v4.2.1

• Require a BioPortal API key to work with the BioPortal mapper
• Fix issue where empty data frame was filtered by mapping score (closes #61)

v4.2.0

• Updated dependencies (closes #56)
• Minor refactorings

v4.1.4

Minor improvements and bug fixes:

• Update the Command Line Interface to reflect recent changes
• Parameterize the ngram length parameter in the TFIDFMapper
• Ensure inputs to text2term are strings (closes #49)
• Add timers to post-mapping filters for debugging

Note: skipped v4.1.3 due to inability to re-publish a deleted release on PyPI.

v4.1.2

Updated version number and .gitignore for local files

v4.1.1

Fixes several bugs, including:

• Adding tags to a term with no initial tags no longer throws an error.
• Caching is now more effective, as previously the cache would recompile when it was unnecessary
• Increased efficiency of filtering terms

Also adds a preliminary GitHub Action for testing, but it is not currently functional yet.

v4.0.0

Change log

User-facing changes:

• Expose OntologyTermCollector module so that it can be used on its own. For example:

  import text2term
  term_collector = OntologyTermCollector(ontology_iri="EFO")
  efo_terms = term_collector.get_ontology_terms(exclude_deprecated=True)

• Modify OntologyTerm to have a dictionary of complex OWL restrictions of the form :A :contains some/only :C, where the dictionary key would be :contains and the value would be :C, which is either a class or another OWL restriction. In the latter case, we store a string representation of the restriction as the value, e.g., {':contains': 'D and E'}.

  pancreatitis = efo_terms["http://www.ebi.ac.uk/efo/EFO_0000278"]
  
  print(pancreatitis.restrictions)
  >> {'http://www.ebi.ac.uk/efo/EFO_0000784': 'http://purl.obolibrary.org/obo/UBERON_0001264'}
  
  print(efo_terms["http://www.ebi.ac.uk/efo/EFO_0000784"].labels.pop())
  >> has_disease_location
  
  print(efo_terms["http://purl.obolibrary.org/obo/UBERON_0001264"].labels.pop())
  >> pancreas

Developer-facing changes:

• Modify OntologyTermCollector to parse OWL restrictions (using owlready2) and to create a dictionary-based (simplified) representation of the restrictions.
• Modify TermCollector to be able to reuse previously loaded ontology to apply different filters without having to reload.
• Rename tagged_terms to tagged_term and fix some references.
• Remove no longer needed mapping functions in onto_cache.py.
• Create new enumeration OntologyTermType to represent the different types of terms supported: Class, Property, Any.
• Use string enumerations for mapper and term_type throughout.
• Rename variable used both for term type and term type filter.
• Update ontologies with latest releases. Add FoodOn ontology.
• Update setuptools dependency version.
• Fix min_score filter issue.
• Fix issue accessing df to include unmapped terms when mappings df is empty.

Documentation / Logging

• Adds Sphinx Docs to be published in readthedocs.
• Add warning about mapping score when using BioPortal mapper.
• Add documentation about parameters in text2term.map_terms().
• Update documentation of source_terms parameter in map_terms() to clarify the different possibilities currently supported.
• Add logging to t2t module and move some logging from individual mappers to t2t.

Testing

• Turn simple_tests.py into a more modular test suite.
• Add some structure around all tests and documentation about them.
• Add tests for Zooma, Bioportal, TFIDF and syntactic mappers.
• Add min_score filter tests.
• Add include_unmapped feature tests.
• Add OntologyTermCollector tests.

Full Changelog: v3.0.2...v4.0.0

v3.0.2

Updates various dependencies and makes them less strict by allowing compatible versions.
Fixes issue in command-line interface imports and mapping function, due to recent refactoring.

v3.0.1

Fixes a bug that prevents adding tags when the current tag variable is just None

v3.0.0

Removes 'map_file' and 'map_tagged_terms' functions in favor of 'map_terms'. Also adds functionality to ignore certain terms via tagging, and to include unmapped terms in the results.

v2.3.2

Fixes a bug where the min_score argument in "map" was ignored when not using the TFIDF mapper.