v.2.0.6

What's Changed

• 🚀 v.2.0.6 by @RaulFD-creator in #164
• Update pypa/gh-action-pypi-publish digest to cef2210 by @renovate[bot] in #165
• 🛠️ Update name of IBM organisation in HF by @RaulFD-creator in #166

Full Changelog: v.1.0.7...v.2.0.6

v.1.0.7

What's Changed

• 🪲 minor bug fixes by @RaulFD-creator in #157
• 🛠️ Added support for PeptideMTR_lg model by @RaulFD-creator in #158
• 🪲🧹 Clean debugging-only code by @RaulFD-creator in #159
• 🛠️ Check comptability of transformers with molformer + tests by @RaulFD-creator in #160
• Update pypa/gh-action-pypi-publish digest to e2bab26 by @renovate[bot] in #161

Full Changelog: v.2.0.4...v.1.0.7

v.2.0.4

What's Changed

• 🏎️ Remove redundant pipeline call when neg_sampling by @RaulFD-creator in #135
• 🪲 calculate 'sequence' for nc db (#135) by @RaulFD-creator in #136
• Update pypa/gh-action-pypi-publish digest to ed0c539 by @renovate[bot] in #137
• Update actions/setup-python action to v6 by @renovate[bot] in #138
• 🏎️ Reduce FP size to 1024 by default by @RaulFD-creator in #139
• 🪲 Fix sequence_field assignment to apml-smiles or apml-sequence by @RaulFD-creator in #140
• 🪲🕵️‍♂️ SMILEStoSequence handle complex cycles by @RaulFD-creator in #141
• 🪲 Bond indices break program if repeated by @RaulFD-creator in #142
• 🛠️ Minor improvements to metrics and monomer library by @RaulFD-creator in #143
• 🛠️🕵️‍♂️ Add support for PepFuNN #144 by @RaulFD-creator in #145
• Update pypa/gh-action-pypi-publish digest to ab69e43 by @renovate[bot] in #146
• Configure Mend for GitHub Enterprise by @ibm-mend-app[bot] in #147
• Update pypa/gh-action-pypi-publish digest to 987f11e by @renovate[bot] in #150
• Update pypa/gh-action-pypi-publish digest to 3317ede by @renovate[bot] in #151
• Update actions/checkout action to v6 by @renovate[bot] in #152
• 🧹🛠️🪲 Housekeeping + Chemberta-3 + Documentation by @RaulFD-creator in #153
• Update pypa/gh-action-pypi-publish digest to 03f86fe by @renovate[bot] in #155
• 🛠️ Multiple small fixes and functionality improvements by @RaulFD-creator in #156

New Contributors

• @ibm-mend-app[bot] made their first contribution in #147

Full Changelog: v.2.0.3...v.2.0.4

v.2.0.3

What's Changed

• 🛠️ Update survey link + notebook by @RaulFD-creator in #102
• 🛠️ Substitute data for notebook by @RaulFD-creator in #103
• 🪲 New df after negative sampling, reset index @RaulFD-creator by @RaulFD-creator in #104
• 📝 Added instructions to Colab notebook by @RaulFD-creator in #105
• 🏎️ Implemented parallelism swtich (solved #100) by @RaulFD-creator in #106
• 🪲 Handle when no negative sampling by @RaulFD-creator in #107
• 🛠️ Trainer input np.ndarray as well by @RaulFD-creator in #108
• Update pypa/gh-action-pypi-publish digest to d417ba7 by @renovate[bot] in #109
• 🪲 Handle when db_file is not specified by @RaulFD-creator in #110
• 🪲 Handle when db_file is not specified by @RaulFD-creator in #111
• 🪲 Introduce order_monomers. Resolve #113 by @RaulFD-creator in #114
• 🪲 Fix issue with amidation solve #115 by @RaulFD-creator in #116
• 🪲 Solve # 117 by @RaulFD-creator in #118
• 🪲 Peptide ordering bug by @RaulFD-creator in #119
• 🛠️ Flag for flexible CV folds. Solve #112 by @RaulFD-creator in #120
• 🛠️ sim_args flag by @RaulFD-creator in #121
• Update actions/checkout action to v5 by @renovate[bot] in #122
• 🛠️ Add remove_duplicates flag + extra_x by @RaulFD-creator in #124
• 🪲 Only use extra_x if it is not None by @RaulFD-creator in #126
• 🪲 sample_negatives requires preprocessing beforehand by @RaulFD-creator in #128
• 🪲 Discretizer not working properly by @RaulFD-creator in #129
• 🛠️ Update notebook by @RaulFD-creator in #130
• 🪲 Minor bug fixes by @RaulFD-creator in #131
• 🪲 Sampling negatives requires preprocessing afterwards by @RaulFD-creator in #132
• 🪲 Consistent np.array size when fp is not computable by @RaulFD-creator in #133
• 🪲 Non-canonical is smaller than I thought by @RaulFD-creator in #134

Full Changelog: v.2.0.2...v.2.0.3

v.2.0.2

AutoPeptideML Release Notes

🚀 Highlights

• Simplified CLI for building model
• Simplified config.yml files
• Improved results directory including:
  • Clearer structure
  • Additional metadata
  • PDF summary through Quarto
• Clearer AutoPeptideML class interface
• New default pipelines: to-smiles, to-smiles-fast, to-sequences

🛠️ Features and enhacements

• SequenceFromSmiles has the capability of recognizing natural analogs
• VotingEnsemble class provides a clear interface for loading, saving, and using the models
• Introduction of the default pipelines
• Pipelines better output
• Faster BaseElements by substituting multithreading with multiprocessing
• New PDF summaries through Quarto

📚 Documentation

• Updated documentation for new CLI

🪲 Bug fixes

• Minor bug fixes

🧹 Code Maintenance

• Avoid tokenizers parallelism warning
• Simplified usage of databases
• Simplified configuration files
• Cleaner AutoPeptideML class
• Deprecated GridTrainer

🚧 Future plan

More info on: https://github.com/IBM/AutoPeptideML/milestone/2

• Integration with webserver
• Study tradeoffs between multiprocessing and multithreading for the BaseElement
• Improve low-level access to sim_args from .build_model method

Full Changelog: v.2.0.1...v.2.0.2

v.2.0.1

AutoPeptideML v2.0.0 – Release Notes

🚀 Highlights

• Version 2.0.0 released with support for new engine dependencies and improved reproducibility.
• Introduction of modular pipelines for flexible model integration.
• Support added for chemical language models and peptideclm.
• New pipeline elements: CanonicalizeSmiles, SMILES-to-sequence, save_experiments, model saving.

🛠️ Features & Enhancements

• Modular loading of pipeline components.
• New HPO system with custom configurations, grid search, and best set retrieval.
• Support for molecular fingerprints
• Support for chemical and peptide language models
• Support for half-precision inference with molecular (protein, chemical, or peptide) language models
• Custom DB integration
• Predict method and reproducible DB sampling (seeded).
• Updated supported models list
• Clearer CLI interface with prompts for easier model building

🐛 Bug Fixes

• Fixed multiple minor bugs and webserver integration issues.
• Resolved issues with negative label assignments and HPO definitions.
• Improved handling of empty input in parallel processing.

📚 Documentation

• Reformatted and expanded documentation including trainer docs and README for v2.
• Link added to APML v1.0.6 documentation.
• Minor docstring corrections.

🧹 Code Maintenance

• Removed legacy peptipedia dependency.
• Cleaned up debug prints and unused files.
• Improved str outputs for pipelines.

v.1.0.6

What's Changed

• 🪲 Bug fix: Enable MMseqs alignment again by @RaulFD-creator in #44
• 📝 Update README.md by @RaulFD-creator in #45

Full Changelog: v.1.0.5...v.1.0.6

v.1.0.5

What's Changed

• 🧹🛠️🪲: Patch and incorporate new PLMs by @RaulFD-creator in #35
• 🪲 Bug fix: Missing first token in for full sequence embedding by @RaulFD-creator in #36
• 🛠️ Dev: typer CLI instead of argparse by @RaulFD-creator in #37
• 🛠️ Code-dev: Backend for model storage onnx (see #33) by @RaulFD-creator in #40
• 🛠️🧹: Minor patches and code cleanliness improvements by @RaulFD-creator in #42
• New onnx models Zenodo link and Optuna sampler bug fix by @RaulFD-creator in #43

Full Changelog: v.1.0.4...v.1.0.5

v.1.0.4

What's Changed

• 🧹🤖 Simplified requirements, testing in more distributions by @RaulFD-creator in #32
• 🛠️🪲📝: Improvements to prediction mode by @RaulFD-creator in #34

Full Changelog: v.1.0.3...v.1.0.4

v.1.0.3

What's Changed

• 🛠️ Code-dev: Updated homology partitioning to Hestia (v.0.0.32) by @RaulFD-creator in #30
• 🛠️ Code-dev: Update default parameters (see #24) by @RaulFD-creator in #31

Full Changelog: v.1.0.2...v.1.0.3