		'sulfinylamide',
-		'superconductor',
-		'superhard',
-		'superionic conductor',
-		'superoxide',
-		'talc',
-		'tantalcarbide',
-		'tellurantimony',
-		'telluride',
-		'ternary',
-		'tetrahedron',
-		'thermoelectric',
-		'thiocyanate',
-		'thiocyanurate',
-		'thiophosphate',
-		'thiosulfate',
-		'topaz',
-		'topological insulator',
-		'tourmaline',
-		'transitional',
-		'triflate',
-		'tripolyhedral',
-		'triteride',
-		'trithionate',
-		'tritide',
-		'tritium',
-		'trona',
-		'tungstate',
-		'turquoise',
-		'ulvoespinel',
-		'unary',
-		'uranophane',
-		'uranopolycrase',
-		'urate',
-		'urea',
-		'ureate',
-		'van vleck paramagnet',
-		'vanadate',
-		'vertical section',
-		'violurate',
-		'xenotime',
-		'zincobotryogen',
-		'zircon',
-		'zircosulfate',
-	];
-
-	const mpds_props = [
-		'acceptor concentration',
-		'acceptor to donor concentration',
-		'activation energy',
-		'adiabatic bulk modulus',
-		'angle-resolved photoelectron spectra',
-		'atomic structure',
-		'band gap',
-		'birefringence',
-		'bremsstrahlung isochromat spectra',
-		'charge carrier concentration',
-		'charge carrier mobility',
-		'charge transfer',
-		'charge-density wave',
-		'charge-transfer energy',
-		'coefficient of schottky term in heat capacity',
-		'coercive electric field',
-		'coercive field',
-		'coherence length',
-		'cohesive energy',
-		'compressibility',
-		'conductivity',
-		'core-electron contribution to magnetic susceptibility',
-		'critical current density',
-		'critical magnetic field',
-		'crystal electric field parameter',
-		'crystal electric field parameters',
-		'crystal electric field splitting',
-		'crystal electric field',
-		'crystal field level',
-		'crystalline structure',
-		'crystal cell',
-		'crystal structure',
-		'curie coefficient',
-		'curie temperature',
-		'curie-weiss paramagnetism',
-		'debye temperature',
-		'decomposition temperature',
-		'decomposition',
-		'diamagnetic contribution to magnetic susceptibility',
-		'dielectric constant',
-		'dielectric loss tangent',
-		'diffusion',
-		'donor concentration',
-		'donor energy',
-		'effective charge',
-		'effective electron number',
-		'effective mass of electrons to holes ratio',
-		'effective mass of electrons',
-		'effective mass',
-		'einstein temperature',
-		'elastic compliance',
-		'elastic moduli',
-		'elastic stiffness coefficient',
-		'elasticity',
-		'electric field gradient',
-		'electric polarization',
-		'electrical conductivity',
-		'electric properties',
-		'electrical properties',
-		'electrical resistivity',
-		'electrochemical impedance spectroscopy',
-		'electron density maps',
-		'electron density of states at fermi level',
-		'electron density of states',
-		'electron energy band structure',
-		'electron energy loss spectra',
-		'electron grueneisen coefficient',
-		'electron mobility',
-		'electron paramagnetic resonance spectra',
-		'electron spin resonance spectra',
-		'electron-phonon interaction parameter',
-		'electronic contribution to heat capacity',
-		'electronic contribution to thermal conductivity',
-		'electronic energy gap',
-		'electronic heat capacity coefficient',
-		'electronic properties',
-		'energy at fermi level',
-		'energy band structure',
-		'energy gap for direct transition',
-		'energy gap for indirect transition',
-		'energy level diagram',
-		'energy of optical phonon',
-		'energy product',
-		'energy',
-		'enthalpy change at melting point',
-		'enthalpy change at phase transition',
-		'enthalpy change at structural transition',
-		'enthalpy change',
-		'enthalpy of formation',
-		'enthalpy of reaction',
-		'enthalpy',
-		'entropy change at melting point',
-		'entropy change at phase transition',
-		'entropy of formation',
-		'entropy of reaction',
-		'entropy',
-		'eutectoid decomposition',
-		'exchange field',
-		'exchange interaction parameter',
-		'exciton energy',
-		'extended x-ray absorption fine structure',
-		'extraordinary refractive index',
-		'fermi energy',
-		'fermi surface',
-		'ferroelasticity',
-		'ferroelectric curie temperature',
-		'ferroelectric hysteresis',
-		'ferroelectric neel temperature',
-		'ferroelectric transitions',
-		'field dependence of resistivity',
-		'figure of merit',
-		'freezing temperature for spin glass',
-		'fusion',
-		'gibbs energy change',
-		'gibbs energy of formation',
-		'gibbs energy of reaction',
-		'ginzburg-landau parameter',
-		'grueneisen coefficient',
-		'gruneisen coefficient',
-		'hall coefficient',
-		'hall effect',
-		'hall mobility',
-		'hardness',
-		'heat capacity at constant pressure',
-		'heat capacity at constant volume',
-		'heat capacity coefficient',
-		'heat capacity discontinuity at structural transition',
-		'heat capacity discontinuity at superconducting transition',
-		'heat capacity discontinuity',
-		'heat capacity',
-		'high-frequency permittivity',
-		'hole mobility',
-		'hydrogen diffusion',
-		'hyperfine magnetic field',
-		'imaginary part of magnetic susceptibility',
-		'imaginary part of permittivity',
-		'inelastic neutron scattering',
-		'inelastic x-ray scattering',
-		'infrared spectra',
-		'ionic conductivity',
-		'irreversibility field',
-		'isomer shift',
-		'isothermal bulk modulus',
-		'isothermal linear compressibility',
-		'isothermal volume compressibility',
-		'knoop hardness',
-		'kondo behavior',
-		'kondo temperature',
-		'lattice',
-		'linear magnetostriction',
-		'linear thermal expansion coefficient',
-		'longitudinal sound velocity',
-		'longitudinal-mode elastic coefficient',
-		'lorentz number',
-		'lower critical magnetic field',
-		'lowest temperature of investigation',
-		'luminescence lifetime',
-		'luminescence',
-		'magnetic anisotropy field',
-		'magnetic anisotropy',
-		'magnetic circular x-ray dichroism',
-		'magnetic dichroism',
-		'magnetic direction',
-		'magnetic entropy',
-		'magnetic field for magnetic transition',
-		'magnetic field for structural transition',
-		'magnetic heat capacity',
-		'magnetic hysteresis',
-		'magnetic moment',
-		'magnetic order',
-		'magnetic penetration depth',
-		'magnetic phase diagram',
-		'magnetic properties',
-		'magnetic resistivity',
-		'magnetic structure',
-		'magnetic susceptibility',
-		'magnetic transitions',
-		'magnetism',
-		'magnetization',
-		'magneto-optical effects',
-		'magneto-optical kerr effect',
-		'magnetostriction',
-		'mechanical properties',
-		'melting temperature',
-		'microhardness',
-		'moessbauer spectra',
-		'mohs hardness',
-		'molar volume',
-		'molecular field parameter',
-		'muon spin spectra',
-		'neel temperature',
-		'neutron energy loss spectra',
-		'non-linear optical properties',
-		'non-linear optics',
-		'nuclear magnetic resonance spectra',
-		'nuclear quadrupolar resonance spectra',
-		'optical absorption coefficient',
-		'optical absorption',
-		'optical conductivity',
-		'optical phonons',
-		'optical properties',
-		'optical spectra',
-		'orbital magnetic moment',
-		'ordinary refractive index',
-		'paraelectric curie coefficient',
-		'paraelectric curie temperature',
-		'paraelectric state',
-		'paramagnetic curie temperature',
-		'paramagnetic moment',
-		'pauli magnetic susceptibility',
-		'peritectic formation',
-		'peritectoid formation',
-		'permittivity',
-		'perturbed angular correlation',
-		'phase diagram',
-		'phase diagrams',
-		'phase transitions',
-		'phonon contribution to thermal conductivity',
-		'phonon density of states',
-		'phonon dispersion',
-		'phonon grueneisen coefficient',
-		'phonon heat capacity at constant pressure',
-		'phonons',
-		'photo-conductivity data',
-		'photo-conductivity',
-		'photoelectron emission spectra',
-		'photoluminescence spectra',
-		'physical properties',
-		'piezoelectric coefficient',
-		'piezoelectric coefficient',
-		'piezoelectricity',
-		'plasma edge',
-		'poisson ratio',
-		'power factor',
-		'pressure derivative of adiabatic bulk modulus',
-		'pressure derivative of curie temperature',
-		'pressure derivative of elastic stiffness coefficient',
-		'pressure derivative of energy gap',
-		'pressure derivative of isothermal bulk modulus',
-		'pressure derivative of neel temperature',
-		'pressure derivative of superconducting transition temperature',
-		'pressure derivative of transition temperature',
-		'pressure for magnetic transition',
-		'pressure for metal-nonmetal transition',
-		'pressure for structural transition',
-		'pyroelectric coefficient',
-		'pyroelectricity',
-		'quadrupole splitting',
-		'raman spectra',
-		'real part of magnetic permeability',
-		'real part of magnetic susceptibility',
-		'real part of optical conductivity',
-		'real part of permittivity',
-		'reflectivity',
-		'refractive index',
-		'relative cooling power',
-		'remanent induction',
-		'remanent magnetic field',
-		'remanent magnetic moment',
-		'remanent magnetization',
-		'remanent polarization',
-		'residual resistivity ratio',
-		'residual resistivity',
-		'resistivity anisotropy',
-		'resistivity',
-		'resonance spectra',
-		'saturation magnetic moment',
-		'saturation magnetization',
-		'second-harmonic generation',
-		'seebeck coefficient',
-		'shear modulus',
-		'soft-x-ray emission spectra',
-		'sound velocity',
-		'spin contribution to magnetic susceptibility',
-		'spin magnetic moment',
-		'spin-fluctuation temperature',
-		'spin-fluctuation',
-		'spin-orbit splitting of valence band',
-		'spin-resolved electron density of states at fermi level',
-		'spontaneous elastic strain',
-		'spontaneous magnetic moment',
-		'spontaneous magnetization',
-		'spontaneous polarization',
-		'static permittivity',
-		'stoner enhancement factor',
-		'stoner parameter',
-		'stoner product',
-		'structural properties',
-		'structural transition',
-		'structural transitions',
-		'superconducting transition temperature',
-		'superconductivity energy gap',
-		'superconductivity phenomena',
-		'superconductivity',
-		'temperature dependence of resistivity',
-		'temperature dependence of static permittivity',
-		'temperature derivative of elastic stiffness coefficient',
-		'temperature derivative of energy gap',
-		'temperature derivative of resistivity',
-		'temperature derivative of upper critical magnetic field',
-		'temperature for congruent melting',
-		'temperature for eutectoid decomposition',
-		'temperature for ferroelectric reordering',
-		'temperature for magnetic transition',
-		'temperature for metal-nonmetal transition',
-		'temperature for peritectic formation',
-		'temperature for peritectoid formation',
-		'temperature for structural transition',
-		'temperature-independent part of magnetic susceptibility',
-		'thermal cell parameters change',
-		'thermal conductivity',
-		'thermal energy gap',
-		'thermal expansion',
-		'thermal properties',
-		'thermal strain',
-		'thermodynamic properties',
-		'thermodynamics',
-		'thermoelectric figure of merit',
-		'thermoelectric power',
-		'total energy calculation data',
-		'transmittance',
-		'transverse sound velocity',
-		'type of magnetism',
-		'upper critical magnetic field',
-		'vacuum ultraviolet photoemission spectra',
-		'valence',
-		'van vleck contribution to magnetic susceptibility',
-		'vibrational spectra',
-		'vickers hardness number',
-		'volume change at phase transition',
-		'volume change at structural transition',
-		'volume change',
-		'volume magnetostriction',
-		'volume thermal expansion coefficient',
-		'wavelength for luminescence',
-		'wavenumber of longitudinal optical phonon',
-		'wavenumber of optical phonon',
-		'wavenumber of transverse optical phonon',
-		'work function',
-		'x-ray absorption near-edge spectra',
-		'x-ray absorption spectra',
-		'x-ray photoemission spectra',
-		'young modulus',
-	];
-
-	/*
-	 * Methods
-	 */
-	function is_numeric(n) {
-		return !isNaN(parseFloat(n)) && isFinite(n);
-	}
-
-	/*
-	 * Fix chemical formula if needed
-	 */
-	function termify_formulae(input, charred) {
-		if (input.includes('&#')) charred = true;
-		const re = charred ? /&#x208(\d);/g : /%u208(\d)/g;
-		input = charred ? input : escape(input);
-		const matches = getMatchAll(input, re);
-		if (matches) {
-			for (let i = 0; i < matches.length; i++) {
-				input = input.replace(matches[i][0], matches[i][1]);
-			}
-		}
-		return unescape(input).replace(/^\(|\)$/g, '');
-	}
-
-	/*
-	 * User input processing: brute-force similarity check
-	 */
-	function is_like_chem_formula(chk) {
-		//console.log('Checking formula');
-		const len = chk.length;
-
-		let checks;
-
-		if (len > 10) return false;
-		// this cannot be no-index chemical formula
-		else if (len === 2) {
-			checks = [[chk.substr(0, 1), chk.substr(1, 1)]];
-		} else if (len === 3) {
-			checks = [
-				[chk.substr(0, 1), chk.substr(1, 1), chk.substr(2, 1)],
-				[chk.substr(0, 1), chk.substr(1, 2)],
-				[chk.substr(0, 2), chk.substr(2, 1)],
-			];
-		} else if (len === 4) {
-			checks = [
-				[chk.substr(0, 2), chk.substr(2, 2)],
-				[chk.substr(0, 2), chk.substr(2, 1), chk.substr(3, 1)],
-				[chk.substr(0, 1), chk.substr(1, 1), chk.substr(2, 2)],
-				[chk.substr(0, 1), chk.substr(1, 2), chk.substr(3, 1)],
-				[chk.substr(0, 1), chk.substr(1, 1), chk.substr(2, 1), chk.substr(3, 1)],
-			];
-		} else if (len === 5) {
-			checks = [
-				[chk.substr(0, 1), chk.substr(1, 1), chk.substr(2, 1)],
-				[chk.substr(0, 1), chk.substr(1, 1), chk.substr(2, 2)],
-				[chk.substr(0, 2), chk.substr(2, 2), chk.substr(4, 1)],
-				[chk.substr(0, 1), chk.substr(1, 2), chk.substr(3, 2)],
-				[chk.substr(0, 1), chk.substr(1, 2), chk.substr(3, 1)],
-				[chk.substr(0, 2), chk.substr(2, 1), chk.substr(3, 1)],
-				[chk.substr(0, 2), chk.substr(2, 1), chk.substr(3, 2)],
-			];
-		} else {
-			// 6-9
-			checks = [
-				// NB too improbable to have 5 one-symbol elements in row
-				[chk.substr(0, 2), chk.substr(2, 2), chk.substr(4, 2)], // El-El-El
-				[chk.substr(0, 2), chk.substr(2, 1), chk.substr(3, 1), chk.substr(4, 2)], // El-E-E-El
-				[
-					chk.substr(0, 2),
-					chk.substr(2, 1),
-					chk.substr(3, 1),
-					chk.substr(4, 1),
-					chk.substr(5, 1),
-				], // El-E-E-E-E
-				[chk.substr(0, 2), chk.substr(2, 2), chk.substr(4, 1), chk.substr(5, 1)], // El-El-E-E
-				[chk.substr(0, 2), chk.substr(2, 2), chk.substr(4, 1), chk.substr(5, 2)], // El-El-E-El
-				[chk.substr(0, 2), chk.substr(2, 1), chk.substr(3, 2), chk.substr(5, 1)], // El-E-El-E
-				[chk.substr(0, 2), chk.substr(2, 1), chk.substr(3, 2), chk.substr(5, 2)], // El-E-El-El
-				[chk.substr(0, 1), chk.substr(1, 2), chk.substr(3, 1), chk.substr(4, 2)], // E-El-E-El
-				[
-					chk.substr(0, 1),
-					chk.substr(1, 2),
-					chk.substr(3, 1),
-					chk.substr(4, 1),
-					chk.substr(5, 1),
-				], // E-El-E-E-E
-				[chk.substr(0, 1), chk.substr(1, 2), chk.substr(3, 2), chk.substr(5, 1)], // E-El-El-E
-				[chk.substr(0, 1), chk.substr(1, 2), chk.substr(3, 2), chk.substr(5, 2)], // E-El-El-El
-				[
-					chk.substr(0, 1),
-					chk.substr(1, 1),
-					chk.substr(2, 2),
-					chk.substr(4, 1),
-					chk.substr(5, 1),
-				], // E-E-El-E-E
-				[chk.substr(0, 1), chk.substr(1, 1), chk.substr(2, 2), chk.substr(4, 2)], // E-E-El-El
-				[
-					chk.substr(0, 1),
-					chk.substr(1, 1),
-					chk.substr(2, 1),
-					chk.substr(3, 2),
-					chk.substr(5, 1),
-				], // E-E-E-El-E
-				[
-					chk.substr(0, 1),
-					chk.substr(1, 1),
-					chk.substr(2, 1),
-					chk.substr(3, 1),
-					chk.substr(4, 2),
-				], // E-E-E-E-El
-			];
-		}
-		//console.log(checks);
-
-		for (let i = 0; i < checks.length; i++) {
-			let signals = 0;
-			for (let j = 0; j < checks[i].length; j++) {
-				if (periodic_elements_cased.includes(checks[i][j])) signals++;
-
-				if (signals === checks[i].length) {
-					//console.log(checks[i]);
-					return true;
-				}
-			}
-		}
-		return false;
-	}
-
-	/*
-	 * Detect facets: formulae, elements, lattices, and some classes
-	 */
-	function try_uniword_facet(term) {
-		if (term === 'AB' || term === 'ABC' || term === 'ABCD') return ['formulae']; // special case-sensitive anonymous cases
-
-		term = term.toLowerCase();
-
-		const maybe_formula = !is_numeric(term.charAt(0));
-
-		const dmatches = getMatchAll(term, /(\d)/g);
-		if (dmatches && dmatches.length > 1 && maybe_formula) return ['formulae']; // no props with more than one digit
-
-		const imatches = getMatchAll(escape(term), /%u208(\d)/g);
-		if (imatches && imatches.length && maybe_formula) return ['formulae']; // no props with subscripts
-
-		if (periodic_elements.includes(term)) return ['elements', capitalize(term)];
-		else if (periodic_element_names.includes(term))
-			return [ 'elements', capitalize(periodic_elements[periodic_element_names.indexOf(term)]) ];
-
-		if (
-			term.includes('-') &&
-			!term.split('-').some((part) => !periodic_elements.includes(part))
-		) {
-			return [ 'elements', term.split('-').map((el) => capitalize(el)).join('-') ];
-		}
-
-		if (['element', 'elementary', 'unitary'].includes(term)) return ['classes', 'unary'];
-		else if (
-			term === 'quintenary' ||
-			term === 'quinternary' ||
-			term === 'quinternaries' ||
-			term === 'quinaries' ||
-			term === 'pentanary' ||
-			term === 'pentanaries'
-		)
-			return ['classes', 'quinary'];
-		else if (term === 'actinide' || term === 'actinides') return ['classes', 'actinoid'];
-		else if (
-			term === 'lantanide' ||
-			term === 'lantanides' ||
-			term === 'lanthanide' ||
-			term === 'lanthanides' ||
-			term === 'lantanoid' ||
-			term === 'lantanoids'
-		)
-			return ['classes', 'lanthanoid'];
-		else if (term.endsWith('ite') && term.length > 4) return ['classes'];
-
-		const chk = term
-			.replace(' structure', '')
-			.replace(' lattice', '')
-			.replace(' crystalline', '')
-			.replace(' crystal', '');
-		if (lat_fgrs.includes(chk)) return ['lattices', lat_i2p[lat_p2i[chk]]];
-
-		if (term.length <= 9 && dmatches && maybe_formula) return ['formulae']; // no SHORT props with digits (NB L0, E1)
-
-		return false;
-	}
-
-	/*
-	 * Detect facets: classes, props
-	 */
-	function try_multiword_facet(term, queue) {
-		term = term.toLowerCase();
-
-		let candidate = false,
-			combined = false,
-			orig = false;
-
-		if (queue.length) {
-			orig = term;
-			combined = true;
-			queue.forEach(function ({ input }) {
-				term = `${input} ${term}`;
-			});
-		}
-		//console.log("CHECKING TERM FOR MULTI-FACET: "+term);
-
-		candidate = check_category(term, 'classes');
-		if (candidate) {
-			if (combined) candidate.combined = true;
-			return candidate;
-		}
-
-		candidate = check_category(term, 'props');
-		if (candidate) {
-			if (combined) candidate.combined = true;
-			return candidate;
-		}
-
-		let single_chk;
-		if (term.endsWith('s')) {
-			// plural-singular fixups
-			single_chk = term.substr(0, term.length - 1);
-
-			if (!combined && single_chk.endsWith('ite'))
-				return { facet: 'classes', input: single_chk, ready: 1 };
-
-			candidate = check_category(single_chk, 'classes');
-			if (candidate) {
-				if (combined) candidate.combined = true;
-				return candidate;
-			}
-		}
-		if (term.endsWith('es')) {
-			// plural-singular fixups
-			single_chk = term.substr(0, term.length - 2);
-
-			if (single_chk === 'binari') single_chk = 'binary';
-			else if (single_chk === 'ternari') single_chk = 'ternary';
-			else if (single_chk === 'quaternari') single_chk = 'quaternary';
-
-			candidate = check_category(single_chk, 'classes');
-			if (candidate) {
-				if (combined) candidate.combined = true;
-				return candidate;
-			}
-		}
-
-		if (!candidate && orig) return try_multiword_facet(orig, []);
-		return false;
-	}
-
-	/*
-	 * Utility algo
-	 */
-	function check_category(term, category) {
-		const host = category === 'classes' ? mpds_classes : mpds_props;
-
-		if (host.includes(term)) return { facet: category, input: term, ready: 1 };
-
-		const len = host.length;
-		const re = new RegExp(`(?:^|\\s)(${term})(?=\\s|$)`);
-		let idx;
-		for (let i = 0; i < len; i++) {
-			idx = host[i].search(re);
-			if (idx === 0) {
-				//console.log("Found unstrict match in "+category+" with <"+host[i]+">");
-				return { facet: category, input: term, anew: 1 };
-			}
-		}
-		return false;
-	}
-
-	/*
-	 * Chemical formula: ABC3 vs. SrTiO3
-	 */
-	function is_formula_anonymous(formula) {
-		const detect = new RegExp(/A(\d{0,3})B(\d{0,3})(C(\d{0,3})(D(\d{0,3}))?)?\b/);
-		return formula.charAt(0) === 'A' && detect.test(formula);
-	}
-
-	/*
-	 * Get center and ligand information from a string
-	 */
-	function parse_ligand(string, start) {
-
-		const center = string.slice(0, start).toLowerCase();
-
-		if (string.slice(start, start + 1).toLowerCase() === 'x' && string.slice(start, start + 2).toLowerCase() !== 'xe'){
-			if (string.slice(start).length === 1) return [center, 'X'];
-
-			return [center, 'X' + string.slice(start + 1)];
-		}
-
-		if (string.length === start) return [center, 'X'];
-
-		const remainder = string.slice(start);
-
-		if (is_numeric(remainder.slice(0, 1)) && start === 2)
-			return parse_ligand(string, 1);
-
-		return [center, capitalize(remainder)];
-	}
-
-	/*
-	 * Get center and ligand information from a string
-	 */
-	function _parse_aeatoms(string) {
-
-		const pos = string.indexOf('-');
-
-		if (pos !== -1){
-			const center = string.slice(0, pos),
-				ligand = string.slice(pos + 1);
-
-			if (center.length > 2) return false;
-
-			return parse_ligand(center + ligand, center.length);
-		}
-
-		const trials = [2, 1],
-			periodic_elements_xed = ["x"].concat(periodic_elements);
-
-		for (let start = 0; start < 2; start++){
-			if (string.length >= trials[start] && periodic_elements_xed.indexOf(string.slice(0, trials[start]).toLowerCase()) !== -1){
-				return parse_ligand(string, trials[start]);
-			}
-		}
-		return false;
-	}
-
-	/**
-	 * Get center and ligand information from a string
-	 */
-	function parse_aeatoms(string) {
-
-		const parsed = _parse_aeatoms(string);
-
-		if (!parsed) return ['?', '?'];
-
-		return [
-			capitalize(parsed[0]),
-			formula_to_tags(parsed[1])
-		];
-	}
-
-	/*
-	 * Add HTML tags to a chemical formula as a string
-	 */
-	function formula_to_tags(string) {
-
-		let sub = false,
-			html_formula = '';
-
-		for (let i = 0, len = string.length; i < len; i++){
-			if (is_numeric(string[i]) || string[i] === '.'){
-				if (!sub){
-					html_formula += '<sub>';
-					sub = true;
-				}
-			} else {
-				if (sub){
-					html_formula += '</sub>';
-					sub = false;
-				}
-			}
-			html_formula += string[i];
-		}
-		if (sub) html_formula += '</sub>';
-		return html_formula;
-	}
-
-	/*
-	 * User input processing: main algorithm
-	 */
-	function guess(inputstr) {
-
-		// *pseudo_numerics*
-		/*if (inputstr.includes('c/a ') || inputstr.includes('a/b ') || inputstr.includes('b/c ')) {
+  /*
+   * Definitions
+   */
+  const stop_words = [
+    'a',
+    'about',
+    'above',
+    'after',
+    'again',
+    'against',
+    'all',
+    'am',
+    'an',
+    'and',
+    'any',
+    'are',
+    "aren't",
+    'as',
+    'at',
+    'be',
+    'because',
+    'been',
+    'before',
+    'being',
+    'below',
+    'between',
+    'both',
+    'but',
+    'by',
+    "can't",
+    'cannot',
+    'could',
+    "couldn't",
+    'did',
+    "didn't",
+    'do',
+    'does',
+    "doesn't",
+    'doing',
+    "don't",
+    'down',
+    'during',
+    'each',
+    'few',
+    'for',
+    'from',
+    'further',
+    'had',
+    "hadn't",
+    'has',
+    "hasn't",
+    'have',
+    "haven't",
+    'having',
+    'he',
+    "he'd",
+    "he'll",
+    "he's",
+    'her',
+    'here',
+    "here's",
+    'hers',
+    'herself',
+    'him',
+    'himself',
+    'his',
+    'how',
+    "how's",
+    'i',
+    "i'd",
+    "i'll",
+    "i'm",
+    "i've",
+    'if',
+    'in',
+    'into',
+    'is',
+    "isn't",
+    'it',
+    "it's",
+    'its',
+    'itself',
+    "let's",
+    'me',
+    'more',
+    'most',
+    "mustn't",
+    'my',
+    'myself',
+    'no',
+    'nor',
+    'not',
+    'of',
+    'off',
+    'on',
+    'once',
+    'only',
+    'or',
+    'other',
+    'ought',
+    'our',
+    'ours',
+    'ourselves',
+    'out',
+    'over',
+    'own',
+    'same',
+    "shan't",
+    'she',
+    "she'd",
+    "she'll",
+    "she's",
+    'should',
+    "shouldn't",
+    'so',
+    'some',
+    'such',
+    'than',
+    'that',
+    "that's",
+    'the',
+    'their',
+    'theirs',
+    'them',
+    'themselves',
+    'then',
+    'there',
+    "there's",
+    'these',
+    'they',
+    "they'd",
+    "they'll",
+    "they're",
+    "they've",
+    'this',
+    'those',
+    'through',
+    'to',
+    'too',
+    'u',
+    'under',
+    'until',
+    'up',
+    'very',
+    'was',
+    "wasn't",
+    'we',
+    "we'd",
+    "we'll",
+    "we're",
+    "we've",
+    'were',
+    "weren't",
+    'what',
+    "what's",
+    'when',
+    "when's",
+    'where',
+    "where's",
+    'which',
+    'while',
+    'who',
+    "who's",
+    'whom',
+    'why',
+    "why's",
+    'with',
+    "won't",
+    'would',
+    "wouldn't",
+    'you',
+    "you'd",
+    "you'll",
+    "you're",
+    "you've",
+    'your',
+    'yours',
+    'yourself',
+    'yourselves',
+  ]; /* exact */
+
+  const arity_keys = [
+    null,
+    'unary',
+    'binary',
+    'ternary',
+    'quaternary',
+    'quinary',
+    'multinary',
+    'multinary',
+    'multinary',
+    'multinary',
+    'multinary',
+  ]; // NB null is for "0-ary"
+
+  const periodic_elements = [
+    'h',
+    'he',
+    'li',
+    'be',
+    'b',
+    'c',
+    'n',
+    'o',
+    'f',
+    'ne',
+    'na',
+    'mg',
+    'al',
+    'si',
+    'p',
+    's',
+    'cl',
+    'ar',
+    'k',
+    'ca',
+    'sc',
+    'ti',
+    'v',
+    'cr',
+    'mn',
+    'fe',
+    'co',
+    'ni',
+    'cu',
+    'zn',
+    'ga',
+    'ge',
+    'as',
+    'se',
+    'br',
+    'kr',
+    'rb',
+    'sr',
+    'y',
+    'zr',
+    'nb',
+    'mo',
+    'tc',
+    'ru',
+    'rh',
+    'pd',
+    'ag',
+    'cd',
+    'in',
+    'sn',
+    'sb',
+    'te',
+    'i',
+    'xe',
+    'cs',
+    'ba',
+    'la',
+    'ce',
+    'pr',
+    'nd',
+    'pm',
+    'sm',
+    'eu',
+    'gd',
+    'tb',
+    'dy',
+    'ho',
+    'er',
+    'tm',
+    'yb',
+    'lu',
+    'hf',
+    'ta',
+    'w',
+    're',
+    'os',
+    'ir',
+    'pt',
+    'au',
+    'hg',
+    'tl',
+    'pb',
+    'bi',
+    'po',
+    'at',
+    'rn',
+    'fr',
+    'ra',
+    'ac',
+    'th',
+    'pa',
+    'u',
+    'np',
+    'pu',
+    'am',
+    'cm',
+    'bk',
+    'cf',
+    'es',
+    'fm',
+    'md',
+    'no',
+    'lr',
+    'rf',
+    'db',
+    'sg',
+    'bh',
+    'hs',
+    'mt',
+    'ds',
+    'rg',
+    'cn',
+    'nh',
+    'fl',
+    'mc',
+    'lv',
+    'ts',
+    'og',
+  ]; /* exact */
+
+  const periodic_elements_cased = periodic_elements.map(function (x) {
+    return capitalize(x);
+  });
+
+  const periodic_element_names = [
+    'hydrogen',
+    'helium',
+    'lithium',
+    'beryllium',
+    'boron',
+    'carbon',
+    'nitrogen',
+    'oxygen',
+    'fluorine',
+    'neon',
+    'sodium',
+    'magnesium',
+    'aluminium',
+    'silicon',
+    'phosphorus',
+    'sulfur',
+    'chlorine',
+    'argon',
+    'potassium',
+    'calcium',
+    'scandium',
+    'titanium',
+    'vanadium',
+    'chromium',
+    'manganese',
+    'iron',
+    'cobalt',
+    'nickel',
+    'copper',
+    'zinc',
+    'gallium',
+    'germanium',
+    'arsenic',
+    'selenium',
+    'bromine',
+    'krypton',
+    'rubidium',
+    'strontium',
+    'yttrium',
+    'zirconium',
+    'niobium',
+    'molybdenum',
+    'technetium',
+    'ruthenium',
+    'rhodium',
+    'palladium',
+    'silver',
+    'cadmium',
+    'indium',
+    'tin',
+    'antimony',
+    'tellurium',
+    'iodine',
+    'xenon',
+    'caesium',
+    'barium',
+    'lanthanum',
+    'cerium',
+    'praseodymium',
+    'neodymium',
+    'promethium',
+    'samarium',
+    'europium',
+    'gadolinium',
+    'terbium',
+    'dysprosium',
+    'holmium',
+    'erbium',
+    'thulium',
+    'ytterbium',
+    'lutetium',
+    'hafnium',
+    'tantalum',
+    'tungsten',
+    'rhenium',
+    'osmium',
+    'iridium',
+    'platinum',
+    'gold',
+    'mercury',
+    'thallium',
+    'lead',
+    'bismuth',
+    'polonium',
+    'astatine',
+    'radon',
+    'francium',
+    'radium',
+    'actinium',
+    'thorium',
+    'protactinium',
+    'uranium',
+    'neptunium',
+    'plutonium',
+    'americium',
+    'curium',
+    'berkelium',
+    'californium',
+    'einsteinium',
+    'fermium',
+    'mendelevium',
+    'nobelium',
+    'lawrencium',
+    'rutherfordium',
+    'dubnium',
+    'seaborgium',
+    'bohrium',
+    'hassium',
+    'meitnerium',
+    'darmstadium',
+    'roentgenium',
+    'copernicium',
+    'nihonium',
+    'flerovium',
+    'moscovium',
+    'livermorium',
+    'tennessine',
+    'oganesson',
+  ]; /* fuzzy */
+
+  const lat_p2i = {
+    cubic: 1,
+    hexagonal: 2,
+    trigonal: 3,
+    tetragonal: 4,
+    orthorhombic: 5,
+    monoclinic: 6,
+    triclinic: 7,
+    rhombohedral: 3,
+    cub: 1,
+    hex: 2,
+    hexag: 2,
+    trig: 3,
+    tet: 4,
+    tetr: 4,
+    tetrag: 4,
+    orth: 5,
+    ortho: 5,
+    monocl: 6,
+    tric: 7,
+    tricl: 7,
+    rhom: 3,
+    rhomb: 3,
+  };
+  const lat_fgrs = Object.keys(lat_p2i);
+  const lat_i2p = {
+    1: 'cubic',
+    2: 'hexagonal',
+    3: 'trigonal',
+    4: 'tetragonal',
+    5: 'orthorhombic',
+    6: 'monoclinic',
+    7: 'triclinic',
+  };
+
+  const mpds_classes = [
+    'ab initio calculations',
+    'ab initio literature',
+    'actinoid',
+    'adamantane',
+    'aegirine',
+    'alkali',
+    'alkaline',
+    'allargentum',
+    'almandine',
+    'alum',
+    'alunogen',
+    'amide',
+    'analcime',
+    'anatase',
+    'anorpiment',
+    'anorthoclase',
+    'antiferroelectric',
+    'antiferromagnet',
+    'antiferromagnetic',
+    'arsenate',
+    'arsenide',
+    'ashcroftine',
+    'auricupride',
+    'aurocupride',
+    'azide',
+    'baileychlore',
+    'bariopyrochlore',
+    'baryte',
+    'beryl',
+    'beta-alumina',
+    'beta-boron',
+    'biguanide',
+    'binary',
+    'birefringent',
+    'borane',
+    'borate',
+    'borax',
+    'boride',
+    'borocarbide',
+    'borohydride',
+    'boronitride',
+    'botryogen',
+    'bromanilate',
+    'bromide',
+    'bromoimide',
+    'calomel',
+    'carbamate',
+    'carbide',
+    'carbonate',
+    'carbonyl',
+    'carboxylate',
+    'celestine',
+    'cell and atoms',
+    'cell-only',
+    'celsian',
+    'cesiokenopyrochlore',
+    'chalcogen',
+    'charge-density wave state',
+    'chevrel',
+    'chimney-ladder',
+    'chloranilate',
+    'chlorate',
+    'chloride',
+    'chloritoid',
+    'chlorosulfate',
+    'chromate',
+    'chrysoberyl',
+    'chrysotile',
+    'cinnabar',
+    'clathrate',
+    'clinochlore',
+    'clinoclase',
+    'clodronate',
+    'close-packed',
+    'cluster glass',
+    'colossal magnetoresistance',
+    'conductor',
+    'corundum',
+    'cosmochlor',
+    'croconate violet',
+    'croconate',
+    'cryptomelane',
+    'cuprate',
+    'cuspidine',
+    'cyamelurate',
+    'cyanamide',
+    'cyanamidonitrate',
+    'cyananilate',
+    'cyanide',
+    'cyanotetrazolate',
+    'cyanoureate',
+    'cyanurate',
+    'cyprine',
+    'davyne',
+    'deuteride',
+    'deuterium',
+    'devilline',
+    'diamagnetic',
+    'diamond',
+    'diarsenate',
+    'diaspore',
+    'diazanide',
+    'diazenide',
+    'dichromate',
+    'digermanate',
+    'diiodobromide',
+    'dinitramide',
+    'diopside',
+    'dioptase',
+    'dioxobromate',
+    'dioxoiodate',
+    'dioxosulfate',
+    'dioxothiosulfate',
+    'diphosphate',
+    'diphosphonate',
+    'dipolyhedral',
+    'diselenate',
+    'disilicate',
+    'disordered',
+    'disulfate',
+    'dithiocarbamate',
+    'dithiocarbonate',
+    'dithionate',
+    'dithiooxalate',
+    'dithiophosphate',
+    'dithiosquarate',
+    'divanadate',
+    'epidote',
+    'euchlorine',
+    'euclase',
+    'eudialyte',
+    'eulytine',
+    'fermi liquid',
+    'feroxihyte',
+    'feroxyhyte',
+    'ferrielectric',
+    'ferrimagnet',
+    'ferroelastic',
+    'ferroelectric',
+    'ferromagnet',
+    'ferromagnetic',
+    'fluor-schorl',
+    'fluoride',
+    'fluoroborate',
+    'frank-kasper',
+    'friauf-laves',
+    'fulleride',
+    'fulminate',
+    'galena',
+    'gamma-brass',
+    'garnet',
+    'giant magnetocaloric effect',
+    'gismondine',
+    'glaucodot',
+    'glaucophane',
+    'grossular',
+    'guanidinate',
+    'gypsum',
+    'hafnon',
+    'half metal',
+    'halogen',
+    'hard magnet',
+    'harmotome',
+    'haueyne',
+    'heavy fermion',
+    'hedyphane',
+    'helimagnet',
+    'helvine',
+    'hexasulfate',
+    'hexathionate',
+    'host-guest',
+    'humboldtine',
+    'hydrate',
+    'hydride',
+    'hydroxide',
+    'hypercinnabar',
+    'hypophosphate',
+    'ice',
+    'imide',
+    'iminate',
+    'intercalation',
+    'intermediate valence',
+    'intermetallic',
+    'iodate',
+    'iodide',
+    'ionic conductor',
+    'iridium',
+    'isoferroplatinum',
+    'isopolyhedral',
+    'isothermal section',
+    'kornerupine',
+    'kosmochlor',
+    'lanthanoid',
+    'lavendulan',
+    'levyne',
+    'lime',
+    'liquidus projection',
+    'litharge',
+    'lithiophosphate',
+    'luminescent',
+    'machine learning',
+    'machine-learning',
+    'magnesiochloritoid',
+    'magnetoelastic',
+    'magnetoelectric',
+    'manganate',
+    'massicot',
+    'mellitate',
+    'melonate',
+    'metacinnabar',
+    'metal',
+    'metalloid',
+    'metamagnet',
+    'metavoltine',
+    'mica',
+    'microcline',
+    'microline',
+    'mictomagnet',
+    'minium',
+    'molybdate',
+    'multiferroic',
+    'multinary',
+    'nasicon',
+    'natron',
+    'natrophosphate',
+    'natroxalate',
+    'negative thermal expansion',
+    'nepheline',
+    'nickeline',
+    'niobocarbide',
+    'niter',
+    'nitranilate',
+    'nitrate',
+    'nitratine',
+    'nitride',
+    'nitroformate',
+    'noble gas',
+    'non disordered',
+    'non-disordered',
+    'non-linear optics',
+    'nonaflate',
+    'nonmetal',
+    'nordenskioeldine',
+    'nosean',
+    'olivine',
+    'optically isotropic',
+    'organic',
+    'orpiment',
+    'orthoborate',
+    'orthoclase',
+    'orthogermanate',
+    'orthonitrate',
+    'orthophosphate',
+    'orthosilicate',
+    'oxalate',
+    'oxamate',
+    'oxide',
+    'oxoiodate',
+    'oxonitrate',
+    'oxotetrazolate',
+    'oxy-schorl',
+    'ozonide',
+    'palladogermanide',
+    'paracelsian',
+    'paramagnet',
+    'pararealgar',
+    'pauli paramagnet',
+    'peer review',
+    'peer reviewed',
+    'peer-review',
+    'peer-reviewed',
+    'perchlorate',
+    'perhydrate',
+    'periclase',
+    'periodate',
+    'permanganate',
+    'pernitride',
+    'peroxide',
+    'pertechnetate',
+    'phosphate',
+    'phosphide',
+    'phosphinate',
+    'phosphonate',
+    'photocatalyst',
+    'photovoltaic effect',
+    'photovoltaic',
+    'piezoelectric',
+    'pnictogen',
+    'polaron conductor',
+    'polycrase',
+    'potassic',
+    'prism',
+    'prussian blue',
+    'pseudorutile',
+    'pyrochlore',
+    'pyroelectric',
+    'pyrope',
+    'quadridavyne',
+    'quartz',
+    'quasicrystal',
+    'quaternary',
+    'quinary',
+    'radioactive',
+    'rare earth',
+    'realgar',
+    'refractory',
+    'relaxor',
+    'retzian',
+    'rhodarsenide',
+    'rhodizonate',
+    'rhomboclase',
+    'rocksalt',
+    'rubicline',
+    'ruddlesden-popper',
+    'rutheniridosmine',
+    'rutherfordine',
+    'rutile',
+    'salammoniac',
+    'sanidine',
+    'sapphirine',
+    'sarcopside',
+    'schorl',
+    'selenide',
+    'selenidel',
+    'semiconductor',
+    'semimetal',
+    'shape memory effect',
+    'shape memory',
+    'siderotil',
+    'silanide',
+    'silicate',
+    'silicide',
+    'sillen-aurivillius',
+    'skyrmion',
+    'solidus projection',
+    'solvus projection',
+    'spessartine',
+    'spin glass',
+    'spinel',
+    'spodumene',
+    'squarate',
+    'steenstrupine',
+    'stibarsen',
+    'subsolidus relations',
+    'sulfamate',
+    'sulfamide',
+    'sulfate',
+    'sulfide',
+    'sulfidel',
+    'sulfinylamide',
+    'superconductor',
+    'superhard',
+    'superionic conductor',
+    'superoxide',
+    'talc',
+    'tantalcarbide',
+    'tellurantimony',
+    'telluride',
+    'ternary',
+    'tetrahedron',
+    'thermoelectric',
+    'thiocyanate',
+    'thiocyanurate',
+    'thiophosphate',
+    'thiosulfate',
+    'topaz',
+    'topological insulator',
+    'tourmaline',
+    'transitional',
+    'triflate',
+    'tripolyhedral',
+    'triteride',
+    'trithionate',
+    'tritide',
+    'tritium',
+    'trona',
+    'tungstate',
+    'turquoise',
+    'ulvoespinel',
+    'unary',
+    'uranophane',
+    'uranopolycrase',
+    'urate',
+    'urea',
+    'ureate',
+    'van vleck paramagnet',
+    'vanadate',
+    'vertical section',
+    'violurate',
+    'xenotime',
+    'zincobotryogen',
+    'zircon',
+    'zircosulfate',
+  ];
+
+  // Mapping of common element groups / periods to element symbol arrays
+  const ELEMENT_GROUPS_MAP = {
+    'period 1': ['H', 'He'],
+    'period 2': ['Li', 'Be', 'B', 'C', 'N', 'O', 'F', 'Ne'],
+    'period 3': ['Na', 'Mg', 'Al', 'Si', 'P', 'S', 'Cl', 'Ar'],
+    'period 4': [
+      'K',
+      'Ca',
+      'Sc',
+      'Ti',
+      'V',
+      'Cr',
+      'Mn',
+      'Fe',
+      'Co',
+      'Ni',
+      'Cu',
+      'Zn',
+      'Ga',
+      'Ge',
+      'As',
+      'Se',
+      'Br',
+      'Kr',
+    ],
+    'period 5': [
+      'Rb',
+      'Sr',
+      'Y',
+      'Zr',
+      'Nb',
+      'Mo',
+      'Tc',
+      'Ru',
+      'Rh',
+      'Pd',
+      'Ag',
+      'Cd',
+      'In',
+      'Sn',
+      'Sb',
+      'Te',
+      'I',
+      'Xe',
+    ],
+    'period 6': [
+      'Cs',
+      'Ba',
+      'Lu',
+      'Hf',
+      'Ta',
+      'W',
+      'Re',
+      'Os',
+      'Ir',
+      'Pt',
+      'Au',
+      'Hg',
+      'Tl',
+      'Pb',
+      'Bi',
+      'Po',
+      'At',
+      'Rn',
+    ],
+    'period 7': ['Fr', 'Ra', 'Lr'],
+
+    lanthanide: [
+      'La',
+      'Ce',
+      'Pr',
+      'Nd',
+      'Pm',
+      'Sm',
+      'Eu',
+      'Gd',
+      'Tb',
+      'Dy',
+      'Ho',
+      'Er',
+      'Tm',
+      'Yb',
+    ],
+    actinide: [
+      'Ac',
+      'Th',
+      'Pa',
+      'U',
+      'Np',
+      'Pu',
+      'Am',
+      'Cm',
+      'Bk',
+      'Cf',
+      'Es',
+      'Fm',
+      'Md',
+      'No',
+    ],
+
+    alkali: ['H', 'Li', 'Na', 'K', 'Rb', 'Cs', 'Fr'],
+    alkaline: ['Be', 'Mg', 'Ca', 'Sr', 'Ba', 'Ra'],
+
+    'group 3': ['Sc', 'Y', 'Lu', 'Lr'],
+    'group 4': ['Ti', 'Zr', 'Hf'],
+    'group 5': ['V', 'Nb', 'Ta'],
+    'group 6': ['Cr', 'Mo', 'W'],
+    'group 7': ['Mn', 'Tc', 'Re'],
+    'group 8': ['Fe', 'Ru', 'Os'],
+    'group 9': ['Co', 'Rh', 'Ir'],
+    'group 10': ['Ni', 'Pd', 'Pt'],
+    'group 11': ['Cu', 'Ag', 'Au', 'Rg'],
+    'group 12': ['Zn', 'Cd', 'Hg'],
+
+    triels: ['B', 'Al', 'Ga', 'In', 'Tl'],
+    triel: ['B', 'Al', 'Ga', 'In', 'Tl'],
+    tetrels: ['C', 'Si', 'Ge', 'Sn', 'Pb', 'Fl'],
+    tetrel: ['C', 'Si', 'Ge', 'Sn', 'Pb', 'Fl'],
+    pnictogen: ['N', 'P', 'As', 'Sb', 'Bi'],
+    chalcogen: ['O', 'S', 'Se', 'Te', 'Po'],
+    halogen: ['F', 'Cl', 'Br', 'I', 'At'],
+    'noble gas': ['He', 'Ne', 'Ar', 'Kr', 'Xe', 'Rn'],
+  };
+
+  function getGroupElements(name) {
+    if (!name) return null;
+    const key = name.toLowerCase().trim();
+    // normalize simple plural forms
+    if (ELEMENT_GROUPS_MAP[key]) return ELEMENT_GROUPS_MAP[key];
+    const singular = key.replace(/s$/i, '');
+    if (ELEMENT_GROUPS_MAP[singular]) return ELEMENT_GROUPS_MAP[singular];
+    return null;
+  }
+
+  const mpds_props = [
+    'acceptor concentration',
+    'acceptor to donor concentration',
+    'activation energy',
+    'adiabatic bulk modulus',
+    'angle-resolved photoelectron spectra',
+    'atomic structure',
+    'band gap',
+    'birefringence',
+    'bremsstrahlung isochromat spectra',
+    'charge carrier concentration',
+    'charge carrier mobility',
+    'charge transfer',
+    'charge-density wave',
+    'charge-transfer energy',
+    'coefficient of schottky term in heat capacity',
+    'coercive electric field',
+    'coercive field',
+    'coherence length',
+    'cohesive energy',
+    'compressibility',
+    'conductivity',
+    'core-electron contribution to magnetic susceptibility',
+    'critical current density',
+    'critical magnetic field',
+    'crystal electric field parameter',
+    'crystal electric field parameters',
+    'crystal electric field splitting',
+    'crystal electric field',
+    'crystal field level',
+    'crystalline structure',
+    'crystal cell',
+    'crystal structure',
+    'curie coefficient',
+    'curie temperature',
+    'curie-weiss paramagnetism',
+    'debye temperature',
+    'decomposition temperature',
+    'decomposition',
+    'diamagnetic contribution to magnetic susceptibility',
+    'dielectric constant',
+    'dielectric loss tangent',
+    'diffusion',
+    'donor concentration',
+    'donor energy',
+    'effective charge',
+    'effective electron number',
+    'effective mass of electrons to holes ratio',
+    'effective mass of electrons',
+    'effective mass',
+    'einstein temperature',
+    'elastic compliance',
+    'elastic moduli',
+    'elastic stiffness coefficient',
+    'elasticity',
+    'electric field gradient',
+    'electric polarization',
+    'electrical conductivity',
+    'electric properties',
+    'electrical properties',
+    'electrical resistivity',
+    'electrochemical impedance spectroscopy',
+    'electron density maps',
+    'electron density of states at fermi level',
+    'electron density of states',
+    'electron energy band structure',
+    'electron energy loss spectra',
+    'electron grueneisen coefficient',
+    'electron mobility',
+    'electron paramagnetic resonance spectra',
+    'electron spin resonance spectra',
+    'electron-phonon interaction parameter',
+    'electronic contribution to heat capacity',
+    'electronic contribution to thermal conductivity',
+    'electronic energy gap',
+    'electronic heat capacity coefficient',
+    'electronic properties',
+    'energy at fermi level',
+    'energy band structure',
+    'energy gap for direct transition',
+    'energy gap for indirect transition',
+    'energy level diagram',
+    'energy of optical phonon',
+    'energy product',
+    'energy',
+    'enthalpy change at melting point',
+    'enthalpy change at phase transition',
+    'enthalpy change at structural transition',
+    'enthalpy change',
+    'enthalpy of formation',
+    'enthalpy of reaction',
+    'enthalpy',
+    'entropy change at melting point',
+    'entropy change at phase transition',
+    'entropy of formation',
+    'entropy of reaction',
+    'entropy',
+    'eutectoid decomposition',
+    'exchange field',
+    'exchange interaction parameter',
+    'exciton energy',
+    'extended x-ray absorption fine structure',
+    'extraordinary refractive index',
+    'fermi energy',
+    'fermi surface',
+    'ferroelasticity',
+    'ferroelectric curie temperature',
+    'ferroelectric hysteresis',
+    'ferroelectric neel temperature',
+    'ferroelectric transitions',
+    'field dependence of resistivity',
+    'figure of merit',
+    'freezing temperature for spin glass',
+    'fusion',
+    'gibbs energy change',
+    'gibbs energy of formation',
+    'gibbs energy of reaction',
+    'ginzburg-landau parameter',
+    'grueneisen coefficient',
+    'gruneisen coefficient',
+    'hall coefficient',
+    'hall effect',
+    'hall mobility',
+    'hardness',
+    'heat capacity at constant pressure',
+    'heat capacity at constant volume',
+    'heat capacity coefficient',
+    'heat capacity discontinuity at structural transition',
+    'heat capacity discontinuity at superconducting transition',
+    'heat capacity discontinuity',
+    'heat capacity',
+    'high-frequency permittivity',
+    'hole mobility',
+    'hydrogen diffusion',
+    'hyperfine magnetic field',
+    'imaginary part of magnetic susceptibility',
+    'imaginary part of permittivity',
+    'inelastic neutron scattering',
+    'inelastic x-ray scattering',
+    'infrared spectra',
+    'ionic conductivity',
+    'irreversibility field',
+    'isomer shift',
+    'isothermal bulk modulus',
+    'isothermal linear compressibility',
+    'isothermal volume compressibility',
+    'knoop hardness',
+    'kondo behavior',
+    'kondo temperature',
+    'lattice',
+    'linear magnetostriction',
+    'linear thermal expansion coefficient',
+    'longitudinal sound velocity',
+    'longitudinal-mode elastic coefficient',
+    'lorentz number',
+    'lower critical magnetic field',
+    'lowest temperature of investigation',
+    'luminescence lifetime',
+    'luminescence',
+    'magnetic anisotropy field',
+    'magnetic anisotropy',
+    'magnetic circular x-ray dichroism',
+    'magnetic dichroism',
+    'magnetic direction',
+    'magnetic entropy',
+    'magnetic field for magnetic transition',
+    'magnetic field for structural transition',
+    'magnetic heat capacity',
+    'magnetic hysteresis',
+    'magnetic moment',
+    'magnetic order',
+    'magnetic penetration depth',
+    'magnetic phase diagram',
+    'magnetic properties',
+    'magnetic resistivity',
+    'magnetic structure',
+    'magnetic susceptibility',
+    'magnetic transitions',
+    'magnetism',
+    'magnetization',
+    'magneto-optical effects',
+    'magneto-optical kerr effect',
+    'magnetostriction',
+    'mechanical properties',
+    'melting temperature',
+    'microhardness',
+    'moessbauer spectra',
+    'mohs hardness',
+    'molar volume',
+    'molecular field parameter',
+    'muon spin spectra',
+    'neel temperature',
+    'neutron energy loss spectra',
+    'non-linear optical properties',
+    'non-linear optics',
+    'nuclear magnetic resonance spectra',
+    'nuclear quadrupolar resonance spectra',
+    'optical absorption coefficient',
+    'optical absorption',
+    'optical conductivity',
+    'optical phonons',
+    'optical properties',
+    'optical spectra',
+    'orbital magnetic moment',
+    'ordinary refractive index',
+    'paraelectric curie coefficient',
+    'paraelectric curie temperature',
+    'paraelectric state',
+    'paramagnetic curie temperature',
+    'paramagnetic moment',
+    'pauli magnetic susceptibility',
+    'peritectic formation',
+    'peritectoid formation',
+    'permittivity',
+    'perturbed angular correlation',
+    'phase diagram',
+    'phase diagrams',
+    'phase transitions',
+    'phonon contribution to thermal conductivity',
+    'phonon density of states',
+    'phonon dispersion',
+    'phonon grueneisen coefficient',
+    'phonon heat capacity at constant pressure',
+    'phonons',
+    'photo-conductivity data',
+    'photo-conductivity',
+    'photoelectron emission spectra',
+    'photoluminescence spectra',
+    'physical properties',
+    'piezoelectric coefficient',
+    'piezoelectric coefficient',
+    'piezoelectricity',
+    'plasma edge',
+    'poisson ratio',
+    'power factor',
+    'pressure derivative of adiabatic bulk modulus',
+    'pressure derivative of curie temperature',
+    'pressure derivative of elastic stiffness coefficient',
+    'pressure derivative of energy gap',
+    'pressure derivative of isothermal bulk modulus',
+    'pressure derivative of neel temperature',
+    'pressure derivative of superconducting transition temperature',
+    'pressure derivative of transition temperature',
+    'pressure for magnetic transition',
+    'pressure for metal-nonmetal transition',
+    'pressure for structural transition',
+    'pyroelectric coefficient',
+    'pyroelectricity',
+    'quadrupole splitting',
+    'raman spectra',
+    'real part of magnetic permeability',
+    'real part of magnetic susceptibility',
+    'real part of optical conductivity',
+    'real part of permittivity',
+    'reflectivity',
+    'refractive index',
+    'relative cooling power',
+    'remanent induction',
+    'remanent magnetic field',
+    'remanent magnetic moment',
+    'remanent magnetization',
+    'remanent polarization',
+    'residual resistivity ratio',
+    'residual resistivity',
+    'resistivity anisotropy',
+    'resistivity',
+    'resonance spectra',
+    'saturation magnetic moment',
+    'saturation magnetization',
+    'second-harmonic generation',
+    'seebeck coefficient',
+    'shear modulus',
+    'soft-x-ray emission spectra',
+    'sound velocity',
+    'spin contribution to magnetic susceptibility',
+    'spin magnetic moment',
+    'spin-fluctuation temperature',
+    'spin-fluctuation',
+    'spin-orbit splitting of valence band',
+    'spin-resolved electron density of states at fermi level',
+    'spontaneous elastic strain',
+    'spontaneous magnetic moment',
+    'spontaneous magnetization',
+    'spontaneous polarization',
+    'static permittivity',
+    'stoner enhancement factor',
+    'stoner parameter',
+    'stoner product',
+    'structural properties',
+    'structural transition',
+    'structural transitions',
+    'superconducting transition temperature',
+    'superconductivity energy gap',
+    'superconductivity phenomena',
+    'superconductivity',
+    'temperature dependence of resistivity',
+    'temperature dependence of static permittivity',
+    'temperature derivative of elastic stiffness coefficient',
+    'temperature derivative of energy gap',
+    'temperature derivative of resistivity',
+    'temperature derivative of upper critical magnetic field',
+    'temperature for congruent melting',
+    'temperature for eutectoid decomposition',
+    'temperature for ferroelectric reordering',
+    'temperature for magnetic transition',
+    'temperature for metal-nonmetal transition',
+    'temperature for peritectic formation',
+    'temperature for peritectoid formation',
+    'temperature for structural transition',
+    'temperature-independent part of magnetic susceptibility',
+    'thermal cell parameters change',
+    'thermal conductivity',
+    'thermal energy gap',
+    'thermal expansion',
+    'thermal properties',
+    'thermal strain',
+    'thermodynamic properties',
+    'thermodynamics',
+    'thermoelectric figure of merit',
+    'thermoelectric power',
+    'total energy calculation data',
+    'transmittance',
+    'transverse sound velocity',
+    'type of magnetism',
+    'upper critical magnetic field',
+    'vacuum ultraviolet photoemission spectra',
+    'valence',
+    'van vleck contribution to magnetic susceptibility',
+    'vibrational spectra',
+    'vickers hardness number',
+    'volume change at phase transition',
+    'volume change at structural transition',
+    'volume change',
+    'volume magnetostriction',
+    'volume thermal expansion coefficient',
+    'wavelength for luminescence',
+    'wavenumber of longitudinal optical phonon',
+    'wavenumber of optical phonon',
+    'wavenumber of transverse optical phonon',
+    'work function',
+    'x-ray absorption near-edge spectra',
+    'x-ray absorption spectra',
+    'x-ray photoemission spectra',
+    'young modulus',
+  ];
+
+  /*
+   * Methods
+   */
+  function is_numeric(n) {
+    return !isNaN(parseFloat(n)) && isFinite(n);
+  }
+
+  /*
+   * Fix chemical formula if needed
+   */
+  function termify_formulae(input, charred) {
+    if (input.includes('&#')) charred = true;
+    const re = charred ? /&#x208(\d);/g : /%u208(\d)/g;
+    input = charred ? input : escape(input);
+    const matches = getMatchAll(input, re);
+    if (matches) {
+      for (let i = 0; i < matches.length; i++) {
+        input = input.replace(matches[i][0], matches[i][1]);
+      }
+    }
+    return unescape(input).replace(/^\(|\)$/g, '');
+  }
+
+  /*
+   * User input processing: brute-force similarity check
+   */
+  function is_like_chem_formula(chk) {
+    //console.log('Checking formula');
+    const len = chk.length;
+
+    let checks;
+
+    if (len > 10) return false;
+    // this cannot be no-index chemical formula
+    else if (len === 2) {
+      checks = [[chk.substr(0, 1), chk.substr(1, 1)]];
+    } else if (len === 3) {
+      checks = [
+        [chk.substr(0, 1), chk.substr(1, 1), chk.substr(2, 1)],
+        [chk.substr(0, 1), chk.substr(1, 2)],
+        [chk.substr(0, 2), chk.substr(2, 1)],
+      ];
+    } else if (len === 4) {
+      checks = [
+        [chk.substr(0, 2), chk.substr(2, 2)],
+        [chk.substr(0, 2), chk.substr(2, 1), chk.substr(3, 1)],
+        [chk.substr(0, 1), chk.substr(1, 1), chk.substr(2, 2)],
+        [chk.substr(0, 1), chk.substr(1, 2), chk.substr(3, 1)],
+        [
+          chk.substr(0, 1),
+          chk.substr(1, 1),
+          chk.substr(2, 1),
+          chk.substr(3, 1),
+        ],
+      ];
+    } else if (len === 5) {
+      checks = [
+        [chk.substr(0, 1), chk.substr(1, 1), chk.substr(2, 1)],
+        [chk.substr(0, 1), chk.substr(1, 1), chk.substr(2, 2)],
+        [chk.substr(0, 2), chk.substr(2, 2), chk.substr(4, 1)],
+        [chk.substr(0, 1), chk.substr(1, 2), chk.substr(3, 2)],
+        [chk.substr(0, 1), chk.substr(1, 2), chk.substr(3, 1)],
+        [chk.substr(0, 2), chk.substr(2, 1), chk.substr(3, 1)],
+        [chk.substr(0, 2), chk.substr(2, 1), chk.substr(3, 2)],
+      ];
+    } else {
+      // 6-9
+      checks = [
+        // NB too improbable to have 5 one-symbol elements in row
+        [chk.substr(0, 2), chk.substr(2, 2), chk.substr(4, 2)], // El-El-El
+        [
+          chk.substr(0, 2),
+          chk.substr(2, 1),
+          chk.substr(3, 1),
+          chk.substr(4, 2),
+        ], // El-E-E-El
+        [
+          chk.substr(0, 2),
+          chk.substr(2, 1),
+          chk.substr(3, 1),
+          chk.substr(4, 1),
+          chk.substr(5, 1),
+        ], // El-E-E-E-E
+        [
+          chk.substr(0, 2),
+          chk.substr(2, 2),
+          chk.substr(4, 1),
+          chk.substr(5, 1),
+        ], // El-El-E-E
+        [
+          chk.substr(0, 2),
+          chk.substr(2, 2),
+          chk.substr(4, 1),
+          chk.substr(5, 2),
+        ], // El-El-E-El
+        [
+          chk.substr(0, 2),
+          chk.substr(2, 1),
+          chk.substr(3, 2),
+          chk.substr(5, 1),
+        ], // El-E-El-E
+        [
+          chk.substr(0, 2),
+          chk.substr(2, 1),
+          chk.substr(3, 2),
+          chk.substr(5, 2),
+        ], // El-E-El-El
+        [
+          chk.substr(0, 1),
+          chk.substr(1, 2),
+          chk.substr(3, 1),
+          chk.substr(4, 2),
+        ], // E-El-E-El
+        [
+          chk.substr(0, 1),
+          chk.substr(1, 2),
+          chk.substr(3, 1),
+          chk.substr(4, 1),
+          chk.substr(5, 1),
+        ], // E-El-E-E-E
+        [
+          chk.substr(0, 1),
+          chk.substr(1, 2),
+          chk.substr(3, 2),
+          chk.substr(5, 1),
+        ], // E-El-El-E
+        [
+          chk.substr(0, 1),
+          chk.substr(1, 2),
+          chk.substr(3, 2),
+          chk.substr(5, 2),
+        ], // E-El-El-El
+        [
+          chk.substr(0, 1),
+          chk.substr(1, 1),
+          chk.substr(2, 2),
+          chk.substr(4, 1),
+          chk.substr(5, 1),
+        ], // E-E-El-E-E
+        [
+          chk.substr(0, 1),
+          chk.substr(1, 1),
+          chk.substr(2, 2),
+          chk.substr(4, 2),
+        ], // E-E-El-El
+        [
+          chk.substr(0, 1),
+          chk.substr(1, 1),
+          chk.substr(2, 1),
+          chk.substr(3, 2),
+          chk.substr(5, 1),
+        ], // E-E-E-El-E
+        [
+          chk.substr(0, 1),
+          chk.substr(1, 1),
+          chk.substr(2, 1),
+          chk.substr(3, 1),
+          chk.substr(4, 2),
+        ], // E-E-E-E-El
+      ];
+    }
+    //console.log(checks);
+
+    for (let i = 0; i < checks.length; i++) {
+      let signals = 0;
+      for (let j = 0; j < checks[i].length; j++) {
+        if (periodic_elements_cased.includes(checks[i][j])) signals++;
+
+        if (signals === checks[i].length) {
+          //console.log(checks[i]);
+          return true;
+        }
+      }
+    }
+    return false;
+  }
+
+  /*
+   * Detect facets: formulae, elements, lattices, and some classes
+   */
+  function try_uniword_facet(term) {
+    if (term === 'AB' || term === 'ABC' || term === 'ABCD') return ['formulae']; // special case-sensitive anonymous cases
+
+    term = term.toLowerCase();
+
+    const maybe_formula = !is_numeric(term.charAt(0));
+
+    const dmatches = getMatchAll(term, /(\d)/g);
+    if (dmatches && dmatches.length > 1 && maybe_formula) return ['formulae']; // no props with more than one digit
+
+    const imatches = getMatchAll(escape(term), /%u208(\d)/g);
+    if (imatches && imatches.length && maybe_formula) return ['formulae']; // no props with subscripts
+
+    if (periodic_elements.includes(term)) return ['elements', capitalize(term)];
+    else if (periodic_element_names.includes(term))
+      return [
+        'elements',
+        capitalize(periodic_elements[periodic_element_names.indexOf(term)]),
+      ];
+
+    if (
+      term.includes('-') &&
+      !term.split('-').some((part) => !periodic_elements.includes(part))
+    ) {
+      return [
+        'elements',
+        term
+          .split('-')
+          .map((el) => capitalize(el))
+          .join('-'),
+      ];
+    }
+
+    if (['element', 'elementary', 'unitary'].includes(term))
+      return ['classes', 'unary'];
+    else if (
+      term === 'quintenary' ||
+      term === 'quinternary' ||
+      term === 'quinternaries' ||
+      term === 'quinaries' ||
+      term === 'pentanary' ||
+      term === 'pentanaries'
+    )
+      return ['classes', 'quinary'];
+    else if (term === 'actinide' || term === 'actinides')
+      return ['classes', 'actinoid'];
+    else if (
+      term === 'lantanide' ||
+      term === 'lantanides' ||
+      term === 'lanthanide' ||
+      term === 'lanthanides' ||
+      term === 'lantanoid' ||
+      term === 'lantanoids'
+    )
+      return ['classes', 'lanthanoid'];
+    else if (term.endsWith('ite') && term.length > 4) return ['classes'];
+
+    // direct element-group single-word matches (tetrels, triels, chalcogen, etc.)
+    const groupEls = getGroupElements(term);
+    if (groupEls) return ['elements', groupEls.join('-')];
+
+    const chk = term
+      .replace(' structure', '')
+      .replace(' lattice', '')
+      .replace(' crystalline', '')
+      .replace(' crystal', '');
+    if (lat_fgrs.includes(chk)) return ['lattices', lat_i2p[lat_p2i[chk]]];
+
+    if (term.length <= 9 && dmatches && maybe_formula) return ['formulae']; // no SHORT props with digits (NB L0, E1)
+
+    return false;
+  }
+
+  /*
+   * Detect facets: classes, props
+   */
+  function try_multiword_facet(term, queue) {
+    term = term.toLowerCase();
+
+    let candidate = false,
+      combined = false,
+      orig = false;
+
+    if (queue.length) {
+      orig = term;
+      combined = true;
+      queue.forEach(function ({ input }) {
+        term = `${input} ${term}`;
+      });
+    }
+    //console.log("CHECKING TERM FOR MULTI-FACET: "+term);
+
+    // Special handling: phrases like 'period 2' or 'group 11'
+    const pgMatch = term.match(/^(period|group)\s+(\d{1,2})$/);
+    if (pgMatch) {
+      const kind = pgMatch[1];
+      const num = pgMatch[2];
+      const key = `${kind} ${num}`;
+      const els = getGroupElements(key);
+      if (els) return { facet: 'elements', input: els.join('-'), ready: 1 };
+    }
+
+    candidate = check_category(term, 'classes');
+    if (candidate) {
+      if (combined) candidate.combined = true;
+      return candidate;
+    }
+
+    candidate = check_category(term, 'props');
+    if (candidate) {
+      if (combined) candidate.combined = true;
+      return candidate;
+    }
+
+    let single_chk;
+    if (term.endsWith('s')) {
+      // plural-singular fixups
+      single_chk = term.substr(0, term.length - 1);
+
+      if (!combined && single_chk.endsWith('ite'))
+        return { facet: 'classes', input: single_chk, ready: 1 };
+
+      candidate = check_category(single_chk, 'classes');
+      if (candidate) {
+        if (combined) candidate.combined = true;
+        return candidate;
+      }
+    }
+    if (term.endsWith('es')) {
+      // plural-singular fixups
+      single_chk = term.substr(0, term.length - 2);
+
+      if (single_chk === 'binari') single_chk = 'binary';
+      else if (single_chk === 'ternari') single_chk = 'ternary';
+      else if (single_chk === 'quaternari') single_chk = 'quaternary';
+
+      candidate = check_category(single_chk, 'classes');
+      if (candidate) {
+        if (combined) candidate.combined = true;
+        return candidate;
+      }
+    }
+
+    if (!candidate && orig) return try_multiword_facet(orig, []);
+    return false;
+  }
+
+  /*
+   * Utility algo
+   */
+  function check_category(term, category) {
+    const host = category === 'classes' ? mpds_classes : mpds_props;
+
+    if (host.includes(term)) return { facet: category, input: term, ready: 1 };
+
+    const len = host.length;
+    const re = new RegExp(`(?:^|\\s)(${term})(?=\\s|$)`);
+    let idx;
+    for (let i = 0; i < len; i++) {
+      idx = host[i].search(re);
+      if (idx === 0) {
+        //console.log("Found unstrict match in "+category+" with <"+host[i]+">");
+        return { facet: category, input: term, anew: 1 };
+      }
+    }
+    return false;
+  }
+
+  /*
+   * Chemical formula: ABC3 vs. SrTiO3
+   */
+  function is_formula_anonymous(formula) {
+    const detect = new RegExp(
+      /A(\d{0,3})B(\d{0,3})(C(\d{0,3})(D(\d{0,3}))?)?\b/
+    );
+    return formula.charAt(0) === 'A' && detect.test(formula);
+  }
+
+  /*
+   * Get center and ligand information from a string
+   */
+  function parse_ligand(string, start) {
+    const center = string.slice(0, start).toLowerCase();
+
+    if (
+      string.slice(start, start + 1).toLowerCase() === 'x' &&
+      string.slice(start, start + 2).toLowerCase() !== 'xe'
+    ) {
+      if (string.slice(start).length === 1) return [center, 'X'];
+
+      return [center, 'X' + string.slice(start + 1)];
+    }
+
+    if (string.length === start) return [center, 'X'];
+
+    const remainder = string.slice(start);
+
+    if (is_numeric(remainder.slice(0, 1)) && start === 2)
+      return parse_ligand(string, 1);
+
+    return [center, capitalize(remainder)];
+  }
+
+  /*
+   * Get center and ligand information from a string
+   */
+  function _parse_aeatoms(string) {
+    const pos = string.indexOf('-');
+
+    if (pos !== -1) {
+      const center = string.slice(0, pos),
+        ligand = string.slice(pos + 1);
+
+      if (center.length > 2) return false;
+
+      return parse_ligand(center + ligand, center.length);
+    }
+
+    const trials = [2, 1],
+      periodic_elements_xed = ['x'].concat(periodic_elements);
+
+    for (let start = 0; start < 2; start++) {
+      if (
+        string.length >= trials[start] &&
+        periodic_elements_xed.indexOf(
+          string.slice(0, trials[start]).toLowerCase()
+        ) !== -1
+      ) {
+        return parse_ligand(string, trials[start]);
+      }
+    }
+    return false;
+  }
+
+  /**
+   * Get center and ligand information from a string
+   */
+  function parse_aeatoms(string) {
+    const parsed = _parse_aeatoms(string);
+
+    if (!parsed) return ['?', '?'];
+
+    return [capitalize(parsed[0]), formula_to_tags(parsed[1])];
+  }
+
+  /*
+   * Add HTML tags to a chemical formula as a string
+   */
+  function formula_to_tags(string) {
+    let sub = false,
+      html_formula = '';
+
+    for (let i = 0, len = string.length; i < len; i++) {
+      if (is_numeric(string[i]) || string[i] === '.') {
+        if (!sub) {
+          html_formula += '<sub>';
+          sub = true;
+        }
+      } else {
+        if (sub) {
+          html_formula += '</sub>';
+          sub = false;
+        }
+      }
+      html_formula += string[i];
+    }
+    if (sub) html_formula += '</sub>';
+    return html_formula;
+  }
+
+  /*
+   * User input processing: main algorithm
+   */
+  function guess(inputstr) {
+    // *pseudo_numerics*
+    /*if (inputstr.includes('c/a ') || inputstr.includes('a/b ') || inputstr.includes('b/c ')) {
 			// FIXME slashes in names
 			if (inputstr.includes('c/a ')) inputstr = inputstr.replace('c/a ', 'c--a ');
 			if (inputstr.includes('a/b ')) inputstr = inputstr.replace('a/b ', 'a--b ');
 			if (inputstr.includes('b/c ')) inputstr = inputstr.replace('b/c ', 'b--c ');
 		}*/
 
-		if (inputstr.includes('\"')) return {'ignored': inputstr}; // Optimade guard
-
-		const tokens = inputstr
-			.replace(new RegExp('\\+|\\!|\\?', 'g'), '')
-			.replace(new RegExp(',|/', 'g'), ' ')
-			.replace(new RegExp('<', 'g'), ' < ')
-			.replace(new RegExp('>', 'g'), ' > ')
-			.replace(new RegExp('=', 'g'), ' = ')
-			.split(/\s+/);
-
-		const result = {};
-		let n_terms = 0;
-		let n_toks = 1;
-		let queue = [];
-		const ignored = [];
-		//console.log(tokens);
-		// TODO: gracefully discard brackets
-
-		tokens.forEach(function (input) {
-			let facet = false,
-				simple = false;
-			input = input.trim();
-
-			if (
-				input.includes('<') ||
-				input.includes('>') ||
-				input.includes('=') ||
-				is_numeric(input)
-			) {
-				// numeric searches
-				if (
-					input.indexOf('<') === 0 ||
-					input.indexOf('>') === 0 ||
-					input.indexOf('=') === 0
-				) {
-					// separated op sign
-					if (result.props) {
-						if (!result.numeric) result.numeric = [];
-						result.numeric.push([result.props, input.substr(0, 1)]);
-						delete result.props;
-					} else if (result.numeric && result.numeric[result.numeric.length - 1]) {
-						result.numeric[result.numeric.length - 1][1] = input.substr(0, 1); // TODO account <=, =<, >=, =>
-					} else ignored.push(input);
-				} else if (is_numeric(input)) {
-					if (result.props) {
-						if (!result.numeric) result.numeric = [];
-						result.numeric.push([result.props, '=', parseFloat(input)]);
-						delete result.props;
-					} else if (result.numeric && result.numeric[result.numeric.length - 1]) {
-						result.numeric[result.numeric.length - 1][2] = parseFloat(input); // NB no commas!
-						if (!result.numeric[result.numeric.length - 1][1])
-							result.numeric[result.numeric.length - 1][1] = '=';
-					} else ignored.push(input);
-				}
-				return;
-			}
-
-			if (!queue.length || (input != 'at' && input != 'in')) {
-				// FIXME TODO special treatment, words vs. chemical symbols
-				simple = try_uniword_facet(input);
-			}
-
-			if (simple) {
-				facet = simple[0];
-				if (simple[1]) input = simple[1];
-				if (queue.length && !queue[queue.length - 1].ready)
-					ignored.push(...queue.map((obj) => obj.input));
-				queue = [];
-				//console.log(input + ": found simple facet " + simple[0]);
-			} else {
-				if (!stop_words.includes(input)) {
-					input = input
-						.replace(new RegExp('\\(', 'g'), '')
-						.replace(new RegExp('\\)', 'g'), '')
-						.replace(new RegExp('\\[', 'g'), '')
-						.replace(new RegExp('\\]', 'g'), '');
-
-					const candidate = try_multiword_facet(input, queue);
-					//console.log(candidate);
-
-					if (candidate.combined) queue = [];
-
-					if (candidate.ready) {
-						// term found either alone or with previous
-						facet = candidate.facet;
-						input = candidate.input;
-						queue = [candidate];
-					} else if (candidate.anew) {
-						// token anew
-						if (queue.length && !queue[queue.length - 1].ready)
-							ignored.push(...queue.map((obj) => obj.input));
-						queue = [candidate];
-					} else if (!candidate) {
-						// token unknown
-						if (queue.length && !queue[queue.length - 1].ready)
-							ignored.push(...queue.map((obj) => obj.input));
-						queue = [];
-						if (is_like_chem_formula(input)) {
-							facet = 'formulae';
-						}
-						if (!facet) {
-							ignored.push(input);
-						}
-					}
-				}
-			}
-
-			if (n_toks === tokens.length) {
-				// token at the end, terminating
-				if (queue.length && !queue[queue.length - 1].ready)
-					ignored.push(...queue.map((obj) => obj.input));
-				queue = [];
-			}
-
-			if (facet) {
-				if (facet === 'formulae') input = termify_formulae(input);
-
-				if (result[facet]) {
-					// What to do with the found term of the same category?
-					//console.log('Compare: '+result[facet]+' vs. '+input);
-
-					if (facet === 'formulae') {
-						ignored.push(input);
-					} else if (facet === 'elements') {
-						result[facet] += `-${input}`;
-					} else if (facet === 'classes') {
-						result[facet] += `, ${input}`; //escape(input);
-					} else if (facet === 'props') {
-						if (input.includes(result[facet])) {
-							//console.log('Smaller match '+result[facet]+' was thrown away');
-							result[facet] = input;
-						} else {
-							ignored.push(input);
-						}
-					}
-				} else result[facet] = input; //escape(input);
-
-				n_terms++;
-			}
-			n_toks++;
-		});
-
-		result.ignored = ignored;
-		return result;
-	}
-
-	/*
-	 * Chemical formulae parsing, courtesy of Nathan Leung, MIT license
-	 */
-	function tokenizeInitial(formula, tokens = []) {
-		if (formula.length === 0) {
-			return tokens;
-		}
-
-		// Tokenize parentheses
-		if (formula[0] === '(' || formula[0] === '[') {
-			return tokenizeInitial(formula.substr(1), [
-				...tokens, {
-					type: 'parenthesis',
-					value: 'open',
-				},
-			]);
-		}
-		if (formula[0] === ')' || formula[0] === ']') {
-			return tokenizeAfterElement(formula.substr(1), [
-				...tokens, {
-					type: 'parenthesis',
-					value: 'close',
-				},
-			]);
-		}
-
-		// Tokenize coefficient
-		let num = '';
-		for (let i = 0; i < formula.length; i++) {
-			// If the char is a number
-			if (!isNaN(Number(formula[i]))) {
-				num = num + formula[i];
-			} else {
-				// End loop if char is not a number
-				break;
-			}
-		}
-		// If a coefficient exists
-		if (num !== '') {
-			return tokenizeInitial(formula.substr(num.length), [
-				...tokens, {
-					type: 'coefficient',
-					value: Number(num),
-				},
-			]);
-		}
-
-		// Tokenize element
-		// Check if current char is uppercase letter
-		const char = formula[0];
-		const code = char.charCodeAt(0);
-		if (code >= 65 && code <= 90) {
-			// If next char is a lowercase letter
-			if (formula.length > 1) {
-				const nextChar = formula[1];
-				const nextCode = nextChar.charCodeAt(0);
-				if (nextCode >= 97 && nextCode <= 122) {
-					return tokenizeAfterElement(formula.substr(2), [
-						...tokens, {
-							type: 'element',
-							value: char + nextChar,
-						},
-					]);
-				}
-			}
-			return tokenizeAfterElement(formula.substr(1), [
-				...tokens, {
-					type: 'element',
-					value: char,
-				},
-			]);
-		}
-
-		// If it doesn't match anything
-		throw Error(`There was an error parsing formula. We were able to get to here:\n${JSON.stringify(tokens, null, 2)}
+    if (inputstr.includes('"')) return { ignored: inputstr }; // Optimade guard
+
+    const tokens = inputstr
+      .replace(new RegExp('\\+|\\!|\\?', 'g'), '')
+      .replace(new RegExp(',|/', 'g'), ' ')
+      .replace(new RegExp('<', 'g'), ' < ')
+      .replace(new RegExp('>', 'g'), ' > ')
+      .replace(new RegExp('=', 'g'), ' = ')
+      .split(/\s+/);
+
+    const result = {};
+    let n_terms = 0;
+    let n_toks = 1;
+    let queue = [];
+    const ignored = [];
+    //console.log(tokens);
+    // TODO: gracefully discard brackets
+
+    tokens.forEach(function (input) {
+      let facet = false,
+        simple = false;
+      input = input.trim();
+      const linput = input.toLowerCase();
+
+      // Allow 'period' and 'group' to be queued so they can combine with a following number
+      if (linput === 'period' || linput === 'group') {
+        // push a not-ready candidate to queue and continue
+        queue.push({ input: linput, ready: 0 });
+        n_toks++;
+        return;
+      }
+
+      if (
+        input.includes('<') ||
+        input.includes('>') ||
+        input.includes('=') ||
+        is_numeric(input)
+      ) {
+        // numeric searches
+        // If this numeric token is actually the second part of 'period N' or 'group N',
+        // don't treat it as a numeric filter but allow multiword facet combining.
+        if (
+          is_numeric(input) &&
+          queue.length &&
+          (queue[queue.length - 1].input === 'period' ||
+            queue[queue.length - 1].input === 'group')
+        ) {
+          // fall-through to normal multiword processing
+        } else if (
+          input.indexOf('<') === 0 ||
+          input.indexOf('>') === 0 ||
+          input.indexOf('=') === 0
+        ) {
+          // separated op sign
+          if (result.props) {
+            if (!result.numeric) result.numeric = [];
+            result.numeric.push([result.props, input.substr(0, 1)]);
+            delete result.props;
+          } else if (
+            result.numeric &&
+            result.numeric[result.numeric.length - 1]
+          ) {
+            result.numeric[result.numeric.length - 1][1] = input.substr(0, 1); // TODO account <=, =<, >=, =>
+          } else ignored.push(input);
+        } else if (is_numeric(input)) {
+          if (result.props) {
+            if (!result.numeric) result.numeric = [];
+            result.numeric.push([result.props, '=', parseFloat(input)]);
+            delete result.props;
+          } else if (
+            result.numeric &&
+            result.numeric[result.numeric.length - 1]
+          ) {
+            result.numeric[result.numeric.length - 1][2] = parseFloat(input); // NB no commas!
+            if (!result.numeric[result.numeric.length - 1][1])
+              result.numeric[result.numeric.length - 1][1] = '=';
+          } else ignored.push(input);
+        }
+        return;
+      }
+
+      if (!queue.length || (input != 'at' && input != 'in')) {
+        // FIXME TODO special treatment, words vs. chemical symbols
+        simple = try_uniword_facet(input);
+      }
+
+      if (simple) {
+        facet = simple[0];
+        if (simple[1]) input = simple[1];
+        if (queue.length && !queue[queue.length - 1].ready)
+          ignored.push(...queue.map((obj) => obj.input));
+        queue = [];
+        //console.log(input + ": found simple facet " + simple[0]);
+      } else {
+        if (!stop_words.includes(input)) {
+          input = input
+            .replace(new RegExp('\\(', 'g'), '')
+            .replace(new RegExp('\\)', 'g'), '')
+            .replace(new RegExp('\\[', 'g'), '')
+            .replace(new RegExp('\\]', 'g'), '');
+
+          const candidate = try_multiword_facet(input, queue);
+          //console.log(candidate);
+
+          if (candidate.combined) queue = [];
+
+          if (candidate.ready) {
+            // term found either alone or with previous
+            facet = candidate.facet;
+            input = candidate.input;
+            queue = [candidate];
+          } else if (candidate.anew) {
+            // token anew
+            if (queue.length && !queue[queue.length - 1].ready)
+              ignored.push(...queue.map((obj) => obj.input));
+            queue = [candidate];
+          } else if (!candidate) {
+            // token unknown
+            if (queue.length && !queue[queue.length - 1].ready)
+              ignored.push(...queue.map((obj) => obj.input));
+            queue = [];
+            if (is_like_chem_formula(input)) {
+              facet = 'formulae';
+            }
+            if (!facet) {
+              ignored.push(input);
+            }
+          }
+        }
+      }
+
+      if (n_toks === tokens.length) {
+        // token at the end, terminating
+        if (queue.length && !queue[queue.length - 1].ready)
+          ignored.push(...queue.map((obj) => obj.input));
+        queue = [];
+      }
+
+      if (facet) {
+        if (facet === 'formulae') input = termify_formulae(input);
+
+        if (result[facet]) {
+          // What to do with the found term of the same category?
+          //console.log('Compare: '+result[facet]+' vs. '+input);
+
+          if (facet === 'formulae') {
+            ignored.push(input);
+          } else if (facet === 'elements') {
+            result[facet] += `-${input}`;
+          } else if (facet === 'classes') {
+            result[facet] += `, ${input}`; //escape(input);
+          } else if (facet === 'props') {
+            if (input.includes(result[facet])) {
+              //console.log('Smaller match '+result[facet]+' was thrown away');
+              result[facet] = input;
+            } else {
+              ignored.push(input);
+            }
+          }
+        } else result[facet] = input; //escape(input);
+
+        n_terms++;
+      }
+      n_toks++;
+    });
+
+    result.ignored = ignored;
+    return result;
+  }
+
+  /*
+   * Chemical formulae parsing, courtesy of Nathan Leung, MIT license
+   */
+  function tokenizeInitial(formula, tokens = []) {
+    if (formula.length === 0) {
+      return tokens;
+    }
+
+    // Tokenize parentheses
+    if (formula[0] === '(' || formula[0] === '[') {
+      return tokenizeInitial(formula.substr(1), [
+        ...tokens,
+        {
+          type: 'parenthesis',
+          value: 'open',
+        },
+      ]);
+    }
+    if (formula[0] === ')' || formula[0] === ']') {
+      return tokenizeAfterElement(formula.substr(1), [
+        ...tokens,
+        {
+          type: 'parenthesis',
+          value: 'close',
+        },
+      ]);
+    }
+
+    // Tokenize coefficient
+    let num = '';
+    for (let i = 0; i < formula.length; i++) {
+      // If the char is a number
+      if (!isNaN(Number(formula[i]))) {
+        num = num + formula[i];
+      } else {
+        // End loop if char is not a number
+        break;
+      }
+    }
+    // If a coefficient exists
+    if (num !== '') {
+      return tokenizeInitial(formula.substr(num.length), [
+        ...tokens,
+        {
+          type: 'coefficient',
+          value: Number(num),
+        },
+      ]);
+    }
+
+    // Tokenize element
+    // Check if current char is uppercase letter
+    const char = formula[0];
+    const code = char.charCodeAt(0);
+    if (code >= 65 && code <= 90) {
+      // If next char is a lowercase letter
+      if (formula.length > 1) {
+        const nextChar = formula[1];
+        const nextCode = nextChar.charCodeAt(0);
+        if (nextCode >= 97 && nextCode <= 122) {
+          return tokenizeAfterElement(formula.substr(2), [
+            ...tokens,
+            {
+              type: 'element',
+              value: char + nextChar,
+            },
+          ]);
+        }
+      }
+      return tokenizeAfterElement(formula.substr(1), [
+        ...tokens,
+        {
+          type: 'element',
+          value: char,
+        },
+      ]);
+    }
+
+    // If it doesn't match anything
+    throw Error(`There was an error parsing formula. We were able to get to here:\n${JSON.stringify(
+      tokens,
+      null,
+      2
+    )}
 			\n\nRemaining formula: ${formula}`);
-	}
-
-	/*
-	 * Chemical formulae parsing, courtesy of Nathan Leung, MIT license
-	 */
-	function tokenizeAfterElement(formula, tokens) {
-		if (formula.length === 0) {
-			return tokenizeInitial(formula, tokens);
-		}
-
-		// Tokenize subscript
-		// This is repeated from tokenizeInitial, perhaps
-		// take it out
-		let num = '';
-		for (let i = 0; i < formula.length; i++) {
-			// If the char is a number
-			if (!isNaN(Number(formula[i]))) {
-				num = num + formula[i];
-			} else {
-				// End loop if char is not a number
-				break;
-			}
-		}
-		// If a subscript exists
-		if (num !== '') {
-			return tokenizeInitial(formula.substr(num.length), [
-				...tokens, {
-					type: 'subscript',
-					value: Number(num),
-				},
-			]);
-		}
-
-		// If it's not anything, then pass back to tokenizeInitial
-		return tokenizeInitial(formula, tokens);
-	}
-
-	/*
-	 * Chemical formulae parsing, courtesy of Nathan Leung, MIT license
-	 */
-	function countElements(tokens, elements = {}) {
-		if (tokens.length === 0) {
-			return elements;
-		}
-		// Create keys for each element present in the tokens array
-		if (Object.keys(elements).length === 0) {
-			for (let i = 0; i < tokens.length; i++) {
-				if (tokens[i].type === 'element') {
-					if (typeof elements[tokens[i].value] === 'undefined') {
-						elements[tokens[i].value] = 0;
-					}
-				}
-			}
-		}
-
-		for (let i = 0; i < tokens.length; i++) {
-			if (tokens[i].type === 'coefficient') {
-				return countInGroup(tokens.slice(1), elements, tokens[i].value);
-			} else if (tokens[i].type === 'element') {
-				return countInGroup(tokens, elements);
-			} else if (tokens[i].type === 'parenthesis') {
-				if (tokens[i].value === 'open') {
-					for (let j = i; j < tokens.length; j++) {
-						if (tokens[j].type === 'parenthesis' && tokens[j].value === 'close') {
-							// There will always be a subscript after a parenthetical group
-							if (
-								tokens.length > (j + 1) &&
-								tokens[j].type === 'parenthesis' &&
-								tokens[j].value === 'close'
-							) {
-								if (tokens[j + 1].type === 'subscript') {
-									return countInGroup(tokens.slice(1), elements, tokens[j + 1].value);
-								}
-							}
-							//console.error(JSON.stringify(tokens, null, 2));
-							//console.error(JSON.stringify(elements, null, 2))
-							throw Error('Parenthetical group must have subscript');
-						}
-					}
-					return countInGroup(tokens.slice(1), elements);
-				} else {
-					// Skip closing parenthesis and subscript
-					return countInGroup(tokens.slice(2), elements);
-				}
-			}
-		}
-
-		// If none of the above cases are covered then throw an error
-		throw Error(`There was an error:\n\n${JSON.stringify(tokens, null, 2)}
+  }
+
+  /*
+   * Chemical formulae parsing, courtesy of Nathan Leung, MIT license
+   */
+  function tokenizeAfterElement(formula, tokens) {
+    if (formula.length === 0) {
+      return tokenizeInitial(formula, tokens);
+    }
+
+    // Tokenize subscript
+    // This is repeated from tokenizeInitial, perhaps
+    // take it out
+    let num = '';
+    for (let i = 0; i < formula.length; i++) {
+      // If the char is a number
+      if (!isNaN(Number(formula[i]))) {
+        num = num + formula[i];
+      } else {
+        // End loop if char is not a number
+        break;
+      }
+    }
+    // If a subscript exists
+    if (num !== '') {
+      return tokenizeInitial(formula.substr(num.length), [
+        ...tokens,
+        {
+          type: 'subscript',
+          value: Number(num),
+        },
+      ]);
+    }
+
+    // If it's not anything, then pass back to tokenizeInitial
+    return tokenizeInitial(formula, tokens);
+  }
+
+  /*
+   * Chemical formulae parsing, courtesy of Nathan Leung, MIT license
+   */
+  function countElements(tokens, elements = {}) {
+    if (tokens.length === 0) {
+      return elements;
+    }
+    // Create keys for each element present in the tokens array
+    if (Object.keys(elements).length === 0) {
+      for (let i = 0; i < tokens.length; i++) {
+        if (tokens[i].type === 'element') {
+          if (typeof elements[tokens[i].value] === 'undefined') {
+            elements[tokens[i].value] = 0;
+          }
+        }
+      }
+    }
+
+    for (let i = 0; i < tokens.length; i++) {
+      if (tokens[i].type === 'coefficient') {
+        return countInGroup(tokens.slice(1), elements, tokens[i].value);
+      } else if (tokens[i].type === 'element') {
+        return countInGroup(tokens, elements);
+      } else if (tokens[i].type === 'parenthesis') {
+        if (tokens[i].value === 'open') {
+          for (let j = i; j < tokens.length; j++) {
+            if (
+              tokens[j].type === 'parenthesis' &&
+              tokens[j].value === 'close'
+            ) {
+              // There will always be a subscript after a parenthetical group
+              if (
+                tokens.length > j + 1 &&
+                tokens[j].type === 'parenthesis' &&
+                tokens[j].value === 'close'
+              ) {
+                if (tokens[j + 1].type === 'subscript') {
+                  return countInGroup(
+                    tokens.slice(1),
+                    elements,
+                    tokens[j + 1].value
+                  );
+                }
+              }
+              //console.error(JSON.stringify(tokens, null, 2));
+              //console.error(JSON.stringify(elements, null, 2))
+              throw Error('Parenthetical group must have subscript');
+            }
+          }
+          return countInGroup(tokens.slice(1), elements);
+        } else {
+          // Skip closing parenthesis and subscript
+          return countInGroup(tokens.slice(2), elements);
+        }
+      }
+    }
+
+    // If none of the above cases are covered then throw an error
+    throw Error(`There was an error:\n\n${JSON.stringify(tokens, null, 2)}
 			\n\n${JSON.stringify(elements, null, 2)}`);
-	}
-
-	/*
-	 * Chemical formulae parsing, courtesy of Nathan Leung, MIT license
-	 */
-	function countInGroup(tokens, elements, coefficient = 1) {
-		if (tokens.length === 0) {
-			return countElements(tokens, elements);
-		}
-
-		if (tokens[0].type !== 'element') {
-			// If this happens this is likely a mistake, pass back to countElements
-			return countElements(tokens, elements);
-		}
-
-		let i = 0;
-		while (i < tokens.length) {
-			if (tokens[i].type === 'element') {
-				if (tokens.length > (i + 1) && tokens[i + 1].type === 'subscript') {
-					elements[tokens[i].value] += (tokens[i + 1].value * coefficient);
-					i += 2;
-					continue;
-				}
-				elements[tokens[i].value] += coefficient;
-				i++;
-				continue;
-			}
-			break;
-		}
-		return countElements(tokens.slice(i), elements);
-	}
-
-	/*
-	 * Sorting the formulae alphabetically for Optimade
-	 */
-	function sort_formula(text) {
-		let parsed = false;
-		try {
-			parsed = countElements(tokenizeInitial(text));
-		} catch (err) {
-			//console.error(err);
-			return text;
-		}
-		let sorted = '',
-			els = Object.keys(parsed);
-		els.sort();
-		els.forEach(function (el) {
-			sorted += el + (parsed[el] === 1 ? '' : parsed[el]);
-		});
-		return sorted;
-	}
-
-	/*
-	 * Convert MPDS search query object notation into the Optimade filter
-	 */
-	function to_optimade(parsed) {
-		const filter = [];
-
-		['formulae', 'elements', 'props', 'classes'].forEach(function (categ) {
-			if (!parsed[categ]) return;
-			else if (categ === 'formulae') {
-				if (is_formula_anonymous(parsed[categ]))
-					filter.push(`chemical_formula_anonymous="${parsed[categ]}"`);
-				else filter.push(`chemical_formula_reduced="${sort_formula(parsed[categ])}"`);
-			} else if (categ === 'elements') {
-				filter.push(`elements HAS ALL "${parsed[categ].split('-').join('","')}"`);
-			} else if (categ === 'props') {
-				filter.push(`_mpds_${parsed[categ].replaceAll(' ', '_')} IS KNOWN`);
-			} else if (categ === 'classes') {
-				parsed[categ].split(', ').forEach(function (item) {
-					const arity = arity_keys.indexOf(item);
-					if (arity > 0) filter.push(`nelements=${arity}`);
-				});
-			}
-		});
-
-		return filter.join(' AND ');
-	}
-
-	/*
-	 * API
-	 */
-	return {
-		guess,
-		to_optimade,
-
-		is_formula_anonymous,
-		sort_formula,
-		parse_aeatoms,
-		termify_formulae,
-		is_numeric,
-
-		arity_keys,
-		periodic_elements,
-		periodic_elements_cased,
-		periodic_element_names
-	};
+  }
+
+  /*
+   * Chemical formulae parsing, courtesy of Nathan Leung, MIT license
+   */
+  function countInGroup(tokens, elements, coefficient = 1) {
+    if (tokens.length === 0) {
+      return countElements(tokens, elements);
+    }
+
+    if (tokens[0].type !== 'element') {
+      // If this happens this is likely a mistake, pass back to countElements
+      return countElements(tokens, elements);
+    }
+
+    let i = 0;
+    while (i < tokens.length) {
+      if (tokens[i].type === 'element') {
+        if (tokens.length > i + 1 && tokens[i + 1].type === 'subscript') {
+          elements[tokens[i].value] += tokens[i + 1].value * coefficient;
+          i += 2;
+          continue;
+        }
+        elements[tokens[i].value] += coefficient;
+        i++;
+        continue;
+      }
+      break;
+    }
+    return countElements(tokens.slice(i), elements);
+  }
+
+  /*
+   * Sorting the formulae alphabetically for Optimade
+   */
+  function sort_formula(text) {
+    let parsed = false;
+    try {
+      parsed = countElements(tokenizeInitial(text));
+    } catch (err) {
+      //console.error(err);
+      return text;
+    }
+    let sorted = '',
+      els = Object.keys(parsed);
+    els.sort();
+    els.forEach(function (el) {
+      sorted += el + (parsed[el] === 1 ? '' : parsed[el]);
+    });
+    return sorted;
+  }
+
+  /*
+   * Convert MPDS search query object notation into the Optimade filter
+   */
+  function to_optimade(parsed) {
+    const filter = [];
+
+    ['formulae', 'elements', 'props', 'classes'].forEach(function (categ) {
+      if (!parsed[categ]) return;
+      else if (categ === 'formulae') {
+        if (is_formula_anonymous(parsed[categ]))
+          filter.push(`chemical_formula_anonymous="${parsed[categ]}"`);
+        else
+          filter.push(
+            `chemical_formula_reduced="${sort_formula(parsed[categ])}"`
+          );
+      } else if (categ === 'elements') {
+        filter.push(
+          `elements HAS ALL "${parsed[categ].split('-').join('","')}"`
+        );
+      } else if (categ === 'props') {
+        filter.push(`_mpds_${parsed[categ].replaceAll(' ', '_')} IS KNOWN`);
+      } else if (categ === 'classes') {
+        parsed[categ].split(', ').forEach(function (item) {
+          const arity = arity_keys.indexOf(item);
+          if (arity > 0) {
+            filter.push(`nelements=${arity}`);
+          } else if (mpds_classes.includes(item)) {
+            filter.push(`_mpds_classes HAS ALL "${item}"`);
+          }
+        });
+      }
+    });
+
+    if (parsed.numeric && Array.isArray(parsed.numeric)) {
+      parsed.numeric.forEach(function (numfilt) {
+        // numfilt: [property, operator, value]
+        if (
+          numfilt.length === 3 &&
+          numfilt[0] &&
+          numfilt[1] &&
+          typeof numfilt[2] !== 'undefined'
+        ) {
+          // Remove _mpds_ prefix if present
+          let prop = numfilt[0].replace(/^_mpds_/, '').replace(/ /g, '_');
+          filter.push(`${prop}${numfilt[1]}${numfilt[2]}`);
+        }
+      });
+    }
+
+    return filter.join(' AND ');
+  }
+
+  /*
+   * API
+   */
+  return {
+    guess,
+    to_optimade,
+
+    is_formula_anonymous,
+    sort_formula,
+    parse_aeatoms,
+    termify_formulae,
+    is_numeric,
+
+    arity_keys,
+    periodic_elements,
+    periodic_elements_cased,
+    periodic_element_names,
+  };
 };
 
 if (typeof module !== 'undefined' && module.exports) {
-	module.exports = OptimadeNLP;
-} else if (typeof require === 'function' && typeof require.specified === 'function') {
-	define(function () {
-		return OptimadeNLP;
-	});
+  module.exports = OptimadeNLP;
+} else if (
+  typeof require === 'function' &&
+  typeof require.specified === 'function'
+) {
+  define(function () {
+    return OptimadeNLP;
+  });
 } else if (window !== undefined) {
-	window.OptimadeNLP = OptimadeNLP
+  window.OptimadeNLP = OptimadeNLP;
 }
diff --git a/test_nlp.json b/test_nlp.json
index 12844ac..bd61327 100755
--- a/test_nlp.json
+++ b/test_nlp.json
@@ -1,216 +1,330 @@
 [
-	["tio2", { "formulae": "tio2" }],
-	["al2o3", { "formulae": "al2o3" }],
-	["band gap, chlorides", { "props": "band gap", "classes": "chloride" }],
-	["ti o", { "elements": "Ti-O" }],
-	["tio2, band gap", { "formulae": "tio2", "props": "band gap" }],
-	["organic pb i", { "classes": "organic", "elements": "Pb-I" }],
-	["organic pb-i perovskite", { "elements": "Pb-I", "classes": "organic, perovskite" }],
-	["ternary oxide", { "classes": "ternary, oxide" }],
-	["zeolite ite", { "classes": "zeolite" }],
-	["CuLaMnSeO", { "formulae": "CuLaMnSeO" }],
-	["AgGaTeSe", { "formulae": "AgGaTeSe" }],
-	["MnOCaSeRu", { "formulae": "MnOCaSeRu" }],
-	["LaZnOAsAm", { "formulae": "LaZnOAsAm" }],
-	["MnCaOSeRu", { "formulae": "MnCaOSeRu" }],
-	["MnCaSeBRu", { "formulae": "MnCaSeBRu" }],
-	["MnCaSeRu", { "formulae": "MnCaSeRu" }],
-	["BCNOAc", { "formulae": "BCNOAc" }],
-	["LaZnOAs", { "formulae": "LaZnOAs" }],
-	["LaZnAsO", { "formulae": "LaZnAsO" }],
-	["LaOZnAs", { "formulae": "LaOZnAs" }],
-	["OLaZnAs", { "formulae": "OLaZnAs" }],
-	["OLaZnY", { "formulae": "OLaZnY" }],
-	["OLaZnYB", { "formulae": "OLaZnYB" }],
-	["OLaZnYBr", { "formulae": "OLaZnYBr" }],
-	["OLaSZnAs", { "formulae": "OLaSZnAs" }],
-	["OLaOZnAs", { "formulae": "OLaOZnAs" }],
-	["OLaZnPAs", { "formulae": "OLaZnPAs" }],
-	["OLaKZnPAs", { "formulae": "OLaKZnPAs" }],
-	["SYPMnBr", { "formulae": "SYPMnBr" }],
-	["SYMnPBr", { "formulae": "SYMnPBr" }],
-	["RbMnTaO", { "formulae": "RbMnTaO" }],
-	["RoMnTaO", {}],
-	["CPZnOY", { "formulae": "CPZnOY" }],
-	["CPOZnY", { "formulae": "CPOZnY" }],
-	["BaLaMnRuO6", { "formulae": "BaLaMnRuO6" }],
-	["phase diagrams", { "props": "phase diagram" }],
-	[
-		"thermodynamics superconductor fe",
-		{ "props": "thermodynamics", "classes": "superconductor", "elements": "Fe" }
-	],
-	["cub", { "lattices": "cubic" }],
-	[
-		"tetragonal srtio3 ternary",
-		{ "lattices": "tetragonal", "formulae": "srtio3", "classes": "ternary" }
-	],
-	[
-		"SrTiO&#x2083; tet phonons",
-		{ "lattices": "tetragonal", "formulae": "SrTiO3", "props": "phonons" }
-	],
-	["O3Al2 elastic properties", { "formulae": "O3Al2" }],
-	["optical properties, LiKSO4", { "props": "optical properties", "formulae": "LiKSO4" }],
-	["band gap ZnO hex", { "props": "band gap", "lattices": "hexagonal", "formulae": "ZnO" }],
-	["geo", {}],
-	["GeO", { "formulae": "GeO" }],
-	["beo", {}],
-	["BeO", { "formulae": "BeO" }],
-	["GdNiIn conductivity", { "formulae": "GdNiIn", "props": "conductivity" }],
-	["indium binary oxide metal", { "elements": "In", "classes": "binary, oxide, metal" }],
-	["Pd hydride lattice", { "elements": "Pd", "classes": "hydride", "props": "lattice" }],
-	["Mn halogens Be", { "elements": "Mn-Be", "classes": "halogen" }],
-	[
-		"carbide semiconductor electronic properties",
-		{ "classes": "carbide, semiconductor", "props": "electronic properties" }
-	],
-	[
-		"hexagonal carbide semiconductor electronic properties",
-		{
-			"lattices": "hexagonal",
-			"classes": "carbide, semiconductor",
-			"props": "electronic properties"
-		}
-	],
-	[
-		"rare earth chalcogens organics magnetism",
-		{ "classes": "rare earth, chalcogen, organic", "props": "magnetism" }
-	],
-	["Te metals transitional Sn", { "elements": "Te-Sn", "classes": "metal, transitional" }],
-	[
-		"radioactive nonmetals conductivity",
-		{ "classes": "radioactive, nonmetal", "props": "conductivity" }
-	],
-	["noble gases superconductivity", { "classes": "noble gas", "props": "superconductivity" }],
-	[
-		"W-Mo ternary, phase diagram",
-		{ "elements": "W-Mo", "classes": "ternary", "props": "phase diagram" }
-	],
-	[
-		"quaternary perovskites, thermodynamics",
-		{ "classes": "quaternary, perovskite", "props": "thermodynamics" }
-	],
-	["lanthanoid element", { "classes": "lanthanoid, unary" }],
-	[
-		"lanthanoid transitional electronic properties",
-		{ "classes": "lanthanoid, transitional", "props": "electronic properties" }
-	],
-	[
-		"lattice of intermetallic binary cubic",
-		{ "props": "lattice", "classes": "intermetallic, binary", "lattices": "cubic" }
-	],
-	["oxygen", { "elements": "O" }],
-	[
-		"Y oxide lattice ternary",
-		{ "elements": "Y", "classes": "oxide, ternary", "props": "lattice" }
-	],
-	["phases for ternary ferromagnet", { "props": "phases", "classes": "ternary, ferromagnet" }],
-	[
-		"superconductivity superconductors C",
-		{ "props": "superconductivity", "classes": "superconductor", "elements": "C" }
-	],
-	[
-		"magnetism, ferromagnet, chalcogen, monocl",
-		{ "props": "magnetism", "classes": "ferromagnet, chalcogen", "lattices": "monoclinic" }
-	],
-	["Rg", { "elements": "Rg" }],
-	["ErSi", { "formulae": "ErSi" }],
-	["ErSI", { "formulae": "ErSI" }],
-	["Ca3(PO4)2", { "formulae": "Ca3(PO4)2" }],
-	["Ca3[PO4]2", { "formulae": "Ca3[PO4]2" }],
-	["conductors", { "classes": "conductor" }],
-	["chromium, fe, tricl", { "elements": "Cr-Fe", "lattices": "triclinic" }],
-	["astatine, cell", { "elements": "At" }],
-	["Gruneisen coefficient", { "props": "gruneisen coefficients" }],
-	["Seebeck coefficient", { "props": "seebeck coefficient" }],
-	["figure of merit", { "props": "figure of merit" }],
-	["conductivity", { "props": "conductivity" }],
-	[
-		"crystalline structure of binaries",
-		{ "props": "crystalline structure", "classes": "binary" }
-	],
-	["crystal cell for binary compound", { "props": "crystal cell", "classes": "binary" }],
-	["valence", { "props": "valence" }],
-	["conductivity", { "props": "conductivity" }],
-	["La magnetic phase diagram", { "elements": "La", "props": "magnetic phase diagram" }],
-	["info on phase diagrams plots", { "props": "phase diagram plots" }],
-	["optical conductivity", { "props": "optical conductivity" }],
-	["cell and atoms", { "classes": "cell and atoms" }],
-	["conductor superconductor", { "classes": "conductor, superconductor" }],
-	["metal nonmetal", { "classes": "metal, nonmetal" }],
-	["metal or nonmetal character", { "props": "metal or nonmetal character" }],
-	["electrical resistance values", { "props": "electrical properties" }],
-	["temperature derivative by pressure", { "props": "pressure" }],
-	["temperature derivative by composition", {}],
-	["optical phonon frequency", { "props": "optical properties" }],
-	["enthalpy conductivity", { "props": "values of enthalpy" }],
-	["decomposis of the magneto", {}],
-	["TiO2 SrTiO3", { "formulae": "TiO2" }],
-	["average number of 4f electrons", {}],
-	["225aaa!", {}],
-	["he-he-he", { "elements": "He-He-He" }],
-	[
-		"electronic contribution to thermal conductivity",
-		{ "props": "electronic contribution to thermal conductivity" }
-	],
-	[
-		"electronic contribution to superconducting heat capacity",
-		{ "props": "electronic contribution to superconducting heat capacity" }
-	],
-	["residual resistivity", { "props": "residual resistivity ratio" }],
-	["optical absorption no", { "props": "optical absorption coefficient", "elements": "No" }],
-	["energy of optical phonon", { "props": "energy of optical phonon" }],
-	["luminescence", { "props": "values of luminescence" }],
-	["values of luminescence", { "props": "values of luminescence" }],
-	["magnetization", { "props": "magnetization" }],
-	["values of magnetization", { "props": "values of magnetization" }],
-	["energy of longitudinal optical phonon", { "props": "energy of longitudinal optical phonon" }],
-	[
-		"permittivity (dielectric constant) of perovskites ",
-		{ "props": "permittivity", "classes": "perovskite" }
-	],
-	["density > 1", { "numeric": [["density", ">", 1]], "phased": true }],
-	["InFO", { "formulae": "InFO" }],
-	["info", {}],
-	["Goodbye.", {}],
-	["iiii", {}],
-	["...", {}],
-	["What'd I do?", { "elements": "I" }],
-	["OK, let's go", { "formulae": "OK" }],
-	["Ginny Danburry's here.  Look for her", {}],
-	["I think u make me laughing.", { "elements": "I-U" }],
-	["physical properties", { "props": "physical properties" }],
-	["physical property", { "props": "physical properties" }],
-	["electric property", { "props": "electric properties" }],
-	["ab initio", {}],
-	["ab initio calculations,", { "classes": "ab initio calculations" }],
-	["ab initio literature,", { "classes": "ab initio literature" }],
-	[
-		"ab initio literature, ab initio calculations,",
-		{ "classes": "ab initio literature, ab initio calculations" }
-	],
-	["ab initio calculations P K", { "elements": "P-K", "classes": "ab initio calculations" }],
-	["ab initio literature P K", { "elements": "P-K", "classes": "ab initio literature" }],
-	[
-		"ab initio, ab initio calculations, P K",
-		{ "elements": "P-K", "classes": "ab initio calculations" }
-	],
-	[
-		"ab initio literature, ab initio calculations, P K",
-		{ "elements": "P-K", "classes": "ab initio literature, ab initio calculations" }
-	],
-	["A1B2C3", { "formulae": "A1B2C3" }],
-	["AB", { "formulae": "AB" }],
-	["ab datum", {}],
-	["AB ab initio calculations", { "formulae": "AB", "classes": "ab initio calculations" }],
-	["ABC7D9", { "formulae": "ABC7D9" }],
-	["ABCD", { "formulae": "ABCD" }],
-	["ABCDE", {}],
-	["filter=author=\"Sąžininga Žąsis\"", {}],
-	["(elements HAS \"Ac\" AND nelements=1) OR (elements HAS \"Ac\" AND nsites=1)", {}],
-	["nelements=1", {}],
-	["elements HAS \"Ti\"", {}],
-	["elements HAS ALL \"C\",\"N\",\"O\",\"H\"", {}],
-	["elements HAS \"Ti\" AND nelements>3", {}],
-	["chemical_formula_reduced=\"Li7Sn2\"", {}],
-	["chemical_formula_anonymous=\"ABC\"", {}]
+  ["tio2", { "formulae": "tio2" }],
+  ["al2o3", { "formulae": "al2o3" }],
+  ["band gap, chlorides", { "props": "band gap", "classes": "chloride" }],
+  ["ti o", { "elements": "Ti-O" }],
+  ["tio2, band gap", { "formulae": "tio2", "props": "band gap" }],
+  ["organic pb i", { "classes": "organic", "elements": "Pb-I" }],
+  [
+    "organic pb-i perovskite",
+    { "elements": "Pb-I", "classes": "organic, perovskite" }
+  ],
+  ["ternary oxide", { "classes": "ternary, oxide" }],
+  ["zeolite ite", { "classes": "zeolite" }],
+  ["CuLaMnSeO", { "formulae": "CuLaMnSeO" }],
+  ["AgGaTeSe", { "formulae": "AgGaTeSe" }],
+  ["MnOCaSeRu", { "formulae": "MnOCaSeRu" }],
+  ["LaZnOAsAm", { "formulae": "LaZnOAsAm" }],
+  ["MnCaOSeRu", { "formulae": "MnCaOSeRu" }],
+  ["MnCaSeBRu", { "formulae": "MnCaSeBRu" }],
+  ["MnCaSeRu", { "formulae": "MnCaSeRu" }],
+  ["BCNOAc", { "formulae": "BCNOAc" }],
+  ["LaZnOAs", { "formulae": "LaZnOAs" }],
+  ["LaZnAsO", { "formulae": "LaZnAsO" }],
+  ["LaOZnAs", { "formulae": "LaOZnAs" }],
+  ["OLaZnAs", { "formulae": "OLaZnAs" }],
+  ["OLaZnY", { "formulae": "OLaZnY" }],
+  ["OLaZnYB", { "formulae": "OLaZnYB" }],
+  ["OLaZnYBr", { "formulae": "OLaZnYBr" }],
+  ["OLaSZnAs", { "formulae": "OLaSZnAs" }],
+  ["OLaOZnAs", { "formulae": "OLaOZnAs" }],
+  ["OLaZnPAs", { "formulae": "OLaZnPAs" }],
+  ["OLaKZnPAs", { "formulae": "OLaKZnPAs" }],
+  ["SYPMnBr", { "formulae": "SYPMnBr" }],
+  ["SYMnPBr", { "formulae": "SYMnPBr" }],
+  ["RbMnTaO", { "formulae": "RbMnTaO" }],
+  ["RoMnTaO", {}],
+  ["CPZnOY", { "formulae": "CPZnOY" }],
+  ["CPOZnY", { "formulae": "CPOZnY" }],
+  ["BaLaMnRuO6", { "formulae": "BaLaMnRuO6" }],
+  ["phase diagrams", { "props": "phase diagram" }],
+  [
+    "thermodynamics superconductor fe",
+    { "props": "thermodynamics", "classes": "superconductor", "elements": "Fe" }
+  ],
+  ["cub", { "lattices": "cubic" }],
+  [
+    "tetragonal srtio3 ternary",
+    { "lattices": "tetragonal", "formulae": "srtio3", "classes": "ternary" }
+  ],
+  [
+    "SrTiO&#x2083; tet phonons",
+    { "lattices": "tetragonal", "formulae": "SrTiO3", "props": "phonons" }
+  ],
+  ["O3Al2 elastic properties", { "formulae": "O3Al2" }],
+  [
+    "optical properties, LiKSO4",
+    { "props": "optical properties", "formulae": "LiKSO4" }
+  ],
+  [
+    "band gap ZnO hex",
+    { "props": "band gap", "lattices": "hexagonal", "formulae": "ZnO" }
+  ],
+  ["geo", {}],
+  ["GeO", { "formulae": "GeO" }],
+  ["beo", {}],
+  ["BeO", { "formulae": "BeO" }],
+  ["GdNiIn conductivity", { "formulae": "GdNiIn", "props": "conductivity" }],
+  [
+    "indium binary oxide metal",
+    { "elements": "In", "classes": "binary, oxide, metal" }
+  ],
+  [
+    "Pd hydride lattice",
+    { "elements": "Pd", "classes": "hydride", "props": "lattice" }
+  ],
+  ["Mn halogens Be", { "elements": "Mn-Be", "classes": "halogen" }],
+  [
+    "carbide semiconductor electronic properties",
+    { "classes": "carbide, semiconductor", "props": "electronic properties" }
+  ],
+  [
+    "hexagonal carbide semiconductor electronic properties",
+    {
+      "lattices": "hexagonal",
+      "classes": "carbide, semiconductor",
+      "props": "electronic properties"
+    }
+  ],
+  [
+    "rare earth chalcogens organics magnetism",
+    { "classes": "rare earth, chalcogen, organic", "props": "magnetism" }
+  ],
+  [
+    "Te metals transitional Sn",
+    { "elements": "Te-Sn", "classes": "metal, transitional" }
+  ],
+  [
+    "radioactive nonmetals conductivity",
+    { "classes": "radioactive, nonmetal", "props": "conductivity" }
+  ],
+  [
+    "noble gases superconductivity",
+    { "classes": "noble gas", "props": "superconductivity" }
+  ],
+  [
+    "W-Mo ternary, phase diagram",
+    { "elements": "W-Mo", "classes": "ternary", "props": "phase diagram" }
+  ],
+  [
+    "quaternary perovskites, thermodynamics",
+    { "classes": "quaternary, perovskite", "props": "thermodynamics" }
+  ],
+  ["lanthanoid element", { "classes": "lanthanoid, unary" }],
+  [
+    "lanthanoid transitional electronic properties",
+    { "classes": "lanthanoid, transitional", "props": "electronic properties" }
+  ],
+  [
+    "lattice of intermetallic binary cubic",
+    {
+      "props": "lattice",
+      "classes": "intermetallic, binary",
+      "lattices": "cubic"
+    }
+  ],
+  ["oxygen", { "elements": "O" }],
+  [
+    "Y oxide lattice ternary",
+    { "elements": "Y", "classes": "oxide, ternary", "props": "lattice" }
+  ],
+  [
+    "phases for ternary ferromagnet",
+    { "props": "phases", "classes": "ternary, ferromagnet" }
+  ],
+  [
+    "superconductivity superconductors C",
+    {
+      "props": "superconductivity",
+      "classes": "superconductor",
+      "elements": "C"
+    }
+  ],
+  [
+    "magnetism, ferromagnet, chalcogen, monocl",
+    {
+      "props": "magnetism",
+      "classes": "ferromagnet, chalcogen",
+      "lattices": "monoclinic"
+    }
+  ],
+  ["Rg", { "elements": "Rg" }],
+  ["ErSi", { "formulae": "ErSi" }],
+  ["ErSI", { "formulae": "ErSI" }],
+  ["Ca3(PO4)2", { "formulae": "Ca3(PO4)2" }],
+  ["Ca3[PO4]2", { "formulae": "Ca3[PO4]2" }],
+  ["conductors", { "classes": "conductor" }],
+  ["chromium, fe, tricl", { "elements": "Cr-Fe", "lattices": "triclinic" }],
+  ["astatine, cell", { "elements": "At" }],
+  ["Gruneisen coefficient", { "props": "gruneisen coefficients" }],
+  ["Seebeck coefficient", { "props": "seebeck coefficient" }],
+  ["figure of merit", { "props": "figure of merit" }],
+  ["conductivity", { "props": "conductivity" }],
+  [
+    "crystalline structure of binaries",
+    { "props": "crystalline structure", "classes": "binary" }
+  ],
+  [
+    "crystal cell for binary compound",
+    { "props": "crystal cell", "classes": "binary" }
+  ],
+  ["valence", { "props": "valence" }],
+  ["conductivity", { "props": "conductivity" }],
+  [
+    "La magnetic phase diagram",
+    { "elements": "La", "props": "magnetic phase diagram" }
+  ],
+  ["info on phase diagrams plots", { "props": "phase diagram plots" }],
+  ["optical conductivity", { "props": "optical conductivity" }],
+  ["cell and atoms", { "classes": "cell and atoms" }],
+  ["conductor superconductor", { "classes": "conductor, superconductor" }],
+  ["metal nonmetal", { "classes": "metal, nonmetal" }],
+  ["metal or nonmetal character", { "props": "metal or nonmetal character" }],
+  ["electrical resistance values", { "props": "electrical properties" }],
+  ["temperature derivative by pressure", { "props": "pressure" }],
+  ["temperature derivative by composition", {}],
+  ["optical phonon frequency", { "props": "optical properties" }],
+  ["enthalpy conductivity", { "props": "values of enthalpy" }],
+  ["decomposis of the magneto", {}],
+  ["TiO2 SrTiO3", { "formulae": "TiO2" }],
+  ["average number of 4f electrons", {}],
+  ["225aaa!", {}],
+  ["he-he-he", { "elements": "He-He-He" }],
+  [
+    "electronic contribution to thermal conductivity",
+    { "props": "electronic contribution to thermal conductivity" }
+  ],
+  [
+    "electronic contribution to superconducting heat capacity",
+    { "props": "electronic contribution to superconducting heat capacity" }
+  ],
+  ["residual resistivity", { "props": "residual resistivity ratio" }],
+  [
+    "optical absorption no",
+    { "props": "optical absorption coefficient", "elements": "No" }
+  ],
+  ["energy of optical phonon", { "props": "energy of optical phonon" }],
+  ["luminescence", { "props": "values of luminescence" }],
+  ["values of luminescence", { "props": "values of luminescence" }],
+  ["magnetization", { "props": "magnetization" }],
+  ["values of magnetization", { "props": "values of magnetization" }],
+  [
+    "energy of longitudinal optical phonon",
+    { "props": "energy of longitudinal optical phonon" }
+  ],
+  [
+    "permittivity (dielectric constant) of perovskites ",
+    { "props": "permittivity", "classes": "perovskite" }
+  ],
+  ["density > 1", { "numeric": [["density", ">", 1]], "phased": true }],
+  ["InFO", { "formulae": "InFO" }],
+  ["info", {}],
+  ["Goodbye.", {}],
+  ["iiii", {}],
+  ["...", {}],
+  ["What'd I do?", { "elements": "I" }],
+  ["OK, let's go", { "formulae": "OK" }],
+  ["Ginny Danburry's here.  Look for her", {}],
+  ["I think u make me laughing.", { "elements": "I-U" }],
+  ["physical properties", { "props": "physical properties" }],
+  ["physical property", { "props": "physical properties" }],
+  ["electric property", { "props": "electric properties" }],
+  ["ab initio", {}],
+  ["ab initio calculations,", { "classes": "ab initio calculations" }],
+  ["ab initio literature,", { "classes": "ab initio literature" }],
+  [
+    "ab initio literature, ab initio calculations,",
+    { "classes": "ab initio literature, ab initio calculations" }
+  ],
+  [
+    "ab initio calculations P K",
+    { "elements": "P-K", "classes": "ab initio calculations" }
+  ],
+  [
+    "ab initio literature P K",
+    { "elements": "P-K", "classes": "ab initio literature" }
+  ],
+  [
+    "ab initio, ab initio calculations, P K",
+    { "elements": "P-K", "classes": "ab initio calculations" }
+  ],
+  [
+    "ab initio literature, ab initio calculations, P K",
+    {
+      "elements": "P-K",
+      "classes": "ab initio literature, ab initio calculations"
+    }
+  ],
+  ["A1B2C3", { "formulae": "A1B2C3" }],
+  ["AB", { "formulae": "AB" }],
+  ["ab datum", {}],
+  [
+    "AB ab initio calculations",
+    { "formulae": "AB", "classes": "ab initio calculations" }
+  ],
+  ["ABC7D9", { "formulae": "ABC7D9" }],
+  ["ABCD", { "formulae": "ABCD" }],
+  ["ABCDE", {}],
+  ["filter=author=\"Sąžininga Žąsis\"", {}],
+  [
+    "(elements HAS \"Ac\" AND nelements=1) OR (elements HAS \"Ac\" AND nsites=1)",
+    {}
+  ],
+  ["nelements=1", {}],
+  ["elements HAS \"Ti\"", {}],
+  ["elements HAS ALL \"C\",\"N\",\"O\",\"H\"", {}],
+  ["elements HAS \"Ti\" AND nelements>3", {}],
+  ["chemical_formula_reduced=\"Li7Sn2\"", {}],
+  ["chemical_formula_anonymous=\"ABC\"", {}],
+  ["elements HAS \"C\" AND elements HAS \"N\" AND elements HAS \"O\"", {}],
+  ["elements HAS ANY \"C\",\"N\",\"O\"", {}],
+  ["elements HAS \"C\" OR elements HAS \"N\" OR elements HAS \"O\"", {}],
+  ["elements HAS \"C\" AND (elements HAS \"N\" OR elements HAS \"O\")", {}],
+  ["elements HAS ALL \"C\",\"N\",\"O\" AND nelements=3", {}],
+  ["elements HAS ALL \"C\",\"N\",\"O\" AND nelements=4", {}],
+  ["elements HAS ANY \"C\",\"N\",\"O\" AND nelements=3", {}],
+  ["elements HAS ANY \"C\",\"N\",\"O\" AND nelements=4", {}],
+  ["nelements>3 AND elements HAS ANY \"C\",\"N\",\"O\"", {}],
+  ["nelements>4 AND elements HAS ANY \"C\",\"N\",\"O\"", {}],
+  ["nelements>3 AND elements HAS ALL \"C\",\"N\",\"O\"", {}],
+  ["nelements>4 AND elements HAS ALL \"C\",\"N\",\"O\"", {}],
+  ["band gap>1.5", { "numeric": [["band gap", ">", 1.5]], "phased": true }],
+  ["band gap<1.5", { "numeric": [["band gap", "<", 1.5]], "phased": true }],
+  ["band gap>=1.5", { "numeric": [["band gap", ">=", 1.5]], "phased": true }],
+  ["band gap<=1.5", { "numeric": [["band gap", "<=", 1.5]], "phased": true }],
+  ["band gap=1.5", { "numeric": [["band gap", "=", 1.5]], "phased": true }],
+  ["band gap!=1.5", { "numeric": [["band gap", "!=", 1.5]], "phased": true }],
+  [
+    "band gap>1.5 AND band gap<2.0",
+    {
+      "numeric": [
+        ["band gap", ">", 1.5],
+        ["band gap", "<", 2.0]
+      ],
+      "phased": true
+    }
+  ],
+  ["chalcogens", { "elements": "O-S-Se-Te-Po" }],
+  ["period 2", { "elements": "Li-Be-B-C-N-O-F-Ne" }],
+  ["group 11", { "elements": "Cu-Ag-Au-Rg" }],
+  ["tetrels", { "elements": "C-Si-Ge-Sn-Pb-Fl" }],
+  ["tetrel", { "elements": "C-Si-Ge-Sn-Pb-Fl" }],
+  ["all tetrels", { "elements": "C-Si-Ge-Sn-Pb-Fl" }],
+
+  ["all chalcogens", { "elements": "O-S-Se-Te-Po" }],
+  ["Chalcogens", { "elements": "O-S-Se-Te-Po" }],
+  ["chalcogen", { "elements": "O-S-Se-Te-Po" }],
+  ["CHALCOGENS", { "elements": "O-S-Se-Te-Po" }],
+
+  ["period 2 elements", { "elements": "Li-Be-B-C-N-O-F-Ne" }],
+  ["Period 2", { "elements": "Li-Be-B-C-N-O-F-Ne" }],
+  ["PERIOD 2", { "elements": "Li-Be-B-C-N-O-F-Ne" }],
+
+  ["group 11 elements", { "elements": "Cu-Ag-Au-Rg" }],
+  ["Group 11", { "elements": "Cu-Ag-Au-Rg" }],
+  ["ALL GROUP 11", { "elements": "Cu-Ag-Au-Rg" }],
+
+  ["tetrel", { "elements": "C-Si-Ge-Sn-Pb-Fl" }],
+  ["Tetrels", { "elements": "C-Si-Ge-Sn-Pb-Fl" }],
+  ["ALL tetrels", { "elements": "C-Si-Ge-Sn-Pb-Fl" }]
 ]

From d43c7a5f7b4129542ef172daf993b61cb390345c Mon Sep 17 00:00:00 2001
From: luckylionheart <uyscuti0513@gmail.com>
Date: Fri, 17 Oct 2025 00:27:51 -0400
Subject: [PATCH 2/2] update: disable code format(prettier)

---
 index.js      | 4120 ++++++++++++++++++++++++-------------------------
 test_nlp.json |  489 +++---
 2 files changed, 2226 insertions(+), 2383 deletions(-)

diff --git a/index.js b/index.js
index f85d004..af1bc45 100755
--- a/index.js
+++ b/index.js
@@ -13,893 +13,893 @@
  * Own matchAll used for the chemical formulae
  */
 function getMatchAll(inputstr, regexp) {
-  const matches = [];
-  inputstr.replace(regexp, function (...args) {
-    const arr = [].slice.call(args, 0),
-      extras = arr.splice(-2);
-    arr.index = extras[0];
-    arr.input = extras[1];
-    matches.push(arr);
-  });
-  return matches.length ? matches : null;
+	const matches = [];
+	inputstr.replace(regexp, function (...args) {
+		const arr = [].slice.call(args, 0),
+			extras = arr.splice(-2);
+		arr.index = extras[0];
+		arr.input = extras[1];
+		matches.push(arr);
+	});
+	return matches.length ? matches : null;
 }
 
 function capitalize(string) {
-  return string.charAt(0).toUpperCase() + string.slice(1);
+	return string.charAt(0).toUpperCase() + string.slice(1);
 }
 
 const OptimadeNLP = function () {
-  /*
-   * Definitions
-   */
-  const stop_words = [
-    'a',
-    'about',
-    'above',
-    'after',
-    'again',
-    'against',
-    'all',
-    'am',
-    'an',
-    'and',
-    'any',
-    'are',
-    "aren't",
-    'as',
-    'at',
-    'be',
-    'because',
-    'been',
-    'before',
-    'being',
-    'below',
-    'between',
-    'both',
-    'but',
-    'by',
-    "can't",
-    'cannot',
-    'could',
-    "couldn't",
-    'did',
-    "didn't",
-    'do',
-    'does',
-    "doesn't",
-    'doing',
-    "don't",
-    'down',
-    'during',
-    'each',
-    'few',
-    'for',
-    'from',
-    'further',
-    'had',
-    "hadn't",
-    'has',
-    "hasn't",
-    'have',
-    "haven't",
-    'having',
-    'he',
-    "he'd",
-    "he'll",
-    "he's",
-    'her',
-    'here',
-    "here's",
-    'hers',
-    'herself',
-    'him',
-    'himself',
-    'his',
-    'how',
-    "how's",
-    'i',
-    "i'd",
-    "i'll",
-    "i'm",
-    "i've",
-    'if',
-    'in',
-    'into',
-    'is',
-    "isn't",
-    'it',
-    "it's",
-    'its',
-    'itself',
-    "let's",
-    'me',
-    'more',
-    'most',
-    "mustn't",
-    'my',
-    'myself',
-    'no',
-    'nor',
-    'not',
-    'of',
-    'off',
-    'on',
-    'once',
-    'only',
-    'or',
-    'other',
-    'ought',
-    'our',
-    'ours',
-    'ourselves',
-    'out',
-    'over',
-    'own',
-    'same',
-    "shan't",
-    'she',
-    "she'd",
-    "she'll",
-    "she's",
-    'should',
-    "shouldn't",
-    'so',
-    'some',
-    'such',
-    'than',
-    'that',
-    "that's",
-    'the',
-    'their',
-    'theirs',
-    'them',
-    'themselves',
-    'then',
-    'there',
-    "there's",
-    'these',
-    'they',
-    "they'd",
-    "they'll",
-    "they're",
-    "they've",
-    'this',
-    'those',
-    'through',
-    'to',
-    'too',
-    'u',
-    'under',
-    'until',
-    'up',
-    'very',
-    'was',
-    "wasn't",
-    'we',
-    "we'd",
-    "we'll",
-    "we're",
-    "we've",
-    'were',
-    "weren't",
-    'what',
-    "what's",
-    'when',
-    "when's",
-    'where',
-    "where's",
-    'which',
-    'while',
-    'who',
-    "who's",
-    'whom',
-    'why',
-    "why's",
-    'with',
-    "won't",
-    'would',
-    "wouldn't",
-    'you',
-    "you'd",
-    "you'll",
-    "you're",
-    "you've",
-    'your',
-    'yours',
-    'yourself',
-    'yourselves',
-  ]; /* exact */
-
-  const arity_keys = [
-    null,
-    'unary',
-    'binary',
-    'ternary',
-    'quaternary',
-    'quinary',
-    'multinary',
-    'multinary',
-    'multinary',
-    'multinary',
-    'multinary',
-  ]; // NB null is for "0-ary"
-
-  const periodic_elements = [
-    'h',
-    'he',
-    'li',
-    'be',
-    'b',
-    'c',
-    'n',
-    'o',
-    'f',
-    'ne',
-    'na',
-    'mg',
-    'al',
-    'si',
-    'p',
-    's',
-    'cl',
-    'ar',
-    'k',
-    'ca',
-    'sc',
-    'ti',
-    'v',
-    'cr',
-    'mn',
-    'fe',
-    'co',
-    'ni',
-    'cu',
-    'zn',
-    'ga',
-    'ge',
-    'as',
-    'se',
-    'br',
-    'kr',
-    'rb',
-    'sr',
-    'y',
-    'zr',
-    'nb',
-    'mo',
-    'tc',
-    'ru',
-    'rh',
-    'pd',
-    'ag',
-    'cd',
-    'in',
-    'sn',
-    'sb',
-    'te',
-    'i',
-    'xe',
-    'cs',
-    'ba',
-    'la',
-    'ce',
-    'pr',
-    'nd',
-    'pm',
-    'sm',
-    'eu',
-    'gd',
-    'tb',
-    'dy',
-    'ho',
-    'er',
-    'tm',
-    'yb',
-    'lu',
-    'hf',
-    'ta',
-    'w',
-    're',
-    'os',
-    'ir',
-    'pt',
-    'au',
-    'hg',
-    'tl',
-    'pb',
-    'bi',
-    'po',
-    'at',
-    'rn',
-    'fr',
-    'ra',
-    'ac',
-    'th',
-    'pa',
-    'u',
-    'np',
-    'pu',
-    'am',
-    'cm',
-    'bk',
-    'cf',
-    'es',
-    'fm',
-    'md',
-    'no',
-    'lr',
-    'rf',
-    'db',
-    'sg',
-    'bh',
-    'hs',
-    'mt',
-    'ds',
-    'rg',
-    'cn',
-    'nh',
-    'fl',
-    'mc',
-    'lv',
-    'ts',
-    'og',
-  ]; /* exact */
-
-  const periodic_elements_cased = periodic_elements.map(function (x) {
-    return capitalize(x);
-  });
-
-  const periodic_element_names = [
-    'hydrogen',
-    'helium',
-    'lithium',
-    'beryllium',
-    'boron',
-    'carbon',
-    'nitrogen',
-    'oxygen',
-    'fluorine',
-    'neon',
-    'sodium',
-    'magnesium',
-    'aluminium',
-    'silicon',
-    'phosphorus',
-    'sulfur',
-    'chlorine',
-    'argon',
-    'potassium',
-    'calcium',
-    'scandium',
-    'titanium',
-    'vanadium',
-    'chromium',
-    'manganese',
-    'iron',
-    'cobalt',
-    'nickel',
-    'copper',
-    'zinc',
-    'gallium',
-    'germanium',
-    'arsenic',
-    'selenium',
-    'bromine',
-    'krypton',
-    'rubidium',
-    'strontium',
-    'yttrium',
-    'zirconium',
-    'niobium',
-    'molybdenum',
-    'technetium',
-    'ruthenium',
-    'rhodium',
-    'palladium',
-    'silver',
-    'cadmium',
-    'indium',
-    'tin',
-    'antimony',
-    'tellurium',
-    'iodine',
-    'xenon',
-    'caesium',
-    'barium',
-    'lanthanum',
-    'cerium',
-    'praseodymium',
-    'neodymium',
-    'promethium',
-    'samarium',
-    'europium',
-    'gadolinium',
-    'terbium',
-    'dysprosium',
-    'holmium',
-    'erbium',
-    'thulium',
-    'ytterbium',
-    'lutetium',
-    'hafnium',
-    'tantalum',
-    'tungsten',
-    'rhenium',
-    'osmium',
-    'iridium',
-    'platinum',
-    'gold',
-    'mercury',
-    'thallium',
-    'lead',
-    'bismuth',
-    'polonium',
-    'astatine',
-    'radon',
-    'francium',
-    'radium',
-    'actinium',
-    'thorium',
-    'protactinium',
-    'uranium',
-    'neptunium',
-    'plutonium',
-    'americium',
-    'curium',
-    'berkelium',
-    'californium',
-    'einsteinium',
-    'fermium',
-    'mendelevium',
-    'nobelium',
-    'lawrencium',
-    'rutherfordium',
-    'dubnium',
-    'seaborgium',
-    'bohrium',
-    'hassium',
-    'meitnerium',
-    'darmstadium',
-    'roentgenium',
-    'copernicium',
-    'nihonium',
-    'flerovium',
-    'moscovium',
-    'livermorium',
-    'tennessine',
-    'oganesson',
-  ]; /* fuzzy */
-
-  const lat_p2i = {
-    cubic: 1,
-    hexagonal: 2,
-    trigonal: 3,
-    tetragonal: 4,
-    orthorhombic: 5,
-    monoclinic: 6,
-    triclinic: 7,
-    rhombohedral: 3,
-    cub: 1,
-    hex: 2,
-    hexag: 2,
-    trig: 3,
-    tet: 4,
-    tetr: 4,
-    tetrag: 4,
-    orth: 5,
-    ortho: 5,
-    monocl: 6,
-    tric: 7,
-    tricl: 7,
-    rhom: 3,
-    rhomb: 3,
-  };
-  const lat_fgrs = Object.keys(lat_p2i);
-  const lat_i2p = {
-    1: 'cubic',
-    2: 'hexagonal',
-    3: 'trigonal',
-    4: 'tetragonal',
-    5: 'orthorhombic',
-    6: 'monoclinic',
-    7: 'triclinic',
-  };
-
-  const mpds_classes = [
-    'ab initio calculations',
-    'ab initio literature',
-    'actinoid',
-    'adamantane',
-    'aegirine',
-    'alkali',
-    'alkaline',
-    'allargentum',
-    'almandine',
-    'alum',
-    'alunogen',
-    'amide',
-    'analcime',
-    'anatase',
-    'anorpiment',
-    'anorthoclase',
-    'antiferroelectric',
-    'antiferromagnet',
-    'antiferromagnetic',
-    'arsenate',
-    'arsenide',
-    'ashcroftine',
-    'auricupride',
-    'aurocupride',
-    'azide',
-    'baileychlore',
-    'bariopyrochlore',
-    'baryte',
-    'beryl',
-    'beta-alumina',
-    'beta-boron',
-    'biguanide',
-    'binary',
-    'birefringent',
-    'borane',
-    'borate',
-    'borax',
-    'boride',
-    'borocarbide',
-    'borohydride',
-    'boronitride',
-    'botryogen',
-    'bromanilate',
-    'bromide',
-    'bromoimide',
-    'calomel',
-    'carbamate',
-    'carbide',
-    'carbonate',
-    'carbonyl',
-    'carboxylate',
-    'celestine',
-    'cell and atoms',
-    'cell-only',
-    'celsian',
-    'cesiokenopyrochlore',
-    'chalcogen',
-    'charge-density wave state',
-    'chevrel',
-    'chimney-ladder',
-    'chloranilate',
-    'chlorate',
-    'chloride',
-    'chloritoid',
-    'chlorosulfate',
-    'chromate',
-    'chrysoberyl',
-    'chrysotile',
-    'cinnabar',
-    'clathrate',
-    'clinochlore',
-    'clinoclase',
-    'clodronate',
-    'close-packed',
-    'cluster glass',
-    'colossal magnetoresistance',
-    'conductor',
-    'corundum',
-    'cosmochlor',
-    'croconate violet',
-    'croconate',
-    'cryptomelane',
-    'cuprate',
-    'cuspidine',
-    'cyamelurate',
-    'cyanamide',
-    'cyanamidonitrate',
-    'cyananilate',
-    'cyanide',
-    'cyanotetrazolate',
-    'cyanoureate',
-    'cyanurate',
-    'cyprine',
-    'davyne',
-    'deuteride',
-    'deuterium',
-    'devilline',
-    'diamagnetic',
-    'diamond',
-    'diarsenate',
-    'diaspore',
-    'diazanide',
-    'diazenide',
-    'dichromate',
-    'digermanate',
-    'diiodobromide',
-    'dinitramide',
-    'diopside',
-    'dioptase',
-    'dioxobromate',
-    'dioxoiodate',
-    'dioxosulfate',
-    'dioxothiosulfate',
-    'diphosphate',
-    'diphosphonate',
-    'dipolyhedral',
-    'diselenate',
-    'disilicate',
-    'disordered',
-    'disulfate',
-    'dithiocarbamate',
-    'dithiocarbonate',
-    'dithionate',
-    'dithiooxalate',
-    'dithiophosphate',
-    'dithiosquarate',
-    'divanadate',
-    'epidote',
-    'euchlorine',
-    'euclase',
-    'eudialyte',
-    'eulytine',
-    'fermi liquid',
-    'feroxihyte',
-    'feroxyhyte',
-    'ferrielectric',
-    'ferrimagnet',
-    'ferroelastic',
-    'ferroelectric',
-    'ferromagnet',
-    'ferromagnetic',
-    'fluor-schorl',
-    'fluoride',
-    'fluoroborate',
-    'frank-kasper',
-    'friauf-laves',
-    'fulleride',
-    'fulminate',
-    'galena',
-    'gamma-brass',
-    'garnet',
-    'giant magnetocaloric effect',
-    'gismondine',
-    'glaucodot',
-    'glaucophane',
-    'grossular',
-    'guanidinate',
-    'gypsum',
-    'hafnon',
-    'half metal',
-    'halogen',
-    'hard magnet',
-    'harmotome',
-    'haueyne',
-    'heavy fermion',
-    'hedyphane',
-    'helimagnet',
-    'helvine',
-    'hexasulfate',
-    'hexathionate',
-    'host-guest',
-    'humboldtine',
-    'hydrate',
-    'hydride',
-    'hydroxide',
-    'hypercinnabar',
-    'hypophosphate',
-    'ice',
-    'imide',
-    'iminate',
-    'intercalation',
-    'intermediate valence',
-    'intermetallic',
-    'iodate',
-    'iodide',
-    'ionic conductor',
-    'iridium',
-    'isoferroplatinum',
-    'isopolyhedral',
-    'isothermal section',
-    'kornerupine',
-    'kosmochlor',
-    'lanthanoid',
-    'lavendulan',
-    'levyne',
-    'lime',
-    'liquidus projection',
-    'litharge',
-    'lithiophosphate',
-    'luminescent',
-    'machine learning',
-    'machine-learning',
-    'magnesiochloritoid',
-    'magnetoelastic',
-    'magnetoelectric',
-    'manganate',
-    'massicot',
-    'mellitate',
-    'melonate',
-    'metacinnabar',
-    'metal',
-    'metalloid',
-    'metamagnet',
-    'metavoltine',
-    'mica',
-    'microcline',
-    'microline',
-    'mictomagnet',
-    'minium',
-    'molybdate',
-    'multiferroic',
-    'multinary',
-    'nasicon',
-    'natron',
-    'natrophosphate',
-    'natroxalate',
-    'negative thermal expansion',
-    'nepheline',
-    'nickeline',
-    'niobocarbide',
-    'niter',
-    'nitranilate',
-    'nitrate',
-    'nitratine',
-    'nitride',
-    'nitroformate',
-    'noble gas',
-    'non disordered',
-    'non-disordered',
-    'non-linear optics',
-    'nonaflate',
-    'nonmetal',
-    'nordenskioeldine',
-    'nosean',
-    'olivine',
-    'optically isotropic',
-    'organic',
-    'orpiment',
-    'orthoborate',
-    'orthoclase',
-    'orthogermanate',
-    'orthonitrate',
-    'orthophosphate',
-    'orthosilicate',
-    'oxalate',
-    'oxamate',
-    'oxide',
-    'oxoiodate',
-    'oxonitrate',
-    'oxotetrazolate',
-    'oxy-schorl',
-    'ozonide',
-    'palladogermanide',
-    'paracelsian',
-    'paramagnet',
-    'pararealgar',
-    'pauli paramagnet',
-    'peer review',
-    'peer reviewed',
-    'peer-review',
-    'peer-reviewed',
-    'perchlorate',
-    'perhydrate',
-    'periclase',
-    'periodate',
-    'permanganate',
-    'pernitride',
-    'peroxide',
-    'pertechnetate',
-    'phosphate',
-    'phosphide',
-    'phosphinate',
-    'phosphonate',
-    'photocatalyst',
-    'photovoltaic effect',
-    'photovoltaic',
-    'piezoelectric',
-    'pnictogen',
-    'polaron conductor',
-    'polycrase',
-    'potassic',
-    'prism',
-    'prussian blue',
-    'pseudorutile',
-    'pyrochlore',
-    'pyroelectric',
-    'pyrope',
-    'quadridavyne',
-    'quartz',
-    'quasicrystal',
-    'quaternary',
-    'quinary',
-    'radioactive',
-    'rare earth',
-    'realgar',
-    'refractory',
-    'relaxor',
-    'retzian',
-    'rhodarsenide',
-    'rhodizonate',
-    'rhomboclase',
-    'rocksalt',
-    'rubicline',
-    'ruddlesden-popper',
-    'rutheniridosmine',
-    'rutherfordine',
-    'rutile',
-    'salammoniac',
-    'sanidine',
-    'sapphirine',
-    'sarcopside',
-    'schorl',
-    'selenide',
-    'selenidel',
-    'semiconductor',
-    'semimetal',
-    'shape memory effect',
-    'shape memory',
-    'siderotil',
-    'silanide',
-    'silicate',
-    'silicide',
-    'sillen-aurivillius',
-    'skyrmion',
-    'solidus projection',
-    'solvus projection',
-    'spessartine',
-    'spin glass',
-    'spinel',
-    'spodumene',
-    'squarate',
-    'steenstrupine',
-    'stibarsen',
-    'subsolidus relations',
-    'sulfamate',
-    'sulfamide',
-    'sulfate',
-    'sulfide',
-    'sulfidel',
-    'sulfinylamide',
-    'superconductor',
-    'superhard',
-    'superionic conductor',
-    'superoxide',
-    'talc',
-    'tantalcarbide',
-    'tellurantimony',
-    'telluride',
-    'ternary',
-    'tetrahedron',
-    'thermoelectric',
-    'thiocyanate',
-    'thiocyanurate',
-    'thiophosphate',
-    'thiosulfate',
-    'topaz',
-    'topological insulator',
-    'tourmaline',
-    'transitional',
-    'triflate',
-    'tripolyhedral',
-    'triteride',
-    'trithionate',
-    'tritide',
-    'tritium',
-    'trona',
-    'tungstate',
-    'turquoise',
-    'ulvoespinel',
-    'unary',
-    'uranophane',
-    'uranopolycrase',
-    'urate',
-    'urea',
-    'ureate',
-    'van vleck paramagnet',
-    'vanadate',
-    'vertical section',
-    'violurate',
-    'xenotime',
-    'zincobotryogen',
-    'zircon',
-    'zircosulfate',
-  ];
+	/*
+	 * Definitions
+	 */
+	const stop_words = [
+		'a',
+		'about',
+		'above',
+		'after',
+		'again',
+		'against',
+		'all',
+		'am',
+		'an',
+		'and',
+		'any',
+		'are',
+		"aren't",
+		'as',
+		'at',
+		'be',
+		'because',
+		'been',
+		'before',
+		'being',
+		'below',
+		'between',
+		'both',
+		'but',
+		'by',
+		"can't",
+		'cannot',
+		'could',
+		"couldn't",
+		'did',
+		"didn't",
+		'do',
+		'does',
+		"doesn't",
+		'doing',
+		"don't",
+		'down',
+		'during',
+		'each',
+		'few',
+		'for',
+		'from',
+		'further',
+		'had',
+		"hadn't",
+		'has',
+		"hasn't",
+		'have',
+		"haven't",
+		'having',
+		'he',
+		"he'd",
+		"he'll",
+		"he's",
+		'her',
+		'here',
+		"here's",
+		'hers',
+		'herself',
+		'him',
+		'himself',
+		'his',
+		'how',
+		"how's",
+		'i',
+		"i'd",
+		"i'll",
+		"i'm",
+		"i've",
+		'if',
+		'in',
+		'into',
+		'is',
+		"isn't",
+		'it',
+		"it's",
+		'its',
+		'itself',
+		"let's",
+		'me',
+		'more',
+		'most',
+		"mustn't",
+		'my',
+		'myself',
+		'no',
+		'nor',
+		'not',
+		'of',
+		'off',
+		'on',
+		'once',
+		'only',
+		'or',
+		'other',
+		'ought',
+		'our',
+		'ours',
+		'ourselves',
+		'out',
+		'over',
+		'own',
+		'same',
+		"shan't",
+		'she',
+		"she'd",
+		"she'll",
+		"she's",
+		'should',
+		"shouldn't",
+		'so',
+		'some',
+		'such',
+		'than',
+		'that',
+		"that's",
+		'the',
+		'their',
+		'theirs',
+		'them',
+		'themselves',
+		'then',
+		'there',
+		"there's",
+		'these',
+		'they',
+		"they'd",
+		"they'll",
+		"they're",
+		"they've",
+		'this',
+		'those',
+		'through',
+		'to',
+		'too',
+		'u',
+		'under',
+		'until',
+		'up',
+		'very',
+		'was',
+		"wasn't",
+		'we',
+		"we'd",
+		"we'll",
+		"we're",
+		"we've",
+		'were',
+		"weren't",
+		'what',
+		"what's",
+		'when',
+		"when's",
+		'where',
+		"where's",
+		'which',
+		'while',
+		'who',
+		"who's",
+		'whom',
+		'why',
+		"why's",
+		'with',
+		"won't",
+		'would',
+		"wouldn't",
+		'you',
+		"you'd",
+		"you'll",
+		"you're",
+		"you've",
+		'your',
+		'yours',
+		'yourself',
+		'yourselves',
+	]; /* exact */
+
+	const arity_keys = [
+		null,
+		'unary',
+		'binary',
+		'ternary',
+		'quaternary',
+		'quinary',
+		'multinary',
+		'multinary',
+		'multinary',
+		'multinary',
+		'multinary',
+	]; // NB null is for "0-ary"
+
+	const periodic_elements = [
+		'h',
+		'he',
+		'li',
+		'be',
+		'b',
+		'c',
+		'n',
+		'o',
+		'f',
+		'ne',
+		'na',
+		'mg',
+		'al',
+		'si',
+		'p',
+		's',
+		'cl',
+		'ar',
+		'k',
+		'ca',
+		'sc',
+		'ti',
+		'v',
+		'cr',
+		'mn',
+		'fe',
+		'co',
+		'ni',
+		'cu',
+		'zn',
+		'ga',
+		'ge',
+		'as',
+		'se',
+		'br',
+		'kr',
+		'rb',
+		'sr',
+		'y',
+		'zr',
+		'nb',
+		'mo',
+		'tc',
+		'ru',
+		'rh',
+		'pd',
+		'ag',
+		'cd',
+		'in',
+		'sn',
+		'sb',
+		'te',
+		'i',
+		'xe',
+		'cs',
+		'ba',
+		'la',
+		'ce',
+		'pr',
+		'nd',
+		'pm',
+		'sm',
+		'eu',
+		'gd',
+		'tb',
+		'dy',
+		'ho',
+		'er',
+		'tm',
+		'yb',
+		'lu',
+		'hf',
+		'ta',
+		'w',
+		're',
+		'os',
+		'ir',
+		'pt',
+		'au',
+		'hg',
+		'tl',
+		'pb',
+		'bi',
+		'po',
+		'at',
+		'rn',
+		'fr',
+		'ra',
+		'ac',
+		'th',
+		'pa',
+		'u',
+		'np',
+		'pu',
+		'am',
+		'cm',
+		'bk',
+		'cf',
+		'es',
+		'fm',
+		'md',
+		'no',
+		'lr',
+		'rf',
+		'db',
+		'sg',
+		'bh',
+		'hs',
+		'mt',
+		'ds',
+		'rg',
+		'cn',
+		'nh',
+		'fl',
+		'mc',
+		'lv',
+		'ts',
+		'og'
+	]; /* exact */
+
+	const periodic_elements_cased = periodic_elements.map(function (x) {
+		return capitalize(x);
+	});
+
+	const periodic_element_names = [
+		'hydrogen',
+		'helium',
+		'lithium',
+		'beryllium',
+		'boron',
+		'carbon',
+		'nitrogen',
+		'oxygen',
+		'fluorine',
+		'neon',
+		'sodium',
+		'magnesium',
+		'aluminium',
+		'silicon',
+		'phosphorus',
+		'sulfur',
+		'chlorine',
+		'argon',
+		'potassium',
+		'calcium',
+		'scandium',
+		'titanium',
+		'vanadium',
+		'chromium',
+		'manganese',
+		'iron',
+		'cobalt',
+		'nickel',
+		'copper',
+		'zinc',
+		'gallium',
+		'germanium',
+		'arsenic',
+		'selenium',
+		'bromine',
+		'krypton',
+		'rubidium',
+		'strontium',
+		'yttrium',
+		'zirconium',
+		'niobium',
+		'molybdenum',
+		'technetium',
+		'ruthenium',
+		'rhodium',
+		'palladium',
+		'silver',
+		'cadmium',
+		'indium',
+		'tin',
+		'antimony',
+		'tellurium',
+		'iodine',
+		'xenon',
+		'caesium',
+		'barium',
+		'lanthanum',
+		'cerium',
+		'praseodymium',
+		'neodymium',
+		'promethium',
+		'samarium',
+		'europium',
+		'gadolinium',
+		'terbium',
+		'dysprosium',
+		'holmium',
+		'erbium',
+		'thulium',
+		'ytterbium',
+		'lutetium',
+		'hafnium',
+		'tantalum',
+		'tungsten',
+		'rhenium',
+		'osmium',
+		'iridium',
+		'platinum',
+		'gold',
+		'mercury',
+		'thallium',
+		'lead',
+		'bismuth',
+		'polonium',
+		'astatine',
+		'radon',
+		'francium',
+		'radium',
+		'actinium',
+		'thorium',
+		'protactinium',
+		'uranium',
+		'neptunium',
+		'plutonium',
+		'americium',
+		'curium',
+		'berkelium',
+		'californium',
+		'einsteinium',
+		'fermium',
+		'mendelevium',
+		'nobelium',
+		'lawrencium',
+		'rutherfordium',
+		'dubnium',
+		'seaborgium',
+		'bohrium',
+		'hassium',
+		'meitnerium',
+		'darmstadium',
+		'roentgenium',
+		'copernicium',
+		'nihonium',
+		'flerovium',
+		'moscovium',
+		'livermorium',
+		'tennessine',
+		'oganesson'
+	]; /* fuzzy */
+
+	const lat_p2i = {
+		cubic: 1,
+		hexagonal: 2,
+		trigonal: 3,
+		tetragonal: 4,
+		orthorhombic: 5,
+		monoclinic: 6,
+		triclinic: 7,
+		rhombohedral: 3,
+		cub: 1,
+		hex: 2,
+		hexag: 2,
+		trig: 3,
+		tet: 4,
+		tetr: 4,
+		tetrag: 4,
+		orth: 5,
+		ortho: 5,
+		monocl: 6,
+		tric: 7,
+		tricl: 7,
+		rhom: 3,
+		rhomb: 3,
+	};
+	const lat_fgrs = Object.keys(lat_p2i);
+	const lat_i2p = {
+		1: 'cubic',
+		2: 'hexagonal',
+		3: 'trigonal',
+		4: 'tetragonal',
+		5: 'orthorhombic',
+		6: 'monoclinic',
+		7: 'triclinic',
+	};
+
+	const mpds_classes = [
+		'ab initio calculations',
+		'ab initio literature',
+		'actinoid',
+		'adamantane',
+		'aegirine',
+		'alkali',
+		'alkaline',
+		'allargentum',
+		'almandine',
+		'alum',
+		'alunogen',
+		'amide',
+		'analcime',
+		'anatase',
+		'anorpiment',
+		'anorthoclase',
+		'antiferroelectric',
+		'antiferromagnet',
+		'antiferromagnetic',
+		'arsenate',
+		'arsenide',
+		'ashcroftine',
+		'auricupride',
+		'aurocupride',
+		'azide',
+		'baileychlore',
+		'bariopyrochlore',
+		'baryte',
+		'beryl',
+		'beta-alumina',
+		'beta-boron',
+		'biguanide',
+		'binary',
+		'birefringent',
+		'borane',
+		'borate',
+		'borax',
+		'boride',
+		'borocarbide',
+		'borohydride',
+		'boronitride',
+		'botryogen',
+		'bromanilate',
+		'bromide',
+		'bromoimide',
+		'calomel',
+		'carbamate',
+		'carbide',
+		'carbonate',
+		'carbonyl',
+		'carboxylate',
+		'celestine',
+		'cell and atoms',
+		'cell-only',
+		'celsian',
+		'cesiokenopyrochlore',
+		'chalcogen',
+		'charge-density wave state',
+		'chevrel',
+		'chimney-ladder',
+		'chloranilate',
+		'chlorate',
+		'chloride',
+		'chloritoid',
+		'chlorosulfate',
+		'chromate',
+		'chrysoberyl',
+		'chrysotile',
+		'cinnabar',
+		'clathrate',
+		'clinochlore',
+		'clinoclase',
+		'clodronate',
+		'close-packed',
+		'cluster glass',
+		'colossal magnetoresistance',
+		'conductor',
+		'corundum',
+		'cosmochlor',
+		'croconate violet',
+		'croconate',
+		'cryptomelane',
+		'cuprate',
+		'cuspidine',
+		'cyamelurate',
+		'cyanamide',
+		'cyanamidonitrate',
+		'cyananilate',
+		'cyanide',
+		'cyanotetrazolate',
+		'cyanoureate',
+		'cyanurate',
+		'cyprine',
+		'davyne',
+		'deuteride',
+		'deuterium',
+		'devilline',
+		'diamagnetic',
+		'diamond',
+		'diarsenate',
+		'diaspore',
+		'diazanide',
+		'diazenide',
+		'dichromate',
+		'digermanate',
+		'diiodobromide',
+		'dinitramide',
+		'diopside',
+		'dioptase',
+		'dioxobromate',
+		'dioxoiodate',
+		'dioxosulfate',
+		'dioxothiosulfate',
+		'diphosphate',
+		'diphosphonate',
+		'dipolyhedral',
+		'diselenate',
+		'disilicate',
+		'disordered',
+		'disulfate',
+		'dithiocarbamate',
+		'dithiocarbonate',
+		'dithionate',
+		'dithiooxalate',
+		'dithiophosphate',
+		'dithiosquarate',
+		'divanadate',
+		'epidote',
+		'euchlorine',
+		'euclase',
+		'eudialyte',
+		'eulytine',
+		'fermi liquid',
+		'feroxihyte',
+		'feroxyhyte',
+		'ferrielectric',
+		'ferrimagnet',
+		'ferroelastic',
+		'ferroelectric',
+		'ferromagnet',
+		'ferromagnetic',
+		'fluor-schorl',
+		'fluoride',
+		'fluoroborate',
+		'frank-kasper',
+		'friauf-laves',
+		'fulleride',
+		'fulminate',
+		'galena',
+		'gamma-brass',
+		'garnet',
+		'giant magnetocaloric effect',
+		'gismondine',
+		'glaucodot',
+		'glaucophane',
+		'grossular',
+		'guanidinate',
+		'gypsum',
+		'hafnon',
+		'half metal',
+		'halogen',
+		'hard magnet',
+		'harmotome',
+		'haueyne',
+		'heavy fermion',
+		'hedyphane',
+		'helimagnet',
+		'helvine',
+		'hexasulfate',
+		'hexathionate',
+		'host-guest',
+		'humboldtine',
+		'hydrate',
+		'hydride',
+		'hydroxide',
+		'hypercinnabar',
+		'hypophosphate',
+		'ice',
+		'imide',
+		'iminate',
+		'intercalation',
+		'intermediate valence',
+		'intermetallic',
+		'iodate',
+		'iodide',
+		'ionic conductor',
+		'iridium',
+		'isoferroplatinum',
+		'isopolyhedral',
+		'isothermal section',
+		'kornerupine',
+		'kosmochlor',
+		'lanthanoid',
+		'lavendulan',
+		'levyne',
+		'lime',
+		'liquidus projection',
+		'litharge',
+		'lithiophosphate',
+		'luminescent',
+		'machine learning',
+		'machine-learning',
+		'magnesiochloritoid',
+		'magnetoelastic',
+		'magnetoelectric',
+		'manganate',
+		'massicot',
+		'mellitate',
+		'melonate',
+		'metacinnabar',
+		'metal',
+		'metalloid',
+		'metamagnet',
+		'metavoltine',
+		'mica',
+		'microcline',
+		'microline',
+		'mictomagnet',
+		'minium',
+		'molybdate',
+		'multiferroic',
+		'multinary',
+		'nasicon',
+		'natron',
+		'natrophosphate',
+		'natroxalate',
+		'negative thermal expansion',
+		'nepheline',
+		'nickeline',
+		'niobocarbide',
+		'niter',
+		'nitranilate',
+		'nitrate',
+		'nitratine',
+		'nitride',
+		'nitroformate',
+		'noble gas',
+		'non disordered',
+		'non-disordered',
+		'non-linear optics',
+		'nonaflate',
+		'nonmetal',
+		'nordenskioeldine',
+		'nosean',
+		'olivine',
+		'optically isotropic',
+		'organic',
+		'orpiment',
+		'orthoborate',
+		'orthoclase',
+		'orthogermanate',
+		'orthonitrate',
+		'orthophosphate',
+		'orthosilicate',
+		'oxalate',
+		'oxamate',
+		'oxide',
+		'oxoiodate',
+		'oxonitrate',
+		'oxotetrazolate',
+		'oxy-schorl',
+		'ozonide',
+		'palladogermanide',
+		'paracelsian',
+		'paramagnet',
+		'pararealgar',
+		'pauli paramagnet',
+		'peer review',
+		'peer reviewed',
+		'peer-review',
+		'peer-reviewed',
+		'perchlorate',
+		'perhydrate',
+		'periclase',
+		'periodate',
+		'permanganate',
+		'pernitride',
+		'peroxide',
+		'pertechnetate',
+		'phosphate',
+		'phosphide',
+		'phosphinate',
+		'phosphonate',
+		'photocatalyst',
+		'photovoltaic effect',
+		'photovoltaic',
+		'piezoelectric',
+		'pnictogen',
+		'polaron conductor',
+		'polycrase',
+		'potassic',
+		'prism',
+		'prussian blue',
+		'pseudorutile',
+		'pyrochlore',
+		'pyroelectric',
+		'pyrope',
+		'quadridavyne',
+		'quartz',
+		'quasicrystal',
+		'quaternary',
+		'quinary',
+		'radioactive',
+		'rare earth',
+		'realgar',
+		'refractory',
+		'relaxor',
+		'retzian',
+		'rhodarsenide',
+		'rhodizonate',
+		'rhomboclase',
+		'rocksalt',
+		'rubicline',
+		'ruddlesden-popper',
+		'rutheniridosmine',
+		'rutherfordine',
+		'rutile',
+		'salammoniac',
+		'sanidine',
+		'sapphirine',
+		'sarcopside',
+		'schorl',
+		'selenide',
+		'selenidel',
+		'semiconductor',
+		'semimetal',
+		'shape memory effect',
+		'shape memory',
+		'siderotil',
+		'silanide',
+		'silicate',
+		'silicide',
+		'sillen-aurivillius',
+		'skyrmion',
+		'solidus projection',
+		'solvus projection',
+		'spessartine',
+		'spin glass',
+		'spinel',
+		'spodumene',
+		'squarate',
+		'steenstrupine',
+		'stibarsen',
+		'subsolidus relations',
+		'sulfamate',
+		'sulfamide',
+		'sulfate',
+		'sulfide',
+		'sulfidel',
+		'sulfinylamide',
+		'superconductor',
+		'superhard',
+		'superionic conductor',
+		'superoxide',
+		'talc',
+		'tantalcarbide',
+		'tellurantimony',
+		'telluride',
+		'ternary',
+		'tetrahedron',
+		'thermoelectric',
+		'thiocyanate',
+		'thiocyanurate',
+		'thiophosphate',
+		'thiosulfate',
+		'topaz',
+		'topological insulator',
+		'tourmaline',
+		'transitional',
+		'triflate',
+		'tripolyhedral',
+		'triteride',
+		'trithionate',
+		'tritide',
+		'tritium',
+		'trona',
+		'tungstate',
+		'turquoise',
+		'ulvoespinel',
+		'unary',
+		'uranophane',
+		'uranopolycrase',
+		'urate',
+		'urea',
+		'ureate',
+		'van vleck paramagnet',
+		'vanadate',
+		'vertical section',
+		'violurate',
+		'xenotime',
+		'zincobotryogen',
+		'zircon',
+		'zircosulfate',
+	];
 
   // Mapping of common element groups / periods to element symbol arrays
   const ELEMENT_GROUPS_MAP = {
@@ -1035,623 +1035,562 @@ const OptimadeNLP = function () {
     return null;
   }
 
-  const mpds_props = [
-    'acceptor concentration',
-    'acceptor to donor concentration',
-    'activation energy',
-    'adiabatic bulk modulus',
-    'angle-resolved photoelectron spectra',
-    'atomic structure',
-    'band gap',
-    'birefringence',
-    'bremsstrahlung isochromat spectra',
-    'charge carrier concentration',
-    'charge carrier mobility',
-    'charge transfer',
-    'charge-density wave',
-    'charge-transfer energy',
-    'coefficient of schottky term in heat capacity',
-    'coercive electric field',
-    'coercive field',
-    'coherence length',
-    'cohesive energy',
-    'compressibility',
-    'conductivity',
-    'core-electron contribution to magnetic susceptibility',
-    'critical current density',
-    'critical magnetic field',
-    'crystal electric field parameter',
-    'crystal electric field parameters',
-    'crystal electric field splitting',
-    'crystal electric field',
-    'crystal field level',
-    'crystalline structure',
-    'crystal cell',
-    'crystal structure',
-    'curie coefficient',
-    'curie temperature',
-    'curie-weiss paramagnetism',
-    'debye temperature',
-    'decomposition temperature',
-    'decomposition',
-    'diamagnetic contribution to magnetic susceptibility',
-    'dielectric constant',
-    'dielectric loss tangent',
-    'diffusion',
-    'donor concentration',
-    'donor energy',
-    'effective charge',
-    'effective electron number',
-    'effective mass of electrons to holes ratio',
-    'effective mass of electrons',
-    'effective mass',
-    'einstein temperature',
-    'elastic compliance',
-    'elastic moduli',
-    'elastic stiffness coefficient',
-    'elasticity',
-    'electric field gradient',
-    'electric polarization',
-    'electrical conductivity',
-    'electric properties',
-    'electrical properties',
-    'electrical resistivity',
-    'electrochemical impedance spectroscopy',
-    'electron density maps',
-    'electron density of states at fermi level',
-    'electron density of states',
-    'electron energy band structure',
-    'electron energy loss spectra',
-    'electron grueneisen coefficient',
-    'electron mobility',
-    'electron paramagnetic resonance spectra',
-    'electron spin resonance spectra',
-    'electron-phonon interaction parameter',
-    'electronic contribution to heat capacity',
-    'electronic contribution to thermal conductivity',
-    'electronic energy gap',
-    'electronic heat capacity coefficient',
-    'electronic properties',
-    'energy at fermi level',
-    'energy band structure',
-    'energy gap for direct transition',
-    'energy gap for indirect transition',
-    'energy level diagram',
-    'energy of optical phonon',
-    'energy product',
-    'energy',
-    'enthalpy change at melting point',
-    'enthalpy change at phase transition',
-    'enthalpy change at structural transition',
-    'enthalpy change',
-    'enthalpy of formation',
-    'enthalpy of reaction',
-    'enthalpy',
-    'entropy change at melting point',
-    'entropy change at phase transition',
-    'entropy of formation',
-    'entropy of reaction',
-    'entropy',
-    'eutectoid decomposition',
-    'exchange field',
-    'exchange interaction parameter',
-    'exciton energy',
-    'extended x-ray absorption fine structure',
-    'extraordinary refractive index',
-    'fermi energy',
-    'fermi surface',
-    'ferroelasticity',
-    'ferroelectric curie temperature',
-    'ferroelectric hysteresis',
-    'ferroelectric neel temperature',
-    'ferroelectric transitions',
-    'field dependence of resistivity',
-    'figure of merit',
-    'freezing temperature for spin glass',
-    'fusion',
-    'gibbs energy change',
-    'gibbs energy of formation',
-    'gibbs energy of reaction',
-    'ginzburg-landau parameter',
-    'grueneisen coefficient',
-    'gruneisen coefficient',
-    'hall coefficient',
-    'hall effect',
-    'hall mobility',
-    'hardness',
-    'heat capacity at constant pressure',
-    'heat capacity at constant volume',
-    'heat capacity coefficient',
-    'heat capacity discontinuity at structural transition',
-    'heat capacity discontinuity at superconducting transition',
-    'heat capacity discontinuity',
-    'heat capacity',
-    'high-frequency permittivity',
-    'hole mobility',
-    'hydrogen diffusion',
-    'hyperfine magnetic field',
-    'imaginary part of magnetic susceptibility',
-    'imaginary part of permittivity',
-    'inelastic neutron scattering',
-    'inelastic x-ray scattering',
-    'infrared spectra',
-    'ionic conductivity',
-    'irreversibility field',
-    'isomer shift',
-    'isothermal bulk modulus',
-    'isothermal linear compressibility',
-    'isothermal volume compressibility',
-    'knoop hardness',
-    'kondo behavior',
-    'kondo temperature',
-    'lattice',
-    'linear magnetostriction',
-    'linear thermal expansion coefficient',
-    'longitudinal sound velocity',
-    'longitudinal-mode elastic coefficient',
-    'lorentz number',
-    'lower critical magnetic field',
-    'lowest temperature of investigation',
-    'luminescence lifetime',
-    'luminescence',
-    'magnetic anisotropy field',
-    'magnetic anisotropy',
-    'magnetic circular x-ray dichroism',
-    'magnetic dichroism',
-    'magnetic direction',
-    'magnetic entropy',
-    'magnetic field for magnetic transition',
-    'magnetic field for structural transition',
-    'magnetic heat capacity',
-    'magnetic hysteresis',
-    'magnetic moment',
-    'magnetic order',
-    'magnetic penetration depth',
-    'magnetic phase diagram',
-    'magnetic properties',
-    'magnetic resistivity',
-    'magnetic structure',
-    'magnetic susceptibility',
-    'magnetic transitions',
-    'magnetism',
-    'magnetization',
-    'magneto-optical effects',
-    'magneto-optical kerr effect',
-    'magnetostriction',
-    'mechanical properties',
-    'melting temperature',
-    'microhardness',
-    'moessbauer spectra',
-    'mohs hardness',
-    'molar volume',
-    'molecular field parameter',
-    'muon spin spectra',
-    'neel temperature',
-    'neutron energy loss spectra',
-    'non-linear optical properties',
-    'non-linear optics',
-    'nuclear magnetic resonance spectra',
-    'nuclear quadrupolar resonance spectra',
-    'optical absorption coefficient',
-    'optical absorption',
-    'optical conductivity',
-    'optical phonons',
-    'optical properties',
-    'optical spectra',
-    'orbital magnetic moment',
-    'ordinary refractive index',
-    'paraelectric curie coefficient',
-    'paraelectric curie temperature',
-    'paraelectric state',
-    'paramagnetic curie temperature',
-    'paramagnetic moment',
-    'pauli magnetic susceptibility',
-    'peritectic formation',
-    'peritectoid formation',
-    'permittivity',
-    'perturbed angular correlation',
-    'phase diagram',
-    'phase diagrams',
-    'phase transitions',
-    'phonon contribution to thermal conductivity',
-    'phonon density of states',
-    'phonon dispersion',
-    'phonon grueneisen coefficient',
-    'phonon heat capacity at constant pressure',
-    'phonons',
-    'photo-conductivity data',
-    'photo-conductivity',
-    'photoelectron emission spectra',
-    'photoluminescence spectra',
-    'physical properties',
-    'piezoelectric coefficient',
-    'piezoelectric coefficient',
-    'piezoelectricity',
-    'plasma edge',
-    'poisson ratio',
-    'power factor',
-    'pressure derivative of adiabatic bulk modulus',
-    'pressure derivative of curie temperature',
-    'pressure derivative of elastic stiffness coefficient',
-    'pressure derivative of energy gap',
-    'pressure derivative of isothermal bulk modulus',
-    'pressure derivative of neel temperature',
-    'pressure derivative of superconducting transition temperature',
-    'pressure derivative of transition temperature',
-    'pressure for magnetic transition',
-    'pressure for metal-nonmetal transition',
-    'pressure for structural transition',
-    'pyroelectric coefficient',
-    'pyroelectricity',
-    'quadrupole splitting',
-    'raman spectra',
-    'real part of magnetic permeability',
-    'real part of magnetic susceptibility',
-    'real part of optical conductivity',
-    'real part of permittivity',
-    'reflectivity',
-    'refractive index',
-    'relative cooling power',
-    'remanent induction',
-    'remanent magnetic field',
-    'remanent magnetic moment',
-    'remanent magnetization',
-    'remanent polarization',
-    'residual resistivity ratio',
-    'residual resistivity',
-    'resistivity anisotropy',
-    'resistivity',
-    'resonance spectra',
-    'saturation magnetic moment',
-    'saturation magnetization',
-    'second-harmonic generation',
-    'seebeck coefficient',
-    'shear modulus',
-    'soft-x-ray emission spectra',
-    'sound velocity',
-    'spin contribution to magnetic susceptibility',
-    'spin magnetic moment',
-    'spin-fluctuation temperature',
-    'spin-fluctuation',
-    'spin-orbit splitting of valence band',
-    'spin-resolved electron density of states at fermi level',
-    'spontaneous elastic strain',
-    'spontaneous magnetic moment',
-    'spontaneous magnetization',
-    'spontaneous polarization',
-    'static permittivity',
-    'stoner enhancement factor',
-    'stoner parameter',
-    'stoner product',
-    'structural properties',
-    'structural transition',
-    'structural transitions',
-    'superconducting transition temperature',
-    'superconductivity energy gap',
-    'superconductivity phenomena',
-    'superconductivity',
-    'temperature dependence of resistivity',
-    'temperature dependence of static permittivity',
-    'temperature derivative of elastic stiffness coefficient',
-    'temperature derivative of energy gap',
-    'temperature derivative of resistivity',
-    'temperature derivative of upper critical magnetic field',
-    'temperature for congruent melting',
-    'temperature for eutectoid decomposition',
-    'temperature for ferroelectric reordering',
-    'temperature for magnetic transition',
-    'temperature for metal-nonmetal transition',
-    'temperature for peritectic formation',
-    'temperature for peritectoid formation',
-    'temperature for structural transition',
-    'temperature-independent part of magnetic susceptibility',
-    'thermal cell parameters change',
-    'thermal conductivity',
-    'thermal energy gap',
-    'thermal expansion',
-    'thermal properties',
-    'thermal strain',
-    'thermodynamic properties',
-    'thermodynamics',
-    'thermoelectric figure of merit',
-    'thermoelectric power',
-    'total energy calculation data',
-    'transmittance',
-    'transverse sound velocity',
-    'type of magnetism',
-    'upper critical magnetic field',
-    'vacuum ultraviolet photoemission spectra',
-    'valence',
-    'van vleck contribution to magnetic susceptibility',
-    'vibrational spectra',
-    'vickers hardness number',
-    'volume change at phase transition',
-    'volume change at structural transition',
-    'volume change',
-    'volume magnetostriction',
-    'volume thermal expansion coefficient',
-    'wavelength for luminescence',
-    'wavenumber of longitudinal optical phonon',
-    'wavenumber of optical phonon',
-    'wavenumber of transverse optical phonon',
-    'work function',
-    'x-ray absorption near-edge spectra',
-    'x-ray absorption spectra',
-    'x-ray photoemission spectra',
-    'young modulus',
-  ];
-
-  /*
-   * Methods
-   */
-  function is_numeric(n) {
-    return !isNaN(parseFloat(n)) && isFinite(n);
-  }
-
-  /*
-   * Fix chemical formula if needed
-   */
-  function termify_formulae(input, charred) {
-    if (input.includes('&#')) charred = true;
-    const re = charred ? /&#x208(\d);/g : /%u208(\d)/g;
-    input = charred ? input : escape(input);
-    const matches = getMatchAll(input, re);
-    if (matches) {
-      for (let i = 0; i < matches.length; i++) {
-        input = input.replace(matches[i][0], matches[i][1]);
-      }
-    }
-    return unescape(input).replace(/^\(|\)$/g, '');
-  }
-
-  /*
-   * User input processing: brute-force similarity check
-   */
-  function is_like_chem_formula(chk) {
-    //console.log('Checking formula');
-    const len = chk.length;
-
-    let checks;
-
-    if (len > 10) return false;
-    // this cannot be no-index chemical formula
-    else if (len === 2) {
-      checks = [[chk.substr(0, 1), chk.substr(1, 1)]];
-    } else if (len === 3) {
-      checks = [
-        [chk.substr(0, 1), chk.substr(1, 1), chk.substr(2, 1)],
-        [chk.substr(0, 1), chk.substr(1, 2)],
-        [chk.substr(0, 2), chk.substr(2, 1)],
-      ];
-    } else if (len === 4) {
-      checks = [
-        [chk.substr(0, 2), chk.substr(2, 2)],
-        [chk.substr(0, 2), chk.substr(2, 1), chk.substr(3, 1)],
-        [chk.substr(0, 1), chk.substr(1, 1), chk.substr(2, 2)],
-        [chk.substr(0, 1), chk.substr(1, 2), chk.substr(3, 1)],
-        [
-          chk.substr(0, 1),
-          chk.substr(1, 1),
-          chk.substr(2, 1),
-          chk.substr(3, 1),
-        ],
-      ];
-    } else if (len === 5) {
-      checks = [
-        [chk.substr(0, 1), chk.substr(1, 1), chk.substr(2, 1)],
-        [chk.substr(0, 1), chk.substr(1, 1), chk.substr(2, 2)],
-        [chk.substr(0, 2), chk.substr(2, 2), chk.substr(4, 1)],
-        [chk.substr(0, 1), chk.substr(1, 2), chk.substr(3, 2)],
-        [chk.substr(0, 1), chk.substr(1, 2), chk.substr(3, 1)],
-        [chk.substr(0, 2), chk.substr(2, 1), chk.substr(3, 1)],
-        [chk.substr(0, 2), chk.substr(2, 1), chk.substr(3, 2)],
-      ];
-    } else {
-      // 6-9
-      checks = [
-        // NB too improbable to have 5 one-symbol elements in row
-        [chk.substr(0, 2), chk.substr(2, 2), chk.substr(4, 2)], // El-El-El
-        [
-          chk.substr(0, 2),
-          chk.substr(2, 1),
-          chk.substr(3, 1),
-          chk.substr(4, 2),
-        ], // El-E-E-El
-        [
-          chk.substr(0, 2),
-          chk.substr(2, 1),
-          chk.substr(3, 1),
-          chk.substr(4, 1),
-          chk.substr(5, 1),
-        ], // El-E-E-E-E
-        [
-          chk.substr(0, 2),
-          chk.substr(2, 2),
-          chk.substr(4, 1),
-          chk.substr(5, 1),
-        ], // El-El-E-E
-        [
-          chk.substr(0, 2),
-          chk.substr(2, 2),
-          chk.substr(4, 1),
-          chk.substr(5, 2),
-        ], // El-El-E-El
-        [
-          chk.substr(0, 2),
-          chk.substr(2, 1),
-          chk.substr(3, 2),
-          chk.substr(5, 1),
-        ], // El-E-El-E
-        [
-          chk.substr(0, 2),
-          chk.substr(2, 1),
-          chk.substr(3, 2),
-          chk.substr(5, 2),
-        ], // El-E-El-El
-        [
-          chk.substr(0, 1),
-          chk.substr(1, 2),
-          chk.substr(3, 1),
-          chk.substr(4, 2),
-        ], // E-El-E-El
-        [
-          chk.substr(0, 1),
-          chk.substr(1, 2),
-          chk.substr(3, 1),
-          chk.substr(4, 1),
-          chk.substr(5, 1),
-        ], // E-El-E-E-E
-        [
-          chk.substr(0, 1),
-          chk.substr(1, 2),
-          chk.substr(3, 2),
-          chk.substr(5, 1),
-        ], // E-El-El-E
-        [
-          chk.substr(0, 1),
-          chk.substr(1, 2),
-          chk.substr(3, 2),
-          chk.substr(5, 2),
-        ], // E-El-El-El
-        [
-          chk.substr(0, 1),
-          chk.substr(1, 1),
-          chk.substr(2, 2),
-          chk.substr(4, 1),
-          chk.substr(5, 1),
-        ], // E-E-El-E-E
-        [
-          chk.substr(0, 1),
-          chk.substr(1, 1),
-          chk.substr(2, 2),
-          chk.substr(4, 2),
-        ], // E-E-El-El
-        [
-          chk.substr(0, 1),
-          chk.substr(1, 1),
-          chk.substr(2, 1),
-          chk.substr(3, 2),
-          chk.substr(5, 1),
-        ], // E-E-E-El-E
-        [
-          chk.substr(0, 1),
-          chk.substr(1, 1),
-          chk.substr(2, 1),
-          chk.substr(3, 1),
-          chk.substr(4, 2),
-        ], // E-E-E-E-El
-      ];
-    }
-    //console.log(checks);
-
-    for (let i = 0; i < checks.length; i++) {
-      let signals = 0;
-      for (let j = 0; j < checks[i].length; j++) {
-        if (periodic_elements_cased.includes(checks[i][j])) signals++;
-
-        if (signals === checks[i].length) {
-          //console.log(checks[i]);
-          return true;
-        }
-      }
-    }
-    return false;
-  }
-
-  /*
-   * Detect facets: formulae, elements, lattices, and some classes
-   */
-  function try_uniword_facet(term) {
-    if (term === 'AB' || term === 'ABC' || term === 'ABCD') return ['formulae']; // special case-sensitive anonymous cases
-
-    term = term.toLowerCase();
-
-    const maybe_formula = !is_numeric(term.charAt(0));
-
-    const dmatches = getMatchAll(term, /(\d)/g);
-    if (dmatches && dmatches.length > 1 && maybe_formula) return ['formulae']; // no props with more than one digit
-
-    const imatches = getMatchAll(escape(term), /%u208(\d)/g);
-    if (imatches && imatches.length && maybe_formula) return ['formulae']; // no props with subscripts
-
-    if (periodic_elements.includes(term)) return ['elements', capitalize(term)];
-    else if (periodic_element_names.includes(term))
-      return [
-        'elements',
-        capitalize(periodic_elements[periodic_element_names.indexOf(term)]),
-      ];
-
-    if (
-      term.includes('-') &&
-      !term.split('-').some((part) => !periodic_elements.includes(part))
-    ) {
-      return [
-        'elements',
-        term
-          .split('-')
-          .map((el) => capitalize(el))
-          .join('-'),
-      ];
-    }
-
-    if (['element', 'elementary', 'unitary'].includes(term))
-      return ['classes', 'unary'];
-    else if (
-      term === 'quintenary' ||
-      term === 'quinternary' ||
-      term === 'quinternaries' ||
-      term === 'quinaries' ||
-      term === 'pentanary' ||
-      term === 'pentanaries'
-    )
-      return ['classes', 'quinary'];
-    else if (term === 'actinide' || term === 'actinides')
-      return ['classes', 'actinoid'];
-    else if (
-      term === 'lantanide' ||
-      term === 'lantanides' ||
-      term === 'lanthanide' ||
-      term === 'lanthanides' ||
-      term === 'lantanoid' ||
-      term === 'lantanoids'
-    )
-      return ['classes', 'lanthanoid'];
-    else if (term.endsWith('ite') && term.length > 4) return ['classes'];
+	const mpds_props = [
+		'acceptor concentration',
+		'acceptor to donor concentration',
+		'activation energy',
+		'adiabatic bulk modulus',
+		'angle-resolved photoelectron spectra',
+		'atomic structure',
+		'band gap',
+		'birefringence',
+		'bremsstrahlung isochromat spectra',
+		'charge carrier concentration',
+		'charge carrier mobility',
+		'charge transfer',
+		'charge-density wave',
+		'charge-transfer energy',
+		'coefficient of schottky term in heat capacity',
+		'coercive electric field',
+		'coercive field',
+		'coherence length',
+		'cohesive energy',
+		'compressibility',
+		'conductivity',
+		'core-electron contribution to magnetic susceptibility',
+		'critical current density',
+		'critical magnetic field',
+		'crystal electric field parameter',
+		'crystal electric field parameters',
+		'crystal electric field splitting',
+		'crystal electric field',
+		'crystal field level',
+		'crystalline structure',
+		'crystal cell',
+		'crystal structure',
+		'curie coefficient',
+		'curie temperature',
+		'curie-weiss paramagnetism',
+		'debye temperature',
+		'decomposition temperature',
+		'decomposition',
+		'diamagnetic contribution to magnetic susceptibility',
+		'dielectric constant',
+		'dielectric loss tangent',
+		'diffusion',
+		'donor concentration',
+		'donor energy',
+		'effective charge',
+		'effective electron number',
+		'effective mass of electrons to holes ratio',
+		'effective mass of electrons',
+		'effective mass',
+		'einstein temperature',
+		'elastic compliance',
+		'elastic moduli',
+		'elastic stiffness coefficient',
+		'elasticity',
+		'electric field gradient',
+		'electric polarization',
+		'electrical conductivity',
+		'electric properties',
+		'electrical properties',
+		'electrical resistivity',
+		'electrochemical impedance spectroscopy',
+		'electron density maps',
+		'electron density of states at fermi level',
+		'electron density of states',
+		'electron energy band structure',
+		'electron energy loss spectra',
+		'electron grueneisen coefficient',
+		'electron mobility',
+		'electron paramagnetic resonance spectra',
+		'electron spin resonance spectra',
+		'electron-phonon interaction parameter',
+		'electronic contribution to heat capacity',
+		'electronic contribution to thermal conductivity',
+		'electronic energy gap',
+		'electronic heat capacity coefficient',
+		'electronic properties',
+		'energy at fermi level',
+		'energy band structure',
+		'energy gap for direct transition',
+		'energy gap for indirect transition',
+		'energy level diagram',
+		'energy of optical phonon',
+		'energy product',
+		'energy',
+		'enthalpy change at melting point',
+		'enthalpy change at phase transition',
+		'enthalpy change at structural transition',
+		'enthalpy change',
+		'enthalpy of formation',
+		'enthalpy of reaction',
+		'enthalpy',
+		'entropy change at melting point',
+		'entropy change at phase transition',
+		'entropy of formation',
+		'entropy of reaction',
+		'entropy',
+		'eutectoid decomposition',
+		'exchange field',
+		'exchange interaction parameter',
+		'exciton energy',
+		'extended x-ray absorption fine structure',
+		'extraordinary refractive index',
+		'fermi energy',
+		'fermi surface',
+		'ferroelasticity',
+		'ferroelectric curie temperature',
+		'ferroelectric hysteresis',
+		'ferroelectric neel temperature',
+		'ferroelectric transitions',
+		'field dependence of resistivity',
+		'figure of merit',
+		'freezing temperature for spin glass',
+		'fusion',
+		'gibbs energy change',
+		'gibbs energy of formation',
+		'gibbs energy of reaction',
+		'ginzburg-landau parameter',
+		'grueneisen coefficient',
+		'gruneisen coefficient',
+		'hall coefficient',
+		'hall effect',
+		'hall mobility',
+		'hardness',
+		'heat capacity at constant pressure',
+		'heat capacity at constant volume',
+		'heat capacity coefficient',
+		'heat capacity discontinuity at structural transition',
+		'heat capacity discontinuity at superconducting transition',
+		'heat capacity discontinuity',
+		'heat capacity',
+		'high-frequency permittivity',
+		'hole mobility',
+		'hydrogen diffusion',
+		'hyperfine magnetic field',
+		'imaginary part of magnetic susceptibility',
+		'imaginary part of permittivity',
+		'inelastic neutron scattering',
+		'inelastic x-ray scattering',
+		'infrared spectra',
+		'ionic conductivity',
+		'irreversibility field',
+		'isomer shift',
+		'isothermal bulk modulus',
+		'isothermal linear compressibility',
+		'isothermal volume compressibility',
+		'knoop hardness',
+		'kondo behavior',
+		'kondo temperature',
+		'lattice',
+		'linear magnetostriction',
+		'linear thermal expansion coefficient',
+		'longitudinal sound velocity',
+		'longitudinal-mode elastic coefficient',
+		'lorentz number',
+		'lower critical magnetic field',
+		'lowest temperature of investigation',
+		'luminescence lifetime',
+		'luminescence',
+		'magnetic anisotropy field',
+		'magnetic anisotropy',
+		'magnetic circular x-ray dichroism',
+		'magnetic dichroism',
+		'magnetic direction',
+		'magnetic entropy',
+		'magnetic field for magnetic transition',
+		'magnetic field for structural transition',
+		'magnetic heat capacity',
+		'magnetic hysteresis',
+		'magnetic moment',
+		'magnetic order',
+		'magnetic penetration depth',
+		'magnetic phase diagram',
+		'magnetic properties',
+		'magnetic resistivity',
+		'magnetic structure',
+		'magnetic susceptibility',
+		'magnetic transitions',
+		'magnetism',
+		'magnetization',
+		'magneto-optical effects',
+		'magneto-optical kerr effect',
+		'magnetostriction',
+		'mechanical properties',
+		'melting temperature',
+		'microhardness',
+		'moessbauer spectra',
+		'mohs hardness',
+		'molar volume',
+		'molecular field parameter',
+		'muon spin spectra',
+		'neel temperature',
+		'neutron energy loss spectra',
+		'non-linear optical properties',
+		'non-linear optics',
+		'nuclear magnetic resonance spectra',
+		'nuclear quadrupolar resonance spectra',
+		'optical absorption coefficient',
+		'optical absorption',
+		'optical conductivity',
+		'optical phonons',
+		'optical properties',
+		'optical spectra',
+		'orbital magnetic moment',
+		'ordinary refractive index',
+		'paraelectric curie coefficient',
+		'paraelectric curie temperature',
+		'paraelectric state',
+		'paramagnetic curie temperature',
+		'paramagnetic moment',
+		'pauli magnetic susceptibility',
+		'peritectic formation',
+		'peritectoid formation',
+		'permittivity',
+		'perturbed angular correlation',
+		'phase diagram',
+		'phase diagrams',
+		'phase transitions',
+		'phonon contribution to thermal conductivity',
+		'phonon density of states',
+		'phonon dispersion',
+		'phonon grueneisen coefficient',
+		'phonon heat capacity at constant pressure',
+		'phonons',
+		'photo-conductivity data',
+		'photo-conductivity',
+		'photoelectron emission spectra',
+		'photoluminescence spectra',
+		'physical properties',
+		'piezoelectric coefficient',
+		'piezoelectric coefficient',
+		'piezoelectricity',
+		'plasma edge',
+		'poisson ratio',
+		'power factor',
+		'pressure derivative of adiabatic bulk modulus',
+		'pressure derivative of curie temperature',
+		'pressure derivative of elastic stiffness coefficient',
+		'pressure derivative of energy gap',
+		'pressure derivative of isothermal bulk modulus',
+		'pressure derivative of neel temperature',
+		'pressure derivative of superconducting transition temperature',
+		'pressure derivative of transition temperature',
+		'pressure for magnetic transition',
+		'pressure for metal-nonmetal transition',
+		'pressure for structural transition',
+		'pyroelectric coefficient',
+		'pyroelectricity',
+		'quadrupole splitting',
+		'raman spectra',
+		'real part of magnetic permeability',
+		'real part of magnetic susceptibility',
+		'real part of optical conductivity',
+		'real part of permittivity',
+		'reflectivity',
+		'refractive index',
+		'relative cooling power',
+		'remanent induction',
+		'remanent magnetic field',
+		'remanent magnetic moment',
+		'remanent magnetization',
+		'remanent polarization',
+		'residual resistivity ratio',
+		'residual resistivity',
+		'resistivity anisotropy',
+		'resistivity',
+		'resonance spectra',
+		'saturation magnetic moment',
+		'saturation magnetization',
+		'second-harmonic generation',
+		'seebeck coefficient',
+		'shear modulus',
+		'soft-x-ray emission spectra',
+		'sound velocity',
+		'spin contribution to magnetic susceptibility',
+		'spin magnetic moment',
+		'spin-fluctuation temperature',
+		'spin-fluctuation',
+		'spin-orbit splitting of valence band',
+		'spin-resolved electron density of states at fermi level',
+		'spontaneous elastic strain',
+		'spontaneous magnetic moment',
+		'spontaneous magnetization',
+		'spontaneous polarization',
+		'static permittivity',
+		'stoner enhancement factor',
+		'stoner parameter',
+		'stoner product',
+		'structural properties',
+		'structural transition',
+		'structural transitions',
+		'superconducting transition temperature',
+		'superconductivity energy gap',
+		'superconductivity phenomena',
+		'superconductivity',
+		'temperature dependence of resistivity',
+		'temperature dependence of static permittivity',
+		'temperature derivative of elastic stiffness coefficient',
+		'temperature derivative of energy gap',
+		'temperature derivative of resistivity',
+		'temperature derivative of upper critical magnetic field',
+		'temperature for congruent melting',
+		'temperature for eutectoid decomposition',
+		'temperature for ferroelectric reordering',
+		'temperature for magnetic transition',
+		'temperature for metal-nonmetal transition',
+		'temperature for peritectic formation',
+		'temperature for peritectoid formation',
+		'temperature for structural transition',
+		'temperature-independent part of magnetic susceptibility',
+		'thermal cell parameters change',
+		'thermal conductivity',
+		'thermal energy gap',
+		'thermal expansion',
+		'thermal properties',
+		'thermal strain',
+		'thermodynamic properties',
+		'thermodynamics',
+		'thermoelectric figure of merit',
+		'thermoelectric power',
+		'total energy calculation data',
+		'transmittance',
+		'transverse sound velocity',
+		'type of magnetism',
+		'upper critical magnetic field',
+		'vacuum ultraviolet photoemission spectra',
+		'valence',
+		'van vleck contribution to magnetic susceptibility',
+		'vibrational spectra',
+		'vickers hardness number',
+		'volume change at phase transition',
+		'volume change at structural transition',
+		'volume change',
+		'volume magnetostriction',
+		'volume thermal expansion coefficient',
+		'wavelength for luminescence',
+		'wavenumber of longitudinal optical phonon',
+		'wavenumber of optical phonon',
+		'wavenumber of transverse optical phonon',
+		'work function',
+		'x-ray absorption near-edge spectra',
+		'x-ray absorption spectra',
+		'x-ray photoemission spectra',
+		'young modulus',
+	];
+
+	/*
+	 * Methods
+	 */
+	function is_numeric(n) {
+		return !isNaN(parseFloat(n)) && isFinite(n);
+	}
+
+	/*
+	 * Fix chemical formula if needed
+	 */
+	function termify_formulae(input, charred) {
+		if (input.includes('&#')) charred = true;
+		const re = charred ? /&#x208(\d);/g : /%u208(\d)/g;
+		input = charred ? input : escape(input);
+		const matches = getMatchAll(input, re);
+		if (matches) {
+			for (let i = 0; i < matches.length; i++) {
+				input = input.replace(matches[i][0], matches[i][1]);
+			}
+		}
+		return unescape(input).replace(/^\(|\)$/g, '');
+	}
+
+	/*
+	 * User input processing: brute-force similarity check
+	 */
+	function is_like_chem_formula(chk) {
+		//console.log('Checking formula');
+		const len = chk.length;
+
+		let checks;
+
+		if (len > 10) return false;
+		// this cannot be no-index chemical formula
+		else if (len === 2) {
+			checks = [[chk.substr(0, 1), chk.substr(1, 1)]];
+		} else if (len === 3) {
+			checks = [
+				[chk.substr(0, 1), chk.substr(1, 1), chk.substr(2, 1)],
+				[chk.substr(0, 1), chk.substr(1, 2)],
+				[chk.substr(0, 2), chk.substr(2, 1)],
+			];
+		} else if (len === 4) {
+			checks = [
+				[chk.substr(0, 2), chk.substr(2, 2)],
+				[chk.substr(0, 2), chk.substr(2, 1), chk.substr(3, 1)],
+				[chk.substr(0, 1), chk.substr(1, 1), chk.substr(2, 2)],
+				[chk.substr(0, 1), chk.substr(1, 2), chk.substr(3, 1)],
+				[chk.substr(0, 1), chk.substr(1, 1), chk.substr(2, 1), chk.substr(3, 1)],
+			];
+		} else if (len === 5) {
+			checks = [
+				[chk.substr(0, 1), chk.substr(1, 1), chk.substr(2, 1)],
+				[chk.substr(0, 1), chk.substr(1, 1), chk.substr(2, 2)],
+				[chk.substr(0, 2), chk.substr(2, 2), chk.substr(4, 1)],
+				[chk.substr(0, 1), chk.substr(1, 2), chk.substr(3, 2)],
+				[chk.substr(0, 1), chk.substr(1, 2), chk.substr(3, 1)],
+				[chk.substr(0, 2), chk.substr(2, 1), chk.substr(3, 1)],
+				[chk.substr(0, 2), chk.substr(2, 1), chk.substr(3, 2)],
+			];
+		} else {
+			// 6-9
+			checks = [
+				// NB too improbable to have 5 one-symbol elements in row
+				[chk.substr(0, 2), chk.substr(2, 2), chk.substr(4, 2)], // El-El-El
+				[chk.substr(0, 2), chk.substr(2, 1), chk.substr(3, 1), chk.substr(4, 2)], // El-E-E-El
+				[
+					chk.substr(0, 2),
+					chk.substr(2, 1),
+					chk.substr(3, 1),
+					chk.substr(4, 1),
+					chk.substr(5, 1),
+				], // El-E-E-E-E
+				[chk.substr(0, 2), chk.substr(2, 2), chk.substr(4, 1), chk.substr(5, 1)], // El-El-E-E
+				[chk.substr(0, 2), chk.substr(2, 2), chk.substr(4, 1), chk.substr(5, 2)], // El-El-E-El
+				[chk.substr(0, 2), chk.substr(2, 1), chk.substr(3, 2), chk.substr(5, 1)], // El-E-El-E
+				[chk.substr(0, 2), chk.substr(2, 1), chk.substr(3, 2), chk.substr(5, 2)], // El-E-El-El
+				[chk.substr(0, 1), chk.substr(1, 2), chk.substr(3, 1), chk.substr(4, 2)], // E-El-E-El
+				[
+					chk.substr(0, 1),
+					chk.substr(1, 2),
+					chk.substr(3, 1),
+					chk.substr(4, 1),
+					chk.substr(5, 1),
+				], // E-El-E-E-E
+				[chk.substr(0, 1), chk.substr(1, 2), chk.substr(3, 2), chk.substr(5, 1)], // E-El-El-E
+				[chk.substr(0, 1), chk.substr(1, 2), chk.substr(3, 2), chk.substr(5, 2)], // E-El-El-El
+				[
+					chk.substr(0, 1),
+					chk.substr(1, 1),
+					chk.substr(2, 2),
+					chk.substr(4, 1),
+					chk.substr(5, 1),
+				], // E-E-El-E-E
+				[chk.substr(0, 1), chk.substr(1, 1), chk.substr(2, 2), chk.substr(4, 2)], // E-E-El-El
+				[
+					chk.substr(0, 1),
+					chk.substr(1, 1),
+					chk.substr(2, 1),
+					chk.substr(3, 2),
+					chk.substr(5, 1),
+				], // E-E-E-El-E
+				[
+					chk.substr(0, 1),
+					chk.substr(1, 1),
+					chk.substr(2, 1),
+					chk.substr(3, 1),
+					chk.substr(4, 2),
+				], // E-E-E-E-El
+			];
+		}
+		//console.log(checks);
+
+		for (let i = 0; i < checks.length; i++) {
+			let signals = 0;
+			for (let j = 0; j < checks[i].length; j++) {
+				if (periodic_elements_cased.includes(checks[i][j])) signals++;
+
+				if (signals === checks[i].length) {
+					//console.log(checks[i]);
+					return true;
+				}
+			}
+		}
+		return false;
+	}
+
+	/*
+	 * Detect facets: formulae, elements, lattices, and some classes
+	 */
+	function try_uniword_facet(term) {
+		if (term === 'AB' || term === 'ABC' || term === 'ABCD') return ['formulae']; // special case-sensitive anonymous cases
+
+		term = term.toLowerCase();
+
+		const maybe_formula = !is_numeric(term.charAt(0));
+
+		const dmatches = getMatchAll(term, /(\d)/g);
+		if (dmatches && dmatches.length > 1 && maybe_formula) return ['formulae']; // no props with more than one digit
+
+		const imatches = getMatchAll(escape(term), /%u208(\d)/g);
+		if (imatches && imatches.length && maybe_formula) return ['formulae']; // no props with subscripts
+
+		if (periodic_elements.includes(term)) return ['elements', capitalize(term)];
+		else if (periodic_element_names.includes(term))
+			return [ 'elements', capitalize(periodic_elements[periodic_element_names.indexOf(term)]) ];
+
+		if (
+			term.includes('-') &&
+			!term.split('-').some((part) => !periodic_elements.includes(part))
+		) {
+			return [ 'elements', term.split('-').map((el) => capitalize(el)).join('-') ];
+		}
+
+		if (['element', 'elementary', 'unitary'].includes(term)) return ['classes', 'unary'];
+		else if (
+			term === 'quintenary' ||
+			term === 'quinternary' ||
+			term === 'quinternaries' ||
+			term === 'quinaries' ||
+			term === 'pentanary' ||
+			term === 'pentanaries'
+		)
+			return ['classes', 'quinary'];
+		else if (term === 'actinide' || term === 'actinides') return ['classes', 'actinoid'];
+		else if (
+			term === 'lantanide' ||
+			term === 'lantanides' ||
+			term === 'lanthanide' ||
+			term === 'lanthanides' ||
+			term === 'lantanoid' ||
+			term === 'lantanoids'
+		)
+			return ['classes', 'lanthanoid'];
+		else if (term.endsWith('ite') && term.length > 4) return ['classes'];
 
     // direct element-group single-word matches (tetrels, triels, chalcogen, etc.)
     const groupEls = getGroupElements(term);
     if (groupEls) return ['elements', groupEls.join('-')];
 
-    const chk = term
-      .replace(' structure', '')
-      .replace(' lattice', '')
-      .replace(' crystalline', '')
-      .replace(' crystal', '');
-    if (lat_fgrs.includes(chk)) return ['lattices', lat_i2p[lat_p2i[chk]]];
-
-    if (term.length <= 9 && dmatches && maybe_formula) return ['formulae']; // no SHORT props with digits (NB L0, E1)
-
-    return false;
-  }
-
-  /*
-   * Detect facets: classes, props
-   */
-  function try_multiword_facet(term, queue) {
-    term = term.toLowerCase();
-
-    let candidate = false,
-      combined = false,
-      orig = false;
-
-    if (queue.length) {
-      orig = term;
-      combined = true;
-      queue.forEach(function ({ input }) {
-        term = `${input} ${term}`;
-      });
-    }
-    //console.log("CHECKING TERM FOR MULTI-FACET: "+term);
+		const chk = term
+			.replace(' structure', '')
+			.replace(' lattice', '')
+			.replace(' crystalline', '')
+			.replace(' crystal', '');
+		if (lat_fgrs.includes(chk)) return ['lattices', lat_i2p[lat_p2i[chk]]];
+
+		if (term.length <= 9 && dmatches && maybe_formula) return ['formulae']; // no SHORT props with digits (NB L0, E1)
+
+		return false;
+	}
+
+	/*
+	 * Detect facets: classes, props
+	 */
+	function try_multiword_facet(term, queue) {
+		term = term.toLowerCase();
+
+		let candidate = false,
+			combined = false,
+			orig = false;
+
+		if (queue.length) {
+			orig = term;
+			combined = true;
+			queue.forEach(function ({ input }) {
+				term = `${input} ${term}`;
+			});
+		}
+		//console.log("CHECKING TERM FOR MULTI-FACET: "+term);
 
     // Special handling: phrases like 'period 2' or 'group 11'
     const pgMatch = term.match(/^(period|group)\s+(\d{1,2})$/);
@@ -1663,208 +1602,206 @@ const OptimadeNLP = function () {
       if (els) return { facet: 'elements', input: els.join('-'), ready: 1 };
     }
 
-    candidate = check_category(term, 'classes');
-    if (candidate) {
-      if (combined) candidate.combined = true;
-      return candidate;
-    }
-
-    candidate = check_category(term, 'props');
-    if (candidate) {
-      if (combined) candidate.combined = true;
-      return candidate;
-    }
-
-    let single_chk;
-    if (term.endsWith('s')) {
-      // plural-singular fixups
-      single_chk = term.substr(0, term.length - 1);
-
-      if (!combined && single_chk.endsWith('ite'))
-        return { facet: 'classes', input: single_chk, ready: 1 };
-
-      candidate = check_category(single_chk, 'classes');
-      if (candidate) {
-        if (combined) candidate.combined = true;
-        return candidate;
-      }
-    }
-    if (term.endsWith('es')) {
-      // plural-singular fixups
-      single_chk = term.substr(0, term.length - 2);
-
-      if (single_chk === 'binari') single_chk = 'binary';
-      else if (single_chk === 'ternari') single_chk = 'ternary';
-      else if (single_chk === 'quaternari') single_chk = 'quaternary';
-
-      candidate = check_category(single_chk, 'classes');
-      if (candidate) {
-        if (combined) candidate.combined = true;
-        return candidate;
-      }
-    }
-
-    if (!candidate && orig) return try_multiword_facet(orig, []);
-    return false;
-  }
-
-  /*
-   * Utility algo
-   */
-  function check_category(term, category) {
-    const host = category === 'classes' ? mpds_classes : mpds_props;
-
-    if (host.includes(term)) return { facet: category, input: term, ready: 1 };
-
-    const len = host.length;
-    const re = new RegExp(`(?:^|\\s)(${term})(?=\\s|$)`);
-    let idx;
-    for (let i = 0; i < len; i++) {
-      idx = host[i].search(re);
-      if (idx === 0) {
-        //console.log("Found unstrict match in "+category+" with <"+host[i]+">");
-        return { facet: category, input: term, anew: 1 };
-      }
-    }
-    return false;
-  }
-
-  /*
-   * Chemical formula: ABC3 vs. SrTiO3
-   */
-  function is_formula_anonymous(formula) {
-    const detect = new RegExp(
-      /A(\d{0,3})B(\d{0,3})(C(\d{0,3})(D(\d{0,3}))?)?\b/
-    );
-    return formula.charAt(0) === 'A' && detect.test(formula);
-  }
-
-  /*
-   * Get center and ligand information from a string
-   */
-  function parse_ligand(string, start) {
-    const center = string.slice(0, start).toLowerCase();
-
-    if (
-      string.slice(start, start + 1).toLowerCase() === 'x' &&
-      string.slice(start, start + 2).toLowerCase() !== 'xe'
-    ) {
-      if (string.slice(start).length === 1) return [center, 'X'];
-
-      return [center, 'X' + string.slice(start + 1)];
-    }
-
-    if (string.length === start) return [center, 'X'];
-
-    const remainder = string.slice(start);
-
-    if (is_numeric(remainder.slice(0, 1)) && start === 2)
-      return parse_ligand(string, 1);
-
-    return [center, capitalize(remainder)];
-  }
-
-  /*
-   * Get center and ligand information from a string
-   */
-  function _parse_aeatoms(string) {
-    const pos = string.indexOf('-');
-
-    if (pos !== -1) {
-      const center = string.slice(0, pos),
-        ligand = string.slice(pos + 1);
-
-      if (center.length > 2) return false;
-
-      return parse_ligand(center + ligand, center.length);
-    }
-
-    const trials = [2, 1],
-      periodic_elements_xed = ['x'].concat(periodic_elements);
-
-    for (let start = 0; start < 2; start++) {
-      if (
-        string.length >= trials[start] &&
-        periodic_elements_xed.indexOf(
-          string.slice(0, trials[start]).toLowerCase()
-        ) !== -1
-      ) {
-        return parse_ligand(string, trials[start]);
-      }
-    }
-    return false;
-  }
-
-  /**
-   * Get center and ligand information from a string
-   */
-  function parse_aeatoms(string) {
-    const parsed = _parse_aeatoms(string);
-
-    if (!parsed) return ['?', '?'];
-
-    return [capitalize(parsed[0]), formula_to_tags(parsed[1])];
-  }
-
-  /*
-   * Add HTML tags to a chemical formula as a string
-   */
-  function formula_to_tags(string) {
-    let sub = false,
-      html_formula = '';
-
-    for (let i = 0, len = string.length; i < len; i++) {
-      if (is_numeric(string[i]) || string[i] === '.') {
-        if (!sub) {
-          html_formula += '<sub>';
-          sub = true;
-        }
-      } else {
-        if (sub) {
-          html_formula += '</sub>';
-          sub = false;
-        }
-      }
-      html_formula += string[i];
-    }
-    if (sub) html_formula += '</sub>';
-    return html_formula;
-  }
-
-  /*
-   * User input processing: main algorithm
-   */
-  function guess(inputstr) {
-    // *pseudo_numerics*
-    /*if (inputstr.includes('c/a ') || inputstr.includes('a/b ') || inputstr.includes('b/c ')) {
+		candidate = check_category(term, 'classes');
+		if (candidate) {
+			if (combined) candidate.combined = true;
+			return candidate;
+		}
+
+		candidate = check_category(term, 'props');
+		if (candidate) {
+			if (combined) candidate.combined = true;
+			return candidate;
+		}
+
+		let single_chk;
+		if (term.endsWith('s')) {
+			// plural-singular fixups
+			single_chk = term.substr(0, term.length - 1);
+
+			if (!combined && single_chk.endsWith('ite'))
+				return { facet: 'classes', input: single_chk, ready: 1 };
+
+			candidate = check_category(single_chk, 'classes');
+			if (candidate) {
+				if (combined) candidate.combined = true;
+				return candidate;
+			}
+		}
+		if (term.endsWith('es')) {
+			// plural-singular fixups
+			single_chk = term.substr(0, term.length - 2);
+
+			if (single_chk === 'binari') single_chk = 'binary';
+			else if (single_chk === 'ternari') single_chk = 'ternary';
+			else if (single_chk === 'quaternari') single_chk = 'quaternary';
+
+			candidate = check_category(single_chk, 'classes');
+			if (candidate) {
+				if (combined) candidate.combined = true;
+				return candidate;
+			}
+		}
+
+		if (!candidate && orig) return try_multiword_facet(orig, []);
+		return false;
+	}
+
+	/*
+	 * Utility algo
+	 */
+	function check_category(term, category) {
+		const host = category === 'classes' ? mpds_classes : mpds_props;
+
+		if (host.includes(term)) return { facet: category, input: term, ready: 1 };
+
+		const len = host.length;
+		const re = new RegExp(`(?:^|\\s)(${term})(?=\\s|$)`);
+		let idx;
+		for (let i = 0; i < len; i++) {
+			idx = host[i].search(re);
+			if (idx === 0) {
+				//console.log("Found unstrict match in "+category+" with <"+host[i]+">");
+				return { facet: category, input: term, anew: 1 };
+			}
+		}
+		return false;
+	}
+
+	/*
+	 * Chemical formula: ABC3 vs. SrTiO3
+	 */
+	function is_formula_anonymous(formula) {
+		const detect = new RegExp(/A(\d{0,3})B(\d{0,3})(C(\d{0,3})(D(\d{0,3}))?)?\b/);
+		return formula.charAt(0) === 'A' && detect.test(formula);
+	}
+
+	/*
+	 * Get center and ligand information from a string
+	 */
+	function parse_ligand(string, start) {
+
+		const center = string.slice(0, start).toLowerCase();
+
+		if (string.slice(start, start + 1).toLowerCase() === 'x' && string.slice(start, start + 2).toLowerCase() !== 'xe'){
+			if (string.slice(start).length === 1) return [center, 'X'];
+
+			return [center, 'X' + string.slice(start + 1)];
+		}
+
+		if (string.length === start) return [center, 'X'];
+
+		const remainder = string.slice(start);
+
+		if (is_numeric(remainder.slice(0, 1)) && start === 2)
+			return parse_ligand(string, 1);
+
+		return [center, capitalize(remainder)];
+	}
+
+	/*
+	 * Get center and ligand information from a string
+	 */
+	function _parse_aeatoms(string) {
+
+		const pos = string.indexOf('-');
+
+		if (pos !== -1){
+			const center = string.slice(0, pos),
+				ligand = string.slice(pos + 1);
+
+			if (center.length > 2) return false;
+
+			return parse_ligand(center + ligand, center.length);
+		}
+
+		const trials = [2, 1],
+			periodic_elements_xed = ["x"].concat(periodic_elements);
+
+		for (let start = 0; start < 2; start++){
+			if (string.length >= trials[start] && periodic_elements_xed.indexOf(string.slice(0, trials[start]).toLowerCase()) !== -1){
+				return parse_ligand(string, trials[start]);
+			}
+		}
+		return false;
+	}
+
+	/**
+	 * Get center and ligand information from a string
+	 */
+	function parse_aeatoms(string) {
+
+		const parsed = _parse_aeatoms(string);
+
+		if (!parsed) return ['?', '?'];
+
+		return [
+			capitalize(parsed[0]),
+			formula_to_tags(parsed[1])
+		];
+	}
+
+	/*
+	 * Add HTML tags to a chemical formula as a string
+	 */
+	function formula_to_tags(string) {
+
+		let sub = false,
+			html_formula = '';
+
+		for (let i = 0, len = string.length; i < len; i++){
+			if (is_numeric(string[i]) || string[i] === '.'){
+				if (!sub){
+					html_formula += '<sub>';
+					sub = true;
+				}
+			} else {
+				if (sub){
+					html_formula += '</sub>';
+					sub = false;
+				}
+			}
+			html_formula += string[i];
+		}
+		if (sub) html_formula += '</sub>';
+		return html_formula;
+	}
+
+	/*
+	 * User input processing: main algorithm
+	 */
+	function guess(inputstr) {
+
+		// *pseudo_numerics*
+		/*if (inputstr.includes('c/a ') || inputstr.includes('a/b ') || inputstr.includes('b/c ')) {
 			// FIXME slashes in names
 			if (inputstr.includes('c/a ')) inputstr = inputstr.replace('c/a ', 'c--a ');
 			if (inputstr.includes('a/b ')) inputstr = inputstr.replace('a/b ', 'a--b ');
 			if (inputstr.includes('b/c ')) inputstr = inputstr.replace('b/c ', 'b--c ');
 		}*/
 
-    if (inputstr.includes('"')) return { ignored: inputstr }; // Optimade guard
-
-    const tokens = inputstr
-      .replace(new RegExp('\\+|\\!|\\?', 'g'), '')
-      .replace(new RegExp(',|/', 'g'), ' ')
-      .replace(new RegExp('<', 'g'), ' < ')
-      .replace(new RegExp('>', 'g'), ' > ')
-      .replace(new RegExp('=', 'g'), ' = ')
-      .split(/\s+/);
-
-    const result = {};
-    let n_terms = 0;
-    let n_toks = 1;
-    let queue = [];
-    const ignored = [];
-    //console.log(tokens);
-    // TODO: gracefully discard brackets
-
-    tokens.forEach(function (input) {
-      let facet = false,
-        simple = false;
-      input = input.trim();
+		if (inputstr.includes('\"')) return {'ignored': inputstr}; // Optimade guard
+
+		const tokens = inputstr
+			.replace(new RegExp('\\+|\\!|\\?', 'g'), '')
+			.replace(new RegExp(',|/', 'g'), ' ')
+			.replace(new RegExp('<', 'g'), ' < ')
+			.replace(new RegExp('>', 'g'), ' > ')
+			.replace(new RegExp('=', 'g'), ' = ')
+			.split(/\s+/);
+
+		const result = {};
+		let n_terms = 0;
+		let n_toks = 1;
+		let queue = [];
+		const ignored = [];
+		//console.log(tokens);
+		// TODO: gracefully discard brackets
+
+		tokens.forEach(function (input) {
+			let facet = false,
+				simple = false;
+			input = input.trim();
       const linput = input.toLowerCase();
 
       // Allow 'period' and 'group' to be queued so they can combine with a following number
@@ -1875,13 +1812,13 @@ const OptimadeNLP = function () {
         return;
       }
 
-      if (
-        input.includes('<') ||
-        input.includes('>') ||
-        input.includes('=') ||
-        is_numeric(input)
-      ) {
-        // numeric searches
+			if (
+				input.includes('<') ||
+				input.includes('>') ||
+				input.includes('=') ||
+				is_numeric(input)
+			) {
+				// numeric searches
         // If this numeric token is actually the second part of 'period N' or 'group N',
         // don't treat it as a numeric filter but allow multiword facet combining.
         if (
@@ -1892,394 +1829,366 @@ const OptimadeNLP = function () {
         ) {
           // fall-through to normal multiword processing
         } else if (
-          input.indexOf('<') === 0 ||
-          input.indexOf('>') === 0 ||
-          input.indexOf('=') === 0
-        ) {
-          // separated op sign
-          if (result.props) {
-            if (!result.numeric) result.numeric = [];
-            result.numeric.push([result.props, input.substr(0, 1)]);
-            delete result.props;
-          } else if (
-            result.numeric &&
-            result.numeric[result.numeric.length - 1]
-          ) {
-            result.numeric[result.numeric.length - 1][1] = input.substr(0, 1); // TODO account <=, =<, >=, =>
-          } else ignored.push(input);
-        } else if (is_numeric(input)) {
-          if (result.props) {
-            if (!result.numeric) result.numeric = [];
-            result.numeric.push([result.props, '=', parseFloat(input)]);
-            delete result.props;
-          } else if (
-            result.numeric &&
-            result.numeric[result.numeric.length - 1]
-          ) {
-            result.numeric[result.numeric.length - 1][2] = parseFloat(input); // NB no commas!
-            if (!result.numeric[result.numeric.length - 1][1])
-              result.numeric[result.numeric.length - 1][1] = '=';
-          } else ignored.push(input);
-        }
-        return;
-      }
-
-      if (!queue.length || (input != 'at' && input != 'in')) {
-        // FIXME TODO special treatment, words vs. chemical symbols
-        simple = try_uniword_facet(input);
-      }
-
-      if (simple) {
-        facet = simple[0];
-        if (simple[1]) input = simple[1];
-        if (queue.length && !queue[queue.length - 1].ready)
-          ignored.push(...queue.map((obj) => obj.input));
-        queue = [];
-        //console.log(input + ": found simple facet " + simple[0]);
-      } else {
-        if (!stop_words.includes(input)) {
-          input = input
-            .replace(new RegExp('\\(', 'g'), '')
-            .replace(new RegExp('\\)', 'g'), '')
-            .replace(new RegExp('\\[', 'g'), '')
-            .replace(new RegExp('\\]', 'g'), '');
-
-          const candidate = try_multiword_facet(input, queue);
-          //console.log(candidate);
-
-          if (candidate.combined) queue = [];
-
-          if (candidate.ready) {
-            // term found either alone or with previous
-            facet = candidate.facet;
-            input = candidate.input;
-            queue = [candidate];
-          } else if (candidate.anew) {
-            // token anew
-            if (queue.length && !queue[queue.length - 1].ready)
-              ignored.push(...queue.map((obj) => obj.input));
-            queue = [candidate];
-          } else if (!candidate) {
-            // token unknown
-            if (queue.length && !queue[queue.length - 1].ready)
-              ignored.push(...queue.map((obj) => obj.input));
-            queue = [];
-            if (is_like_chem_formula(input)) {
-              facet = 'formulae';
-            }
-            if (!facet) {
-              ignored.push(input);
-            }
-          }
-        }
-      }
-
-      if (n_toks === tokens.length) {
-        // token at the end, terminating
-        if (queue.length && !queue[queue.length - 1].ready)
-          ignored.push(...queue.map((obj) => obj.input));
-        queue = [];
-      }
-
-      if (facet) {
-        if (facet === 'formulae') input = termify_formulae(input);
-
-        if (result[facet]) {
-          // What to do with the found term of the same category?
-          //console.log('Compare: '+result[facet]+' vs. '+input);
-
-          if (facet === 'formulae') {
-            ignored.push(input);
-          } else if (facet === 'elements') {
-            result[facet] += `-${input}`;
-          } else if (facet === 'classes') {
-            result[facet] += `, ${input}`; //escape(input);
-          } else if (facet === 'props') {
-            if (input.includes(result[facet])) {
-              //console.log('Smaller match '+result[facet]+' was thrown away');
-              result[facet] = input;
-            } else {
-              ignored.push(input);
-            }
-          }
-        } else result[facet] = input; //escape(input);
-
-        n_terms++;
-      }
-      n_toks++;
-    });
-
-    result.ignored = ignored;
-    return result;
-  }
-
-  /*
-   * Chemical formulae parsing, courtesy of Nathan Leung, MIT license
-   */
-  function tokenizeInitial(formula, tokens = []) {
-    if (formula.length === 0) {
-      return tokens;
-    }
-
-    // Tokenize parentheses
-    if (formula[0] === '(' || formula[0] === '[') {
-      return tokenizeInitial(formula.substr(1), [
-        ...tokens,
-        {
-          type: 'parenthesis',
-          value: 'open',
-        },
-      ]);
-    }
-    if (formula[0] === ')' || formula[0] === ']') {
-      return tokenizeAfterElement(formula.substr(1), [
-        ...tokens,
-        {
-          type: 'parenthesis',
-          value: 'close',
-        },
-      ]);
-    }
-
-    // Tokenize coefficient
-    let num = '';
-    for (let i = 0; i < formula.length; i++) {
-      // If the char is a number
-      if (!isNaN(Number(formula[i]))) {
-        num = num + formula[i];
-      } else {
-        // End loop if char is not a number
-        break;
-      }
-    }
-    // If a coefficient exists
-    if (num !== '') {
-      return tokenizeInitial(formula.substr(num.length), [
-        ...tokens,
-        {
-          type: 'coefficient',
-          value: Number(num),
-        },
-      ]);
-    }
-
-    // Tokenize element
-    // Check if current char is uppercase letter
-    const char = formula[0];
-    const code = char.charCodeAt(0);
-    if (code >= 65 && code <= 90) {
-      // If next char is a lowercase letter
-      if (formula.length > 1) {
-        const nextChar = formula[1];
-        const nextCode = nextChar.charCodeAt(0);
-        if (nextCode >= 97 && nextCode <= 122) {
-          return tokenizeAfterElement(formula.substr(2), [
-            ...tokens,
-            {
-              type: 'element',
-              value: char + nextChar,
-            },
-          ]);
-        }
-      }
-      return tokenizeAfterElement(formula.substr(1), [
-        ...tokens,
-        {
-          type: 'element',
-          value: char,
-        },
-      ]);
-    }
-
-    // If it doesn't match anything
-    throw Error(`There was an error parsing formula. We were able to get to here:\n${JSON.stringify(
-      tokens,
-      null,
-      2
-    )}
+					input.indexOf('<') === 0 ||
+					input.indexOf('>') === 0 ||
+					input.indexOf('=') === 0
+				) {
+					// separated op sign
+					if (result.props) {
+						if (!result.numeric) result.numeric = [];
+						result.numeric.push([result.props, input.substr(0, 1)]);
+						delete result.props;
+					} else if (result.numeric && result.numeric[result.numeric.length - 1]) {
+						result.numeric[result.numeric.length - 1][1] = input.substr(0, 1); // TODO account <=, =<, >=, =>
+					} else ignored.push(input);
+				} else if (is_numeric(input)) {
+					if (result.props) {
+						if (!result.numeric) result.numeric = [];
+						result.numeric.push([result.props, '=', parseFloat(input)]);
+						delete result.props;
+					} else if (result.numeric && result.numeric[result.numeric.length - 1]) {
+						result.numeric[result.numeric.length - 1][2] = parseFloat(input); // NB no commas!
+						if (!result.numeric[result.numeric.length - 1][1])
+							result.numeric[result.numeric.length - 1][1] = '=';
+					} else ignored.push(input);
+				}
+				return;
+			}
+
+			if (!queue.length || (input != 'at' && input != 'in')) {
+				// FIXME TODO special treatment, words vs. chemical symbols
+				simple = try_uniword_facet(input);
+			}
+
+			if (simple) {
+				facet = simple[0];
+				if (simple[1]) input = simple[1];
+				if (queue.length && !queue[queue.length - 1].ready)
+					ignored.push(...queue.map((obj) => obj.input));
+				queue = [];
+				//console.log(input + ": found simple facet " + simple[0]);
+			} else {
+				if (!stop_words.includes(input)) {
+					input = input
+						.replace(new RegExp('\\(', 'g'), '')
+						.replace(new RegExp('\\)', 'g'), '')
+						.replace(new RegExp('\\[', 'g'), '')
+						.replace(new RegExp('\\]', 'g'), '');
+
+					const candidate = try_multiword_facet(input, queue);
+					//console.log(candidate);
+
+					if (candidate.combined) queue = [];
+
+					if (candidate.ready) {
+						// term found either alone or with previous
+						facet = candidate.facet;
+						input = candidate.input;
+						queue = [candidate];
+					} else if (candidate.anew) {
+						// token anew
+						if (queue.length && !queue[queue.length - 1].ready)
+							ignored.push(...queue.map((obj) => obj.input));
+						queue = [candidate];
+					} else if (!candidate) {
+						// token unknown
+						if (queue.length && !queue[queue.length - 1].ready)
+							ignored.push(...queue.map((obj) => obj.input));
+						queue = [];
+						if (is_like_chem_formula(input)) {
+							facet = 'formulae';
+						}
+						if (!facet) {
+							ignored.push(input);
+						}
+					}
+				}
+			}
+
+			if (n_toks === tokens.length) {
+				// token at the end, terminating
+				if (queue.length && !queue[queue.length - 1].ready)
+					ignored.push(...queue.map((obj) => obj.input));
+				queue = [];
+			}
+
+			if (facet) {
+				if (facet === 'formulae') input = termify_formulae(input);
+
+				if (result[facet]) {
+					// What to do with the found term of the same category?
+					//console.log('Compare: '+result[facet]+' vs. '+input);
+
+					if (facet === 'formulae') {
+						ignored.push(input);
+					} else if (facet === 'elements') {
+						result[facet] += `-${input}`;
+					} else if (facet === 'classes') {
+						result[facet] += `, ${input}`; //escape(input);
+					} else if (facet === 'props') {
+						if (input.includes(result[facet])) {
+							//console.log('Smaller match '+result[facet]+' was thrown away');
+							result[facet] = input;
+						} else {
+							ignored.push(input);
+						}
+					}
+				} else result[facet] = input; //escape(input);
+
+				n_terms++;
+			}
+			n_toks++;
+		});
+
+		result.ignored = ignored;
+		return result;
+	}
+
+	/*
+	 * Chemical formulae parsing, courtesy of Nathan Leung, MIT license
+	 */
+	function tokenizeInitial(formula, tokens = []) {
+		if (formula.length === 0) {
+			return tokens;
+		}
+
+		// Tokenize parentheses
+		if (formula[0] === '(' || formula[0] === '[') {
+			return tokenizeInitial(formula.substr(1), [
+				...tokens, {
+					type: 'parenthesis',
+					value: 'open',
+				},
+			]);
+		}
+		if (formula[0] === ')' || formula[0] === ']') {
+			return tokenizeAfterElement(formula.substr(1), [
+				...tokens, {
+					type: 'parenthesis',
+					value: 'close',
+				},
+			]);
+		}
+
+		// Tokenize coefficient
+		let num = '';
+		for (let i = 0; i < formula.length; i++) {
+			// If the char is a number
+			if (!isNaN(Number(formula[i]))) {
+				num = num + formula[i];
+			} else {
+				// End loop if char is not a number
+				break;
+			}
+		}
+		// If a coefficient exists
+		if (num !== '') {
+			return tokenizeInitial(formula.substr(num.length), [
+				...tokens, {
+					type: 'coefficient',
+					value: Number(num),
+				},
+			]);
+		}
+
+		// Tokenize element
+		// Check if current char is uppercase letter
+		const char = formula[0];
+		const code = char.charCodeAt(0);
+		if (code >= 65 && code <= 90) {
+			// If next char is a lowercase letter
+			if (formula.length > 1) {
+				const nextChar = formula[1];
+				const nextCode = nextChar.charCodeAt(0);
+				if (nextCode >= 97 && nextCode <= 122) {
+					return tokenizeAfterElement(formula.substr(2), [
+						...tokens, {
+							type: 'element',
+							value: char + nextChar,
+						},
+					]);
+				}
+			}
+			return tokenizeAfterElement(formula.substr(1), [
+				...tokens, {
+					type: 'element',
+					value: char,
+				},
+			]);
+		}
+
+		// If it doesn't match anything
+		throw Error(`There was an error parsing formula. We were able to get to here:\n${JSON.stringify(tokens, null, 2)}
 			\n\nRemaining formula: ${formula}`);
-  }
-
-  /*
-   * Chemical formulae parsing, courtesy of Nathan Leung, MIT license
-   */
-  function tokenizeAfterElement(formula, tokens) {
-    if (formula.length === 0) {
-      return tokenizeInitial(formula, tokens);
-    }
-
-    // Tokenize subscript
-    // This is repeated from tokenizeInitial, perhaps
-    // take it out
-    let num = '';
-    for (let i = 0; i < formula.length; i++) {
-      // If the char is a number
-      if (!isNaN(Number(formula[i]))) {
-        num = num + formula[i];
-      } else {
-        // End loop if char is not a number
-        break;
-      }
-    }
-    // If a subscript exists
-    if (num !== '') {
-      return tokenizeInitial(formula.substr(num.length), [
-        ...tokens,
-        {
-          type: 'subscript',
-          value: Number(num),
-        },
-      ]);
-    }
-
-    // If it's not anything, then pass back to tokenizeInitial
-    return tokenizeInitial(formula, tokens);
-  }
-
-  /*
-   * Chemical formulae parsing, courtesy of Nathan Leung, MIT license
-   */
-  function countElements(tokens, elements = {}) {
-    if (tokens.length === 0) {
-      return elements;
-    }
-    // Create keys for each element present in the tokens array
-    if (Object.keys(elements).length === 0) {
-      for (let i = 0; i < tokens.length; i++) {
-        if (tokens[i].type === 'element') {
-          if (typeof elements[tokens[i].value] === 'undefined') {
-            elements[tokens[i].value] = 0;
-          }
-        }
-      }
-    }
-
-    for (let i = 0; i < tokens.length; i++) {
-      if (tokens[i].type === 'coefficient') {
-        return countInGroup(tokens.slice(1), elements, tokens[i].value);
-      } else if (tokens[i].type === 'element') {
-        return countInGroup(tokens, elements);
-      } else if (tokens[i].type === 'parenthesis') {
-        if (tokens[i].value === 'open') {
-          for (let j = i; j < tokens.length; j++) {
-            if (
-              tokens[j].type === 'parenthesis' &&
-              tokens[j].value === 'close'
-            ) {
-              // There will always be a subscript after a parenthetical group
-              if (
-                tokens.length > j + 1 &&
-                tokens[j].type === 'parenthesis' &&
-                tokens[j].value === 'close'
-              ) {
-                if (tokens[j + 1].type === 'subscript') {
-                  return countInGroup(
-                    tokens.slice(1),
-                    elements,
-                    tokens[j + 1].value
-                  );
-                }
-              }
-              //console.error(JSON.stringify(tokens, null, 2));
-              //console.error(JSON.stringify(elements, null, 2))
-              throw Error('Parenthetical group must have subscript');
-            }
-          }
-          return countInGroup(tokens.slice(1), elements);
-        } else {
-          // Skip closing parenthesis and subscript
-          return countInGroup(tokens.slice(2), elements);
-        }
-      }
-    }
-
-    // If none of the above cases are covered then throw an error
-    throw Error(`There was an error:\n\n${JSON.stringify(tokens, null, 2)}
+	}
+
+	/*
+	 * Chemical formulae parsing, courtesy of Nathan Leung, MIT license
+	 */
+	function tokenizeAfterElement(formula, tokens) {
+		if (formula.length === 0) {
+			return tokenizeInitial(formula, tokens);
+		}
+
+		// Tokenize subscript
+		// This is repeated from tokenizeInitial, perhaps
+		// take it out
+		let num = '';
+		for (let i = 0; i < formula.length; i++) {
+			// If the char is a number
+			if (!isNaN(Number(formula[i]))) {
+				num = num + formula[i];
+			} else {
+				// End loop if char is not a number
+				break;
+			}
+		}
+		// If a subscript exists
+		if (num !== '') {
+			return tokenizeInitial(formula.substr(num.length), [
+				...tokens, {
+					type: 'subscript',
+					value: Number(num),
+				},
+			]);
+		}
+
+		// If it's not anything, then pass back to tokenizeInitial
+		return tokenizeInitial(formula, tokens);
+	}
+
+	/*
+	 * Chemical formulae parsing, courtesy of Nathan Leung, MIT license
+	 */
+	function countElements(tokens, elements = {}) {
+		if (tokens.length === 0) {
+			return elements;
+		}
+		// Create keys for each element present in the tokens array
+		if (Object.keys(elements).length === 0) {
+			for (let i = 0; i < tokens.length; i++) {
+				if (tokens[i].type === 'element') {
+					if (typeof elements[tokens[i].value] === 'undefined') {
+						elements[tokens[i].value] = 0;
+					}
+				}
+			}
+		}
+
+		for (let i = 0; i < tokens.length; i++) {
+			if (tokens[i].type === 'coefficient') {
+				return countInGroup(tokens.slice(1), elements, tokens[i].value);
+			} else if (tokens[i].type === 'element') {
+				return countInGroup(tokens, elements);
+			} else if (tokens[i].type === 'parenthesis') {
+				if (tokens[i].value === 'open') {
+					for (let j = i; j < tokens.length; j++) {
+						if (tokens[j].type === 'parenthesis' && tokens[j].value === 'close') {
+							// There will always be a subscript after a parenthetical group
+							if (
+								tokens.length > (j + 1) &&
+								tokens[j].type === 'parenthesis' &&
+								tokens[j].value === 'close'
+							) {
+								if (tokens[j + 1].type === 'subscript') {
+									return countInGroup(tokens.slice(1), elements, tokens[j + 1].value);
+								}
+							}
+							//console.error(JSON.stringify(tokens, null, 2));
+							//console.error(JSON.stringify(elements, null, 2))
+							throw Error('Parenthetical group must have subscript');
+						}
+					}
+					return countInGroup(tokens.slice(1), elements);
+				} else {
+					// Skip closing parenthesis and subscript
+					return countInGroup(tokens.slice(2), elements);
+				}
+			}
+		}
+
+		// If none of the above cases are covered then throw an error
+		throw Error(`There was an error:\n\n${JSON.stringify(tokens, null, 2)}
 			\n\n${JSON.stringify(elements, null, 2)}`);
-  }
-
-  /*
-   * Chemical formulae parsing, courtesy of Nathan Leung, MIT license
-   */
-  function countInGroup(tokens, elements, coefficient = 1) {
-    if (tokens.length === 0) {
-      return countElements(tokens, elements);
-    }
-
-    if (tokens[0].type !== 'element') {
-      // If this happens this is likely a mistake, pass back to countElements
-      return countElements(tokens, elements);
-    }
-
-    let i = 0;
-    while (i < tokens.length) {
-      if (tokens[i].type === 'element') {
-        if (tokens.length > i + 1 && tokens[i + 1].type === 'subscript') {
-          elements[tokens[i].value] += tokens[i + 1].value * coefficient;
-          i += 2;
-          continue;
-        }
-        elements[tokens[i].value] += coefficient;
-        i++;
-        continue;
-      }
-      break;
-    }
-    return countElements(tokens.slice(i), elements);
-  }
-
-  /*
-   * Sorting the formulae alphabetically for Optimade
-   */
-  function sort_formula(text) {
-    let parsed = false;
-    try {
-      parsed = countElements(tokenizeInitial(text));
-    } catch (err) {
-      //console.error(err);
-      return text;
-    }
-    let sorted = '',
-      els = Object.keys(parsed);
-    els.sort();
-    els.forEach(function (el) {
-      sorted += el + (parsed[el] === 1 ? '' : parsed[el]);
-    });
-    return sorted;
-  }
-
-  /*
-   * Convert MPDS search query object notation into the Optimade filter
-   */
-  function to_optimade(parsed) {
-    const filter = [];
-
-    ['formulae', 'elements', 'props', 'classes'].forEach(function (categ) {
-      if (!parsed[categ]) return;
-      else if (categ === 'formulae') {
-        if (is_formula_anonymous(parsed[categ]))
-          filter.push(`chemical_formula_anonymous="${parsed[categ]}"`);
-        else
-          filter.push(
-            `chemical_formula_reduced="${sort_formula(parsed[categ])}"`
-          );
-      } else if (categ === 'elements') {
-        filter.push(
-          `elements HAS ALL "${parsed[categ].split('-').join('","')}"`
-        );
-      } else if (categ === 'props') {
-        filter.push(`_mpds_${parsed[categ].replaceAll(' ', '_')} IS KNOWN`);
-      } else if (categ === 'classes') {
-        parsed[categ].split(', ').forEach(function (item) {
-          const arity = arity_keys.indexOf(item);
+	}
+
+	/*
+	 * Chemical formulae parsing, courtesy of Nathan Leung, MIT license
+	 */
+	function countInGroup(tokens, elements, coefficient = 1) {
+		if (tokens.length === 0) {
+			return countElements(tokens, elements);
+		}
+
+		if (tokens[0].type !== 'element') {
+			// If this happens this is likely a mistake, pass back to countElements
+			return countElements(tokens, elements);
+		}
+
+		let i = 0;
+		while (i < tokens.length) {
+			if (tokens[i].type === 'element') {
+				if (tokens.length > (i + 1) && tokens[i + 1].type === 'subscript') {
+					elements[tokens[i].value] += (tokens[i + 1].value * coefficient);
+					i += 2;
+					continue;
+				}
+				elements[tokens[i].value] += coefficient;
+				i++;
+				continue;
+			}
+			break;
+		}
+		return countElements(tokens.slice(i), elements);
+	}
+
+	/*
+	 * Sorting the formulae alphabetically for Optimade
+	 */
+	function sort_formula(text) {
+		let parsed = false;
+		try {
+			parsed = countElements(tokenizeInitial(text));
+		} catch (err) {
+			//console.error(err);
+			return text;
+		}
+		let sorted = '',
+			els = Object.keys(parsed);
+		els.sort();
+		els.forEach(function (el) {
+			sorted += el + (parsed[el] === 1 ? '' : parsed[el]);
+		});
+		return sorted;
+	}
+
+	/*
+	 * Convert MPDS search query object notation into the Optimade filter
+	 */
+	function to_optimade(parsed) {
+		const filter = [];
+
+		['formulae', 'elements', 'props', 'classes'].forEach(function (categ) {
+			if (!parsed[categ]) return;
+			else if (categ === 'formulae') {
+				if (is_formula_anonymous(parsed[categ]))
+					filter.push(`chemical_formula_anonymous="${parsed[categ]}"`);
+				else filter.push(`chemical_formula_reduced="${sort_formula(parsed[categ])}"`);
+			} else if (categ === 'elements') {
+				filter.push(`elements HAS ALL "${parsed[categ].split('-').join('","')}"`);
+			} else if (categ === 'props') {
+				filter.push(`_mpds_${parsed[categ].replaceAll(' ', '_')} IS KNOWN`);
+			} else if (categ === 'classes') {
+				parsed[categ].split(', ').forEach(function (item) {
+					const arity = arity_keys.indexOf(item);
           if (arity > 0) {
             filter.push(`nelements=${arity}`);
           } else if (mpds_classes.includes(item)) {
             filter.push(`_mpds_classes HAS ALL "${item}"`);
           }
-        });
-      }
-    });
+				});
+			}
+		});
 
     if (parsed.numeric && Array.isArray(parsed.numeric)) {
       parsed.numeric.forEach(function (numfilt) {
@@ -2297,38 +2206,35 @@ const OptimadeNLP = function () {
       });
     }
 
-    return filter.join(' AND ');
-  }
-
-  /*
-   * API
-   */
-  return {
-    guess,
-    to_optimade,
-
-    is_formula_anonymous,
-    sort_formula,
-    parse_aeatoms,
-    termify_formulae,
-    is_numeric,
-
-    arity_keys,
-    periodic_elements,
-    periodic_elements_cased,
-    periodic_element_names,
-  };
+		return filter.join(' AND ');
+	}
+
+	/*
+	 * API
+	 */
+	return {
+		guess,
+		to_optimade,
+
+		is_formula_anonymous,
+		sort_formula,
+		parse_aeatoms,
+		termify_formulae,
+		is_numeric,
+
+		arity_keys,
+		periodic_elements,
+		periodic_elements_cased,
+		periodic_element_names
+	};
 };
 
 if (typeof module !== 'undefined' && module.exports) {
-  module.exports = OptimadeNLP;
-} else if (
-  typeof require === 'function' &&
-  typeof require.specified === 'function'
-) {
-  define(function () {
-    return OptimadeNLP;
-  });
+	module.exports = OptimadeNLP;
+} else if (typeof require === 'function' && typeof require.specified === 'function') {
+	define(function () {
+		return OptimadeNLP;
+	});
 } else if (window !== undefined) {
-  window.OptimadeNLP = OptimadeNLP;
+	window.OptimadeNLP = OptimadeNLP
 }
diff --git a/test_nlp.json b/test_nlp.json
index bd61327..b5aa45a 100755
--- a/test_nlp.json
+++ b/test_nlp.json
@@ -1,280 +1,217 @@
 [
-  ["tio2", { "formulae": "tio2" }],
-  ["al2o3", { "formulae": "al2o3" }],
-  ["band gap, chlorides", { "props": "band gap", "classes": "chloride" }],
-  ["ti o", { "elements": "Ti-O" }],
-  ["tio2, band gap", { "formulae": "tio2", "props": "band gap" }],
-  ["organic pb i", { "classes": "organic", "elements": "Pb-I" }],
-  [
-    "organic pb-i perovskite",
-    { "elements": "Pb-I", "classes": "organic, perovskite" }
-  ],
-  ["ternary oxide", { "classes": "ternary, oxide" }],
-  ["zeolite ite", { "classes": "zeolite" }],
-  ["CuLaMnSeO", { "formulae": "CuLaMnSeO" }],
-  ["AgGaTeSe", { "formulae": "AgGaTeSe" }],
-  ["MnOCaSeRu", { "formulae": "MnOCaSeRu" }],
-  ["LaZnOAsAm", { "formulae": "LaZnOAsAm" }],
-  ["MnCaOSeRu", { "formulae": "MnCaOSeRu" }],
-  ["MnCaSeBRu", { "formulae": "MnCaSeBRu" }],
-  ["MnCaSeRu", { "formulae": "MnCaSeRu" }],
-  ["BCNOAc", { "formulae": "BCNOAc" }],
-  ["LaZnOAs", { "formulae": "LaZnOAs" }],
-  ["LaZnAsO", { "formulae": "LaZnAsO" }],
-  ["LaOZnAs", { "formulae": "LaOZnAs" }],
-  ["OLaZnAs", { "formulae": "OLaZnAs" }],
-  ["OLaZnY", { "formulae": "OLaZnY" }],
-  ["OLaZnYB", { "formulae": "OLaZnYB" }],
-  ["OLaZnYBr", { "formulae": "OLaZnYBr" }],
-  ["OLaSZnAs", { "formulae": "OLaSZnAs" }],
-  ["OLaOZnAs", { "formulae": "OLaOZnAs" }],
-  ["OLaZnPAs", { "formulae": "OLaZnPAs" }],
-  ["OLaKZnPAs", { "formulae": "OLaKZnPAs" }],
-  ["SYPMnBr", { "formulae": "SYPMnBr" }],
-  ["SYMnPBr", { "formulae": "SYMnPBr" }],
-  ["RbMnTaO", { "formulae": "RbMnTaO" }],
-  ["RoMnTaO", {}],
-  ["CPZnOY", { "formulae": "CPZnOY" }],
-  ["CPOZnY", { "formulae": "CPOZnY" }],
-  ["BaLaMnRuO6", { "formulae": "BaLaMnRuO6" }],
-  ["phase diagrams", { "props": "phase diagram" }],
-  [
-    "thermodynamics superconductor fe",
-    { "props": "thermodynamics", "classes": "superconductor", "elements": "Fe" }
-  ],
-  ["cub", { "lattices": "cubic" }],
-  [
-    "tetragonal srtio3 ternary",
-    { "lattices": "tetragonal", "formulae": "srtio3", "classes": "ternary" }
-  ],
-  [
-    "SrTiO&#x2083; tet phonons",
-    { "lattices": "tetragonal", "formulae": "SrTiO3", "props": "phonons" }
-  ],
-  ["O3Al2 elastic properties", { "formulae": "O3Al2" }],
-  [
-    "optical properties, LiKSO4",
-    { "props": "optical properties", "formulae": "LiKSO4" }
-  ],
-  [
-    "band gap ZnO hex",
-    { "props": "band gap", "lattices": "hexagonal", "formulae": "ZnO" }
-  ],
-  ["geo", {}],
-  ["GeO", { "formulae": "GeO" }],
-  ["beo", {}],
-  ["BeO", { "formulae": "BeO" }],
-  ["GdNiIn conductivity", { "formulae": "GdNiIn", "props": "conductivity" }],
-  [
-    "indium binary oxide metal",
-    { "elements": "In", "classes": "binary, oxide, metal" }
-  ],
-  [
-    "Pd hydride lattice",
-    { "elements": "Pd", "classes": "hydride", "props": "lattice" }
-  ],
-  ["Mn halogens Be", { "elements": "Mn-Be", "classes": "halogen" }],
-  [
-    "carbide semiconductor electronic properties",
-    { "classes": "carbide, semiconductor", "props": "electronic properties" }
-  ],
-  [
-    "hexagonal carbide semiconductor electronic properties",
-    {
-      "lattices": "hexagonal",
-      "classes": "carbide, semiconductor",
-      "props": "electronic properties"
-    }
-  ],
-  [
-    "rare earth chalcogens organics magnetism",
-    { "classes": "rare earth, chalcogen, organic", "props": "magnetism" }
-  ],
-  [
-    "Te metals transitional Sn",
-    { "elements": "Te-Sn", "classes": "metal, transitional" }
-  ],
-  [
-    "radioactive nonmetals conductivity",
-    { "classes": "radioactive, nonmetal", "props": "conductivity" }
-  ],
-  [
-    "noble gases superconductivity",
-    { "classes": "noble gas", "props": "superconductivity" }
-  ],
-  [
-    "W-Mo ternary, phase diagram",
-    { "elements": "W-Mo", "classes": "ternary", "props": "phase diagram" }
-  ],
-  [
-    "quaternary perovskites, thermodynamics",
-    { "classes": "quaternary, perovskite", "props": "thermodynamics" }
-  ],
-  ["lanthanoid element", { "classes": "lanthanoid, unary" }],
-  [
-    "lanthanoid transitional electronic properties",
-    { "classes": "lanthanoid, transitional", "props": "electronic properties" }
-  ],
-  [
-    "lattice of intermetallic binary cubic",
-    {
-      "props": "lattice",
-      "classes": "intermetallic, binary",
-      "lattices": "cubic"
-    }
-  ],
-  ["oxygen", { "elements": "O" }],
-  [
-    "Y oxide lattice ternary",
-    { "elements": "Y", "classes": "oxide, ternary", "props": "lattice" }
-  ],
-  [
-    "phases for ternary ferromagnet",
-    { "props": "phases", "classes": "ternary, ferromagnet" }
-  ],
-  [
-    "superconductivity superconductors C",
-    {
-      "props": "superconductivity",
-      "classes": "superconductor",
-      "elements": "C"
-    }
-  ],
-  [
-    "magnetism, ferromagnet, chalcogen, monocl",
-    {
-      "props": "magnetism",
-      "classes": "ferromagnet, chalcogen",
-      "lattices": "monoclinic"
-    }
-  ],
-  ["Rg", { "elements": "Rg" }],
-  ["ErSi", { "formulae": "ErSi" }],
-  ["ErSI", { "formulae": "ErSI" }],
-  ["Ca3(PO4)2", { "formulae": "Ca3(PO4)2" }],
-  ["Ca3[PO4]2", { "formulae": "Ca3[PO4]2" }],
-  ["conductors", { "classes": "conductor" }],
-  ["chromium, fe, tricl", { "elements": "Cr-Fe", "lattices": "triclinic" }],
-  ["astatine, cell", { "elements": "At" }],
-  ["Gruneisen coefficient", { "props": "gruneisen coefficients" }],
-  ["Seebeck coefficient", { "props": "seebeck coefficient" }],
-  ["figure of merit", { "props": "figure of merit" }],
-  ["conductivity", { "props": "conductivity" }],
-  [
-    "crystalline structure of binaries",
-    { "props": "crystalline structure", "classes": "binary" }
-  ],
-  [
-    "crystal cell for binary compound",
-    { "props": "crystal cell", "classes": "binary" }
-  ],
-  ["valence", { "props": "valence" }],
-  ["conductivity", { "props": "conductivity" }],
-  [
-    "La magnetic phase diagram",
-    { "elements": "La", "props": "magnetic phase diagram" }
-  ],
-  ["info on phase diagrams plots", { "props": "phase diagram plots" }],
-  ["optical conductivity", { "props": "optical conductivity" }],
-  ["cell and atoms", { "classes": "cell and atoms" }],
-  ["conductor superconductor", { "classes": "conductor, superconductor" }],
-  ["metal nonmetal", { "classes": "metal, nonmetal" }],
-  ["metal or nonmetal character", { "props": "metal or nonmetal character" }],
-  ["electrical resistance values", { "props": "electrical properties" }],
-  ["temperature derivative by pressure", { "props": "pressure" }],
-  ["temperature derivative by composition", {}],
-  ["optical phonon frequency", { "props": "optical properties" }],
-  ["enthalpy conductivity", { "props": "values of enthalpy" }],
-  ["decomposis of the magneto", {}],
-  ["TiO2 SrTiO3", { "formulae": "TiO2" }],
-  ["average number of 4f electrons", {}],
-  ["225aaa!", {}],
-  ["he-he-he", { "elements": "He-He-He" }],
-  [
-    "electronic contribution to thermal conductivity",
-    { "props": "electronic contribution to thermal conductivity" }
-  ],
-  [
-    "electronic contribution to superconducting heat capacity",
-    { "props": "electronic contribution to superconducting heat capacity" }
-  ],
-  ["residual resistivity", { "props": "residual resistivity ratio" }],
-  [
-    "optical absorption no",
-    { "props": "optical absorption coefficient", "elements": "No" }
-  ],
-  ["energy of optical phonon", { "props": "energy of optical phonon" }],
-  ["luminescence", { "props": "values of luminescence" }],
-  ["values of luminescence", { "props": "values of luminescence" }],
-  ["magnetization", { "props": "magnetization" }],
-  ["values of magnetization", { "props": "values of magnetization" }],
-  [
-    "energy of longitudinal optical phonon",
-    { "props": "energy of longitudinal optical phonon" }
-  ],
-  [
-    "permittivity (dielectric constant) of perovskites ",
-    { "props": "permittivity", "classes": "perovskite" }
-  ],
-  ["density > 1", { "numeric": [["density", ">", 1]], "phased": true }],
-  ["InFO", { "formulae": "InFO" }],
-  ["info", {}],
-  ["Goodbye.", {}],
-  ["iiii", {}],
-  ["...", {}],
-  ["What'd I do?", { "elements": "I" }],
-  ["OK, let's go", { "formulae": "OK" }],
-  ["Ginny Danburry's here.  Look for her", {}],
-  ["I think u make me laughing.", { "elements": "I-U" }],
-  ["physical properties", { "props": "physical properties" }],
-  ["physical property", { "props": "physical properties" }],
-  ["electric property", { "props": "electric properties" }],
-  ["ab initio", {}],
-  ["ab initio calculations,", { "classes": "ab initio calculations" }],
-  ["ab initio literature,", { "classes": "ab initio literature" }],
-  [
-    "ab initio literature, ab initio calculations,",
-    { "classes": "ab initio literature, ab initio calculations" }
-  ],
-  [
-    "ab initio calculations P K",
-    { "elements": "P-K", "classes": "ab initio calculations" }
-  ],
-  [
-    "ab initio literature P K",
-    { "elements": "P-K", "classes": "ab initio literature" }
-  ],
-  [
-    "ab initio, ab initio calculations, P K",
-    { "elements": "P-K", "classes": "ab initio calculations" }
-  ],
-  [
-    "ab initio literature, ab initio calculations, P K",
-    {
-      "elements": "P-K",
-      "classes": "ab initio literature, ab initio calculations"
-    }
-  ],
-  ["A1B2C3", { "formulae": "A1B2C3" }],
-  ["AB", { "formulae": "AB" }],
-  ["ab datum", {}],
-  [
-    "AB ab initio calculations",
-    { "formulae": "AB", "classes": "ab initio calculations" }
-  ],
-  ["ABC7D9", { "formulae": "ABC7D9" }],
-  ["ABCD", { "formulae": "ABCD" }],
-  ["ABCDE", {}],
-  ["filter=author=\"Sąžininga Žąsis\"", {}],
-  [
-    "(elements HAS \"Ac\" AND nelements=1) OR (elements HAS \"Ac\" AND nsites=1)",
-    {}
-  ],
-  ["nelements=1", {}],
-  ["elements HAS \"Ti\"", {}],
-  ["elements HAS ALL \"C\",\"N\",\"O\",\"H\"", {}],
-  ["elements HAS \"Ti\" AND nelements>3", {}],
-  ["chemical_formula_reduced=\"Li7Sn2\"", {}],
+	["tio2", { "formulae": "tio2" }],
+	["al2o3", { "formulae": "al2o3" }],
+	["band gap, chlorides", { "props": "band gap", "classes": "chloride" }],
+	["ti o", { "elements": "Ti-O" }],
+	["tio2, band gap", { "formulae": "tio2", "props": "band gap" }],
+	["organic pb i", { "classes": "organic", "elements": "Pb-I" }],
+	["organic pb-i perovskite", { "elements": "Pb-I", "classes": "organic, perovskite" }],
+	["ternary oxide", { "classes": "ternary, oxide" }],
+	["zeolite ite", { "classes": "zeolite" }],
+	["CuLaMnSeO", { "formulae": "CuLaMnSeO" }],
+	["AgGaTeSe", { "formulae": "AgGaTeSe" }],
+	["MnOCaSeRu", { "formulae": "MnOCaSeRu" }],
+	["LaZnOAsAm", { "formulae": "LaZnOAsAm" }],
+	["MnCaOSeRu", { "formulae": "MnCaOSeRu" }],
+	["MnCaSeBRu", { "formulae": "MnCaSeBRu" }],
+	["MnCaSeRu", { "formulae": "MnCaSeRu" }],
+	["BCNOAc", { "formulae": "BCNOAc" }],
+	["LaZnOAs", { "formulae": "LaZnOAs" }],
+	["LaZnAsO", { "formulae": "LaZnAsO" }],
+	["LaOZnAs", { "formulae": "LaOZnAs" }],
+	["OLaZnAs", { "formulae": "OLaZnAs" }],
+	["OLaZnY", { "formulae": "OLaZnY" }],
+	["OLaZnYB", { "formulae": "OLaZnYB" }],
+	["OLaZnYBr", { "formulae": "OLaZnYBr" }],
+	["OLaSZnAs", { "formulae": "OLaSZnAs" }],
+	["OLaOZnAs", { "formulae": "OLaOZnAs" }],
+	["OLaZnPAs", { "formulae": "OLaZnPAs" }],
+	["OLaKZnPAs", { "formulae": "OLaKZnPAs" }],
+	["SYPMnBr", { "formulae": "SYPMnBr" }],
+	["SYMnPBr", { "formulae": "SYMnPBr" }],
+	["RbMnTaO", { "formulae": "RbMnTaO" }],
+	["RoMnTaO", {}],
+	["CPZnOY", { "formulae": "CPZnOY" }],
+	["CPOZnY", { "formulae": "CPOZnY" }],
+	["BaLaMnRuO6", { "formulae": "BaLaMnRuO6" }],
+	["phase diagrams", { "props": "phase diagram" }],
+	[
+		"thermodynamics superconductor fe",
+		{ "props": "thermodynamics", "classes": "superconductor", "elements": "Fe" }
+	],
+	["cub", { "lattices": "cubic" }],
+	[
+		"tetragonal srtio3 ternary",
+		{ "lattices": "tetragonal", "formulae": "srtio3", "classes": "ternary" }
+	],
+	[
+		"SrTiO&#x2083; tet phonons",
+		{ "lattices": "tetragonal", "formulae": "SrTiO3", "props": "phonons" }
+	],
+	["O3Al2 elastic properties", { "formulae": "O3Al2" }],
+	["optical properties, LiKSO4", { "props": "optical properties", "formulae": "LiKSO4" }],
+	["band gap ZnO hex", { "props": "band gap", "lattices": "hexagonal", "formulae": "ZnO" }],
+	["geo", {}],
+	["GeO", { "formulae": "GeO" }],
+	["beo", {}],
+	["BeO", { "formulae": "BeO" }],
+	["GdNiIn conductivity", { "formulae": "GdNiIn", "props": "conductivity" }],
+	["indium binary oxide metal", { "elements": "In", "classes": "binary, oxide, metal" }],
+	["Pd hydride lattice", { "elements": "Pd", "classes": "hydride", "props": "lattice" }],
+	["Mn halogens Be", { "elements": "Mn-Be", "classes": "halogen" }],
+	[
+		"carbide semiconductor electronic properties",
+		{ "classes": "carbide, semiconductor", "props": "electronic properties" }
+	],
+	[
+		"hexagonal carbide semiconductor electronic properties",
+		{
+			"lattices": "hexagonal",
+			"classes": "carbide, semiconductor",
+			"props": "electronic properties"
+		}
+	],
+	[
+		"rare earth chalcogens organics magnetism",
+		{ "classes": "rare earth, chalcogen, organic", "props": "magnetism" }
+	],
+	["Te metals transitional Sn", { "elements": "Te-Sn", "classes": "metal, transitional" }],
+	[
+		"radioactive nonmetals conductivity",
+		{ "classes": "radioactive, nonmetal", "props": "conductivity" }
+	],
+	["noble gases superconductivity", { "classes": "noble gas", "props": "superconductivity" }],
+	[
+		"W-Mo ternary, phase diagram",
+		{ "elements": "W-Mo", "classes": "ternary", "props": "phase diagram" }
+	],
+	[
+		"quaternary perovskites, thermodynamics",
+		{ "classes": "quaternary, perovskite", "props": "thermodynamics" }
+	],
+	["lanthanoid element", { "classes": "lanthanoid, unary" }],
+	[
+		"lanthanoid transitional electronic properties",
+		{ "classes": "lanthanoid, transitional", "props": "electronic properties" }
+	],
+	[
+		"lattice of intermetallic binary cubic",
+		{ "props": "lattice", "classes": "intermetallic, binary", "lattices": "cubic" }
+	],
+	["oxygen", { "elements": "O" }],
+	[
+		"Y oxide lattice ternary",
+		{ "elements": "Y", "classes": "oxide, ternary", "props": "lattice" }
+	],
+	["phases for ternary ferromagnet", { "props": "phases", "classes": "ternary, ferromagnet" }],
+	[
+		"superconductivity superconductors C",
+		{ "props": "superconductivity", "classes": "superconductor", "elements": "C" }
+	],
+	[
+		"magnetism, ferromagnet, chalcogen, monocl",
+		{ "props": "magnetism", "classes": "ferromagnet, chalcogen", "lattices": "monoclinic" }
+	],
+	["Rg", { "elements": "Rg" }],
+	["ErSi", { "formulae": "ErSi" }],
+	["ErSI", { "formulae": "ErSI" }],
+	["Ca3(PO4)2", { "formulae": "Ca3(PO4)2" }],
+	["Ca3[PO4]2", { "formulae": "Ca3[PO4]2" }],
+	["conductors", { "classes": "conductor" }],
+	["chromium, fe, tricl", { "elements": "Cr-Fe", "lattices": "triclinic" }],
+	["astatine, cell", { "elements": "At" }],
+	["Gruneisen coefficient", { "props": "gruneisen coefficients" }],
+	["Seebeck coefficient", { "props": "seebeck coefficient" }],
+	["figure of merit", { "props": "figure of merit" }],
+	["conductivity", { "props": "conductivity" }],
+	[
+		"crystalline structure of binaries",
+		{ "props": "crystalline structure", "classes": "binary" }
+	],
+	["crystal cell for binary compound", { "props": "crystal cell", "classes": "binary" }],
+	["valence", { "props": "valence" }],
+	["conductivity", { "props": "conductivity" }],
+	["La magnetic phase diagram", { "elements": "La", "props": "magnetic phase diagram" }],
+	["info on phase diagrams plots", { "props": "phase diagram plots" }],
+	["optical conductivity", { "props": "optical conductivity" }],
+	["cell and atoms", { "classes": "cell and atoms" }],
+	["conductor superconductor", { "classes": "conductor, superconductor" }],
+	["metal nonmetal", { "classes": "metal, nonmetal" }],
+	["metal or nonmetal character", { "props": "metal or nonmetal character" }],
+	["electrical resistance values", { "props": "electrical properties" }],
+	["temperature derivative by pressure", { "props": "pressure" }],
+	["temperature derivative by composition", {}],
+	["optical phonon frequency", { "props": "optical properties" }],
+	["enthalpy conductivity", { "props": "values of enthalpy" }],
+	["decomposis of the magneto", {}],
+	["TiO2 SrTiO3", { "formulae": "TiO2" }],
+	["average number of 4f electrons", {}],
+	["225aaa!", {}],
+	["he-he-he", { "elements": "He-He-He" }],
+	[
+		"electronic contribution to thermal conductivity",
+		{ "props": "electronic contribution to thermal conductivity" }
+	],
+	[
+		"electronic contribution to superconducting heat capacity",
+		{ "props": "electronic contribution to superconducting heat capacity" }
+	],
+	["residual resistivity", { "props": "residual resistivity ratio" }],
+	["optical absorption no", { "props": "optical absorption coefficient", "elements": "No" }],
+	["energy of optical phonon", { "props": "energy of optical phonon" }],
+	["luminescence", { "props": "values of luminescence" }],
+	["values of luminescence", { "props": "values of luminescence" }],
+	["magnetization", { "props": "magnetization" }],
+	["values of magnetization", { "props": "values of magnetization" }],
+	["energy of longitudinal optical phonon", { "props": "energy of longitudinal optical phonon" }],
+	[
+		"permittivity (dielectric constant) of perovskites ",
+		{ "props": "permittivity", "classes": "perovskite" }
+	],
+	["density > 1", { "numeric": [["density", ">", 1]], "phased": true }],
+	["InFO", { "formulae": "InFO" }],
+	["info", {}],
+	["Goodbye.", {}],
+	["iiii", {}],
+	["...", {}],
+	["What'd I do?", { "elements": "I" }],
+	["OK, let's go", { "formulae": "OK" }],
+	["Ginny Danburry's here.  Look for her", {}],
+	["I think u make me laughing.", { "elements": "I-U" }],
+	["physical properties", { "props": "physical properties" }],
+	["physical property", { "props": "physical properties" }],
+	["electric property", { "props": "electric properties" }],
+	["ab initio", {}],
+	["ab initio calculations,", { "classes": "ab initio calculations" }],
+	["ab initio literature,", { "classes": "ab initio literature" }],
+	[
+		"ab initio literature, ab initio calculations,",
+		{ "classes": "ab initio literature, ab initio calculations" }
+	],
+	["ab initio calculations P K", { "elements": "P-K", "classes": "ab initio calculations" }],
+	["ab initio literature P K", { "elements": "P-K", "classes": "ab initio literature" }],
+	[
+		"ab initio, ab initio calculations, P K",
+		{ "elements": "P-K", "classes": "ab initio calculations" }
+	],
+	[
+		"ab initio literature, ab initio calculations, P K",
+		{ "elements": "P-K", "classes": "ab initio literature, ab initio calculations" }
+	],
+	["A1B2C3", { "formulae": "A1B2C3" }],
+	["AB", { "formulae": "AB" }],
+	["ab datum", {}],
+	["AB ab initio calculations", { "formulae": "AB", "classes": "ab initio calculations" }],
+	["ABC7D9", { "formulae": "ABC7D9" }],
+	["ABCD", { "formulae": "ABCD" }],
+	["ABCDE", {}],
+	["filter=author=\"Sąžininga Žąsis\"", {}],
+	["(elements HAS \"Ac\" AND nelements=1) OR (elements HAS \"Ac\" AND nsites=1)", {}],
+	["nelements=1", {}],
+	["elements HAS \"Ti\"", {}],
+	["elements HAS ALL \"C\",\"N\",\"O\",\"H\"", {}],
+	["elements HAS \"Ti\" AND nelements>3", {}],
+	["chemical_formula_reduced=\"Li7Sn2\"", {}],
   ["chemical_formula_anonymous=\"ABC\"", {}],
   ["elements HAS \"C\" AND elements HAS \"N\" AND elements HAS \"O\"", {}],
   ["elements HAS ANY \"C\",\"N\",\"O\"", {}],