mokapot v0.7.1

This release updates the build process for mokapot.

Changed

• Updated the build to align with PEP517

mokapot v0.7.0

This release adds significantly more functionality, a small speedup, and expands the documentation 🎉

Added

• Support for downstream peptide and protein quantitation with
  FlashLFQ. This is accomplished
  through the mokapot.to_flashlfq() function or the to_flashlfq() method of
  LinearConfidence objects. Note that to support the FlashLFQ format, you'll
  need to specify additional columns in read_pin() or use a PepXML input file
  (read_pepxml()).
• Added a top-level function for exporting confident PSMs, peptides, and
  proteins from one or more LinearConfidence objects as a tab-delimited file:
  mokapot.to_txt().
• Added a top-level function for reading FASTA files for protein-level
  confidence estimates: mokapot.read_fasta().
• Tests accompanying the support for the features above.
• Added a "mokapot cookbook" to the documentation with helpful code snippets.

Changed

• Corresponding with support for new formats, the mokapot.read_pin() function
  and the LinearPsmDataset constructor now have many new optional parameters.
  These specify the columns containing the metadata needed to write the added
  formats.
• Starting mokapot should be slightly faster for Python >= 3.8. We were able to
  eliminate the runtime call to setuptools, because of the recent addition of
  importlib.metadata to the standard library, saving a few hundred
  milliseconds.

mokapot v0.6.2

This release cleans up mokapot's logging.

Added

• Now checks to verify there are no debugging print statements in the code
  base when linting.

Fixed

• Removed debugging print statements.

mokapot v0.6.1

This release fixes a few bugs.

Fixed

• Parsing Percolator tab-delimited files with a "DefaultDirection" line.
• Label column is now converted to boolean during PIN file parsing.
  Previously, problems occurred if the Label column was of dtype object.
• Parsing modifications from pepXML files were indexed incorrectly on the
  peptide string.

mokapot v0.6.0

This version contains a few bug fixes, improved tests, and adds support for PepXML files as input.

Added

• Support for parsing PSMs from PepXML input files.
• A changelog.

Fixed

• Parsing a FASTA file previously failed if an entry was not followed by a
  sequence. Now, missing sequences are tolerated and a warning is given instead.
• When the learned model was worse than the best feature and the lower scores
  were better for the best feature, assigning confidence would fail.
• Easy access to grouped confidence estimates in the Python API were not working
  due to a typo.
• Deprecation warnings from Pandas about the regex argument.
• Sometimes peptides were removed as shared incorrectly when part of a protein
  group.

Changed

• Refactored and added many new unit and system tests.
• New pull-requests must now improve or maintain test coverage.
• Improved error messages.

mokapot v0.5.1

This release contains a small bug fix.

Bug fixes:

• Fixed a problem when unsigned integer data types were used as features.

mokapot v0.5

This release mainly improves memory usage when parsing PIN files.

Features

• PIN files are now parsed in batches, decreasing the memory required.

Bug fixes:

• The specified decoy_prefix was not properly used when no decoys were present in the FASTA database.

mokapot v0.4

This version adds support for grouped confidence estimates:

• The read_pin() function now has an optional group_column argument.
• The LinearPSMDataset constructor now has an optional group_column argument.

mokapot v0.3

This version adds:

• Support for protein-level confidence estimates using target-only databases.
• Access to decoy confidence estimates either using the decoy_confidence_estimates attribute, or the new decoys argument in the to_txt() method.

mokapot v0.2

The API for mokapot should now be largely stable and feature-complete for the initial release. Further updates toward 1.0 should be bug fixes, tests, and documentation updates.

This release adds full support for protein-level confidence estimates. Additionally, there were major modifications to the Python API: parameters for classes were front-loaded to initialization instead of being distributed throughout the methods. This will hopefully make them easier to use, particularly if the default parameters are suitable.