Changelog

All notable changes to this project will be documented in this file.

The format is based on Keep a Changelog and this project adheres to Semantic Versioning.

v0.4.0 - 2026-04-15

Added

  • Added support for reading various types of compressed files with load_molecules

  • PyJess now exposes transformation matrices! -> bumped requirement to PyJess 0.10.0

  • Added CLI flag --save-transformations in order to save the transformation matrix as .npz in homogeneous coordinates per match. This requires numpy!

  • Unittests for compressed files types .gz, .lz4, .xz, .bz2

  • Unittests for transformation matrices

Fixed

  • Nextflow pipeline updated to handle output types and CLI changes

  • Warnings in load_molecules now adhere to the CLI --warn argument

v0.3.3 - 2026-04-02

Added

  • Added module enzymm.output with FullMatchResults class (to wrap previous output)

  • Added SimpleMatcTable results class (to resemble the table style on the webserver)

  • Added MatchResidueTable results class to map literature annotations per residue

  • Added the ability to write tables to either .tsv files or to create a polars.DataFrame

  • This _BaseTable framework can be easily adapted for different styles!

  • call Tables.create(kind, matches) to create your choice of output representation

  • Added CLI flag --simple-results to return a simple match results table instead of the full one.

  • Added CLI flag --per-residue-results to return residue to annotation mappings table too.

  • Added CLI flag --write-parquet to write results tables as parquet files instead of tsvs. Requires polars!

v0.3.2 - 2026-03-18

Added

  • Added ability to construct template from PDB or CIF structure file

  • Added template load + dump roundtrip to unittests

Fixed

  • Improved Regex pattern for CATH identifiers

v0.3.1 - 2025-12-15

Fixed

  • Added check for molecule id - molecule id will be derived from structure file or failing that the filename

  • Jess object will only be created once per template size batch and passed along

  • Added a cif file as input to the unittests

v0.3.0 - 2025-11-25

Added

  • Added Documentation

  • --include-template flag added. Set it to include the template pdb in the output

  • templates and molecules can now be dumped too (with the ability to transform them) -> bumped requirement to PyJess 0.9.0

Changed

  • breaking different structures (molecules, hits, templates) will now be seperated in MODEL/ENDMDL blocks (compatibility with some molecular viewers).

  • breaking --transform behaviour changed. Setting it will write in the template coordinate system.

  • --transform and --include-query are now compatible. --transform will write one output pdb per matched template

Fixed

  • removed two duplicate 3-residue templates; we are down to 7605 templates in total now

  • fixed some EC numbers which contained apostrophes

  • will now properly recognize CIF structures starting with a comment line as CIF structures

  • outsourced dumping molecules, templates and hits to PyJess

  • HEADER lines in the pdb output were changed to regular REMARK lines.

  • improved various docstrings

  • testing for pdb file output

v0.2.0 - 2025-09-17

Added

  • Support for mmCIF files through gemmi. integration in PyJess version >= 0.7.0

  • Support for Selenocysteine and Pyrrolysine residues in templates

  • Added reference_residue and reference_pdb attributes to AnnotatedResidue Class

  • Added iter and len methods to Residue and Template Classes in line with PyJess

Changed

  • breaking: Command line argument -j/--jess replaced with -p/--parameters

  • breaking: Command line argument -n/--n-jobs replaced with -j/--jobs

  • breaking: Residue.residue_name attribute replaced with Residue.name

  • breaking: Residue.residue_number attribute replaced with Residue.number

  • breaking: Missing input files will now raise errors instead of just a warning

  • breaking: Bugfix related to max_candidates in PyJess where supplying many templates might supress some matches (requires PyJess version >= 0.7.0)

  • Will now match sites across chain interfaces even if not specified in the template (requires PyJess version >= 0.7.0)

Fixed

  • Should run a lot faster due to many optimizations to PyJess version >= 0.6.0

  • Improved handling of logistic regression models and model ensembles through new classes

  • Included PyJess version in the output tsv file

  • Included Command line argument (if run from cli) in the output tsv file

  • enzymm.template.load_templates() will now load supplied templates by default without further kwargs

  • enzymm.template.load_templates() will now by default use only one thread (improves performance)

  • Removed useless argument conservation_score of Matcher. Behaviour unchanged.

v0.1.7 - 2025-08-21

Fixed

  • Improved thread-based parallelism over both molecules and batches of templates to avoid waiting for costly molecule/template searches

  • Fixed the progress bar

  • Pickling of template objects

Added

  • bash entry points for apptainer and docker containers

v0.1.6 - 2025-08-21

Fixed

  • Fixed overwriting the path variable. now appending to it.

v0.1.5 - 2025-08-21

Fixed

  • Installing apptainer manually in apptainer workflow to avoid later issues with apt

v0.1.4 - 2025-08-21

Fixed

  • Error in downloading and unpacking oras

v0.1.3 - 2025-08-21

Fixed

  • Write permissions given to apptainer workflow. Skip pypi upload (silently!) if version tag exists.

v0.1.2 - 2025-08-20

Fixed

  • Github actions workflow for apptainer should now generate a release artifact and upload via oras.

v0.1.1 - 2025-08-20

Fixed

  • Github actions workflow for apptainer runs directly on unbuntu base without –fakeroot. New tag should satisfy pypi.

v0.1.0 - 2025-08-20

Added

  • Unittests for Annotated Templates and residues

  • Added Apptainer via ORAS built in github actions

Changed

  • Disabling checks on multichain template and query pairs for chain relationships !

Fixed

  • Eliminated the wait times until all molecules had been scanned before smaller templates used for searches

  • Removed unnecessary duplicate EMO function tags from residue annotations

  • Fixed some inconsistent ptm residue annotations

  • Eliminated some unnecessary steps in the unittests which added considerable compute time

  • Check in the CLI for pairwise distances for which no prediction models exist if the tag –unfilteredqwas not passed

  • Spelling and badges in the README.md

v0.0.3 - 2025-08-07

Added

  • Added Information to README.md

  • Added Docker Container via Github actions to GHCR

Changed

  • Disabling checks on multichain template and query pairs for chain relationships

Fixed

  • Fixed Github actions url

  • cpu counting on linux systems

  • Fixed attempt at filtering for pairwise distances without determined logistic models

v0.0.2 - 2025-07-20

Fixed

  • Fixed Github actions to properly build from source

v0.0.1 - 2025-07-20

Initial release.