Skip to content

Latest commit

 

History

History
164 lines (134 loc) · 5.93 KB

CHANGELOG.md

File metadata and controls

164 lines (134 loc) · 5.93 KB
Raw

Changelog

1.0.1

  • Misc
    • Better error reporting when database files cannot be found (in CLI and API).
    • Minor documentation updates.

1.0.0

  • New features
    • tree command for generating hierarchical clustering trees from distance matrices.
  • General
    • Preferred extensions for genome database files and signatures files have been changed from .db and .h5 to .gdb and .gs.
  • Performance improvements
    • Use process-based parallelism by default for parsing multiple sequence files (much faster).
    • Speed up gambit dist with -s option applied.
  • CLI
    • Strip directory and extension from input file IDs. This applies to CSV output for querying and distance calculation and IDs in generated signature files.
    • -k and --prefix parameters now default to values used RefSeq database.
    • Add option to specify number of cores to use.
    • Add option to disable progress bar printing.

0.5.1

Minor edits to project README and metadata.

0.5.0

  • New features
    • gambit dist command for calculating distance matrices.
  • CLI
    • Sequence file input
      • Explicitly restrict input to FASTA format only.
      • Files may be gzipped.
      • Read input file lists from text files.
    • Minor changes to options of subcommands in signatures group.
  • API
    • gambit.db subpackage:
      • Database-loading funcs moved to class methods of ReferenceDatabase.
      • Additional taxonomy tree methods.
      • Some additional internal reorganization/refactoring.

0.4.0

  • New features
    • Result reporting
      • Results include list of closest reference genomes. This is only reported in JSON-based output formats.
      • New "next_taxon" attribute, indicating the next most specific taxon for which the threshold was not met.
  • CLI
    • signatures info subcommand uses current reference DB by default.
  • Documentation
    • Some improvements to API docs.
  • API and internals
    • calc_signature() function can take multiple sequences as input.
    • Remove calc_signature_parse() function.
    • Refactoring
      • Rename GAMBITDatabase -> ReferenceDatabase, gambit.db.gambitdb -> .refdb
      • Rename gambit.signatures -> gambit.sigs.
      • Merge gambit.sigs.array, gambit.sigs.meta -> gambit.sigs.base
      • Rename gambit.io.export -> gambit.results
      • Move generic sequence code from gambit.kmers to gambit.seq.
      • Merge gambit.io.seq -> gambit.seq.
      • Rename load_database* funcs -> load_db*.
      • Move gambit.io.json -> gambit.util.json, gambit.io.util -> gambit.util.io, remove gambit.io.
      • Moved some other stuff between modules.
    • Improvements to gambit.sigs.hdf5.HDF5Signatures
      • Improvements to .create() method.
      • Support compression.
    • Format-independent functions for reading/writing signature data.
    • jaccarddist_pairwise() function.
    • Add more tree-based methods to Taxon.
    • gambit.metric changes
      • jaccarddist_array and jaccarddist_matrix functions now accept any sequence type (e.g. list) for the refs argument, but with diminished performance.

0.3.0

  • CLI updates
    • gambit query now accepts query signatures from a signature file.
    • New command group gambit signatures with info and create subcommands.
    • New debug command group (hidden).
  • Performance enhancements
    • Signature calculation for multiple sequence files can be run in parallel.
    • Signature calculation with large k much faster.
    • Benchmarks for signature calculation.
  • Documentation
    • Installation instructions
    • More complete CLI docs
  • API and internals
    • Major refactor to gambit.kmers and gambit.signatures
    • find_kmers() renamed to calc_signature() and moved to gambit.signatures.calc, related functions also renamed and moved.
    • Refactored k-mer search into new find_kmers() function, which finds locations of prefix matches in sequence.
    • Several other classes and functions moved from gambit.kmers to gambit.signatures submodules.
    • Rearrangement of stuff within gambit.signatures.
    • Added required kmerspec attribute to AbstractKmerArray.
    • Renamed some KmerSpec attributes
    • Rename gambit.kmers.reverse_complement() -> revcomp()
    • Refactor of Jaccard functions
    • Removed _sparse from function names
    • Array and matrix functions now calculate distance only, renamed from jaccard_* to jaccarddist_*
  • New features
    • Most functions which take DNA sequences now accept str, bytes, or Bio.Seq.Seq.
    • Convert signatures between compatible KmerSpecs.
    • HDF5Signatures close() method and context manager.
  • Other
    • Updated Cython kmers code.
    • Many updates/improvements to tests.

0.2.2

  • Replace testdb_210126 with testdb_210818. Small enough to include all files, including reference signatures and query sequences, in version control.
  • Store pre-calculated query results for tests.
  • Some other minor test improvements and bug fixes.

0.2.1

  • Added license
  • Add setuptools to runtime dependencies
  • Minor docstring edits

0.2.0

  • User-facing
    • Rework JSON results format to be simpler and hide internal details
    • Add CSV results format (default)
    • Display progress while querying
    • Increase query performance and decrease memory usage
  • Internal
    • Major redesign of GAMBITDatabase and query funcs
    • GAMBITDatabase stores indices of reference signatures instead of loading them all up front
    • Read reference signatures in chunks when calculating distance matrix
    • Maintain reference to SQLAlchemy Session object on GAMBITDatabase.
    • strict classification parameter
    • Enables new behavior of finding and reconciling all matching taxa
    • Defaults to off, which results old behavior of using only closest match
    • Fix bug in consensus_taxon() and add tests
    • Flexible, generic progress monitoring API
    • Add to long-running functions like querying, distance matrix calculation, and k-mer finding
    • "archive" export format for saving full result data.
    • Lots of test improvements
    • More type annotations
    • Update API docs

0.1.0

Initial version.