MASSA Algorithm v. 2.2.1

[NEW RELEASE - v. 2.2.1]:

• Added support for .smi files.
• Updated logo.

MASSA Algorithm v. 2.2.0

[NEW RELEASE - v. 2.2.0]:

Changelog:

• Updated logo.
• Updated requirements.

MASSA Algorithm v. 2.1.0

Version 2.1.0 changes

• Updated README.md
• Fixed drop_errors issue with csv/excel files and change its default behaviour to True
• Fixed multiprocessing issue
• Change the behavior of large_library=True to use MiniBatch KMeans as default.

MASSA Algorithm v. 2.0.0

[NEW RELEASE - v. 2.0.0]:

Changelog:

• New feature: Automatic algorithm switching for large datasets
  For datasets with 10,000 molecules or more, MASSA now (automatically) switches from Hierarchical Clustering Analysis (HCA) to MiniBatch K-Means for improved scalability and performance. This behavior can be controlled with the new flag: -a / --large-datasets. Dendrogram plotting is disabled in this mode, and the -f / --dendrogram-plot flag is automatically ignored.
• New feature: MASSA can now be called from Python scripts using MASSA_Algorithm.pyMASSA.py_massa().
  The feature is already available, but instructions for using it will be provided in upcoming updates/versions.
• Bug fix: Fixed "NaN/None values in y" error.
  This error, caused by duplicate molecule names, is now resolved. Molecules with identical names are automatically renamed by appending their original index to ensure uniqueness.
• Change: MASSA now ensures preservation of original input order, but be aware that molecules with chemical errors are now skipped by default during processing and not included in the output.

MASSA Algorithm v. 1.0.0

[NEW RELEASE - v. 1.0.0]:

Changelog:

• New feature: Added the option -y to split the dataset into training, test, and validation subsets.
• New feature: Added the option -e to ignore molecule chemistry errors (errors are still logged in log.txt).
• Extensive code refactoring.