0.19.0

Merged in release/0.19.0 (pull request #183)

Release/0.19.0

Approved-by: Max Press

0.10.0

[0.10.0] 2021-09-17

Changed

• CDSInterval object now has methods to access the number of codons and codon locations in both chunk-relative and chromosome coordinates. Chromosome accessors will always return the full original CDS.
• If duplicate sequence identifiers are found when parsing GenBank/FASTA files, an exception is raised.
• The scan_codon_locations methods on CDSInterval now operate on two algorithms, one simpler algorithm for canonical transcripts (no programmed frameshifts and no offset frames) and the original more robust algorithm otherwise.

0.9.0

[0.9.0] 2021-09-03

Fixed

• LocusTag parser is now able to handle CDS-only features, and have more informative error reporting for locus tag collisions.
• GenBank parser no longer raises an exception for multi-stranded feature; instead it warns and moves on

0.8.0

[0.8.0]

Fixed

• Do not trust feature type annotations to define coding vs. non-coding when parsing GenBank files; only rely on the presence/absence of CDS intervals associated with the transcript.
• Setup requirements, build tests, and sphinx config updated to allow building of documentation without installing the package.
• GenBank parser was not handling exon features as direct descendants of gene correctly.

Changed

• Sorted parser now sorts features by position, then gene/mRNA/CDS/other. This helps deal with genbank files that are oddly ordered.
• Introduced new Hybrid GenBank parser mode that does both LocusTag and Sorted parsing at the same time.
• All .chromosome_location accessor of Interval objects always return the full length Location, even if the Interval itself is chunk relative such that the underlying Location object cannot represent the full length. As a result of this, the .chromosome_location of a chunk-relative location cannot have associated sequence information.

Added

• TranscriptInterval and FeatureInterval now have accessor methods to get Location objects for their introns/gaps and full span.

0.7.0

[0.7.0]

Changed

• GenBank position-sorted parser can now handle CDS records that are not directly following a gene record.
• Refactor Location, Parent and Sequence to have base classes AbstractLocation, AbstractParent and AbstractSequence that are in the base of the inscripta.biocantor.location module. This greatly helps with resolving circular imports.
• Optimized checking sequence and location members to explicitly check for None. This avoids a call to __len__.
• CompoundInterval._single_intervals is now lazily evaluated, because it is expensive to generate many SingleInterval objects.
• CompoundInterval now stores the positions as two sorted integer lists.
• CompoundInterval constructor accepts tuples in addition to lists of integer values to avoid list construction overhead.
• CompoundInterval.is_overlapping and CompoundInterval.is_contiguous are lazily evaluated.
• CompoundInterval._combine_blocks now always removes empty blocks. The new implementation also avoids producing a new interval if the result is identical to the start.
• unique_value_or_none was pulled out of Parent into its own separate function with an associated cache. This function was optimized to use sets.
• Added __slots__ to all child classes of AbstractLocation, AbstractSequence and AbstractParent.
• Removed unnecessary call to strip_location_info() in Sequence constructor.
• Removed all unnecessary instances of constructing lists, replacing them with iterators and tuples.
• GenBank export now defaults to not updating /translation tag in order to save execution time. The original behavior can be restored by setting update_translations=True.

Fixed

• GenBank parser was not properly handling 0bp intervals, which can be sometimes seen as a way to represent insertions.
• GenBank parser was not capturing CDS qualifiers when parsing eukaryotic style GenBank files that have mRNA level features

0.6.0

[0.6.0]

Changed

• Added raise_on_reserved_attributes flag to GFF3 export that controls whether reserved attributes lead to warnings or exceptions.
• Added more top-level imports to simplify imports
• Try more common identifiers when parsing gene symbols from GFF3 files
• Attempt to infer frame from GFF3 files with null Phase columns on CDS records
• Update Tox tests to have a separate formatting case

0.5.0

[0.5.0]

Changed

• Added ability to parse non-transcribed features from GenBank records without a parent /gene record in the position-sorted parser.
• Added ability to export SeqRecord annotations when writing to GenBank.
• Added methods to FeatureInterval that mirror TranscriptInterval.
• Added support for translating with non-standard codon tables.

0.4.5

[0.4.5]

Changed

• Remove contributor license agreement, which is superseded by the MIT license.

[0.4.4]

Fixed

• Handle genbank files with broken intervals gracefully.
• Fix interval parsing for negative strand features.

Changed

• The tag Name can now be used to identify a feature interval in a GFF3/GenBank file.

[0.4.3]

Fixed

• AnnotationCollection._subset_parent() now uses seq_chunk_to_parent and pulls out the chromosome ID from the chromosome record.
• CDSInterval.from_dict() now passes along the parent provided.

Added

• strict_parent_compare parameter for binary set theory operations.
• AnnotationCollection.query_by_position() has a new boolean flag expand_location_to_children that defaults to False, but if set to True
  will expand the interval to contain the transcripts. When False, it may be the case that transcripts will have their underlying location objects
  sliced down from their original coordinates. The original coordinates are still retained as integer members.
  If the query position is entirely intronic for an isoform, this isoform will have a EmptyLocation chunk relative location,
  but will still retain a chromosome_location.

Changed

• Added a parent-level sequence identifier to the output of biocantor.io.parser.seq_chunk_to_parent().
• Added a strand argument to biocantor.io.parser.seq_chunk_to_parent() that allows for the sequence chunk to be strand-referenced.
• Location.parent_to_relative_location and Location.location_relative_to now has a optimize_blocks flag that defaults to True.
  If this flag is False, then these operations will not collapse adjacent or overlapping blocks.

[0.4.2]

Added

• Location.union_preserve_overlaps() function added. This function produces the union of intervals, while preserving all overlaps.

[0.4.1]

Added

• Improved docstrings on interval objects.
• Location objects now have a full_span optional flag on all intersction, overlaps and contains functions. This flag has compound intervals be treated as their full span, i.e. from start to end, regardless of compound structure. This flag defaults to False in all cases. When two CompoundInterval are compared, they are both always compared in their full spans when this flag is True.
• Interval and IntervalCollection objects now are capable of being lifted to arbitrary coordinate systems, returning a new copy. These operations rely on first lifting to a shared chromosomal coordinate system.

Changed

• New SequenceType enum stores whether interval sequences are chromosome or chunk_relative.
• All objects that accept SequenceType information accepts either the SequenceType enum OR raw strings.
• AnnotationCollection will look at the provided parent_or_seq_chunk_parent to see if the bounds of the object can be inferred from the parent object. This is only performed if no start/end are explicitly provided. If neither are provided, the bounds of the collection are the bounds of its children.
• Refactored CDSInterval to be based on AbstractFeatureInterval. Moved CDSPhase and CDSFrame to accomodate the circular import this introduced.
• All Interval objects are allowed to have chromosome parents without sequence information.
• Removed versioneer in favor of hard coded versions.

Fixed

• Some functions on Interval objects were not operating in chromosome coordinates
• AnnotationCollection.query_by_position() was not returning valid results if the parent was a sequence chunk.
• GFF3 parser was not inferring transcripts for a gene feature with no children.
• Fixed a bug with missing gene biotypes in GFF3 parsing.

[0.4.0]

Added

• All Interval objects now have the ability to be built from subsets of genome sequence (called sequence_chunk).
• Querying AnnotationCollection objects by coordinates produce new objects with sliced sequences with chunk-relative coordinates.
• Interval objects built from sequence subsets can be exported in chunk-relative coordinates to GFF3/GenBank.
• Interval objects have new coordinate translation methods that operate in chunk-relative space. Coordinate methods that operate in genomic coordinate space were retained.
• Non-transcribed feature identifier parsing looks in the note special field for identifiers.

Changed

• All Interval objects now must be built directly from coordinates, and do not accept Location objects.
• All Interval objects now hide their Location member. This is to avoid confusion about what coordinate system the Location may be on.
• All interval collections have __iter__ functions that call __iter_children() functions.
• All Interval objects have their core ._location object hidden, and offer two accessors -- .chromosome_location and .chunk_relative_location. Note that .chromosome_location will not have sequence information attached to it if a sequence chunk was used. Generally, it is advised to not access .location objects directly.

[0.3.1]

Fixed

• Feature interval identifier regex should exactly match qualifier keys

Added

• Unified API for identifiers on all interval objects with new property methods .id and .name.

[0.3.0]

Fixed

• Biotype enum improperly mapped protein_coding and protein-coding to different values. Added mRNA as another synonym for this type.
• GFF3, BED and GenBank export from Interval objects now raise an exception when the sequence name field is null.

Added

• Parse FeatureInterval and FeatureIntervalCollection from GFF3 or GenBank, and write back as well.

Changed

• FeatureInterval now has multiple types, stored as sets. FeatureIntervalCollection stores the union of these types, in addition to optionally having its own type.

[0.2.0] 2021-01-06

Fixed

• CompoundInterval.relative_interval_to_parent_location() in the case of overlapping blocks. Had previously been double counting overlap region.
• CompoundInterval.gap_list() in the case of overlapping blocks. Had been raising an error in that case.

Added

• CDSInterval.scan_codon_locations() method. Returns an iterator over codon locations.
• Implement __hash__() for CompoundInterval and CDSInterval
• Implemented data structures for TranscriptInterval and FeatureInterval that model transcribed and non-transcribed genomic features.
• Implemented data structures for GeneInterval and FeatureIntervalCollection, that model groups of intervals as genes or generic feature groups.
• Implemented a wrapper data structure AnnotationCollection that contains groups of genes and non-transcribed feature groups.
• Implemented the ability to build BioCantor gene models from GFF3 and GenBank files.
• Implemented the ability to export the above data structures as GFF3, GenBank, BED and NCBI TBL formats.
• Implemented Marshmallow dataclasses that allow for serialization and deserialization of the above data structures.
• Copied the bins implementation from gffutils to avoid needing the full dependency set in a minimal install.
• Added a Biotype enumeration that tracks known biotypes.
• Added caching of sequence retrieval to Interval objects.

Changed

• Migrated sphinx documentation from automodapi to autoapi.
• Performance upgrades to interval arithmetic operations.

Removed

• CDSInterval.intersect() method. Frame math was incorrect for complex CDSs and was deemed too difficult to implement correctly.

v0.1.1

Merged in hotfix/0.1.1 (pull request #1)

Hotfix/0.1.1

Approved-by: Michael DePalatis
Approved-by: Joshua Shorenstein
Approved-by: Ian Fiddes

v0.1.0

Initial release