increase pronominal coverage

Author: Tyler Kniess

Makefile

@@ -1,13 +1,14 @@
 all:
-	cat os.lexc > os-copy.lexc
+	cat os.lexc > backups/os.lexc.copy
+	cat os.twol > backups/os.twol.copy
 	hfst-lexc -v os.lexc -o os.lexc.hfst
 	hfst-twolc os.twol -o os.twol.hfst
 	hfst-compose-intersect -v -o os.gen.hfst os.lexc.hfst os.twol.hfst
 	hfst-invert -v os.gen.hfst -o os.mor.hfst	
 	hfst-fst2strings os.mor.hfst --xfst obey-flags | sort | gsed s/':'/'\t'/g|gsed s/'<'/' <'/g | gsed s/'>'/'> '/g > forms.txt
 	python3 gi-prefix.py
 	cat os.lexc | sh visualize_transducer.sh
-	cat forms-prefixed.txt>forms.txt
+	cat forms-prefixed.txt > forms.txt
 	python3 coverage.py
 	cat heliand/clean.txt|while read line; do hfst-lookup os.mor.hfst; done > lookup.txt
 	echo 'The top 20 words yet to be included are:'

OldSaxonLexc.png

@@ -8,7 +8,7 @@
 covered_total_words = []
 words_not_covered = []
 
-heliand = 'heliand/hel-unique-tokens-clean.txt'
+heliand = 'heliand/hel-unique-tokens.txt'
 heliand_total = 'heliand/clean.txt'
 muspilli = 'forms-prefixed.txt'
 

coverage.py

@@ -1,8 +1,8 @@
-The Hêliand contains a total of 46433 forms, 7965 of which are unique.
+The Hêliand contains a total of 46433 forms, 7506 of which are unique.
 
-Muspilli currently contains 429540 synthetically-generated forms.
+Muspilli currently contains 429545 synthetically-generated forms.
 
-Of the 7965 unique tokens in the Hêliand, Muspilli contains 3182 forms, or 39.95%
+Of the 7506 unique tokens in the Hêliand, Muspilli contains 3324 forms, or 44.285%
 
-Of the 46433 total tokens in the Hêliand, Muspilli contains 37157 forms, or 80.023%
+Of the 46433 total tokens in the Hêliand, Muspilli contains 37359 forms, or 80.458%