Snekmer

Description

The Snekmer v1.2.0 release primarily includes the new Snekmer Learn and Snekmer Apply pipelines, which allow users to evaluate the confidence of kmers in annotation and use the learned information to annotate unknown sequences. This release also include code refactoring to fix memory leakage issues in the clustering workflow, remove obsolete/legacy code and specifications, simplify versioning, resolve bugs.

What's Changed
* Merged in new fixes from main by biodataganache in https://github.com/PNNL-CompBio/Snekmer/pull/68
* updates by biodataganache in https://github.com/PNNL-CompBio/Snekmer/pull/76
* Kmer association by jjacobson95 in https://github.com/PNNL-CompBio/Snekmer/pull/109
* docs: change base class for SnekmerModel by christinehc in https://github.com/PNNL-CompBio/Snekmer/pull/111
* Kmer association by jjacobson95 in https://github.com/PNNL-CompBio/Snekmer/pull/115

**Full Changelog**: https://github.com/PNNL-CompBio/Snekmer/compare/v1.0.4...v1.2.0

Contributors

* Jeremy Jacobson (jjacobson95) - primary code development
* Christine Chang (christinehc) - code development
* Jason McDermott (biodataganache) - code development supervision
* Bill Nelson (wichne) - subject matter expert

The Snekmer v1.0.4 release includes minor code refactoring to directly call the flexible `SnekmerModel` object rather than fixing the model type to logistic regression. All Snekmer-generated objects now also store the Snekmer version number as an attribute, important for reproducibility in analytical workflows involving Snekmer.

Changelog

Description

* `snekmer.model.SnekmerModel` is now used to build models, with the default model type remaining as `"logistic"`.
* Snekmer version numbers are stored as attributes in the `snekmer.score.KmerScoreScaler`, `snekmer.vectorize.KmerBasis`, and `snekmer.vectorize.KmerVec` objects (as `self.snekmer_version`).

What's Changed
* [docs] Add Apple silicon / additional instructions by christinehc in https://github.com/PNNL-CompBio/Snekmer/pull/103
* [chore] merge main changes into branch by christinehc in https://github.com/PNNL-CompBio/Snekmer/pull/104
* Update README.md by lnanderson in https://github.com/PNNL-CompBio/Snekmer/pull/106
* chore: combine branches by christinehc in https://github.com/PNNL-CompBio/Snekmer/pull/107
* Model version by christinehc in https://github.com/PNNL-CompBio/Snekmer/pull/105

Contributors

* Christine Chang (christinehc) - code development
* Bill Nelson (wichne) - feature request
* Jason McDermott (biodataganache)
* Jeremy Jacobson (jjacobson95)

New Contributors

* lnanderson made their first contribution in https://github.com/PNNL-CompBio/Snekmer/pull/106

**Full Changelog**: https://github.com/PNNL-CompBio/Snekmer/compare/v1.0.3...v1.0.4

The Snekmer v1.0.3 release includes code refactoring and some changes to improve usability, including improving documentation and logged error messages, updating CI to directly install and execute Snekmer for better error catching, and restricting some CLI arguments to minimize errors.

Changelog

Major Changes

* Certain snakemake args (`--configfile`, `--cores`) no longer produce errors under default conditions.
* Documentation updates now include instructions for the Snekmer Docker container.
* Snekmer prevents automated model-building for files containing an insufficient number of sequences.
* GitHub actions (CLI) workflow now captures the full installation and code execution process instead of indirect testing via calling individual Snakefiles.
* Where appropriate, `run` directive has been replaced with `script` directive to reduce Snakefile size and make formatting/linting/debugging easier.

Minor Changes

* Documentation has been updated, particularly for troubleshooting.
* Nested output directory option is now working again.
* Minor docstring/type hinting updates

Contributors

* Christine Chang (christinehc)
* Jason McDermott (biodataganache)
* Jeremy Jacobson (jjacobson95)

Changelog

Major Changes

* Overhaul to `kmerize.smk`: Now proceeds through a bag-of-words and generator-based approach to reduce memory requirements. Affects `snekmer cluster`, `snekmer model`, and `snekmer search`.
* Overhaul to `snekmer cluster`: Clustering methods and distance matrix computation methods are updated. Now includes option to invoke BSF.
* Tutorial has been overhauled to reflect code changes

Minor Changes

* Documentation now exists via snekmer.readthedocs.io
* Updates/fixes to address the large memory requirements of `snekmer cluster`
* Installation instructions updated to account for various software ecosystems

Contributors

* Christine Chang (christinehc)
* Jason McDermott (biodataganache)
* Abby Jerger (abbyjerger)

Changelog

- Updated clustering module for compatibility with changes to `snekmer.cluster.KmerClustering` object, `snekmer.transform`, and `kmerize.smk`
- Cluster now works locally with small test example

Changelog

- Enable nested directory output structure for snekmer search mode
- e.g. _**output/model/family.pkl**_ -> _**output/model/{alphabet}/{k}/model/family.pkl**_
- Update code formatting/style for all .py and .smk (snakemake) files