Skip to content

v.0.0.35

Compare
Choose a tag to compare
@RaulFD-creator RaulFD-creator released this 05 Dec 12:48
· 53 commits to main since this release
99889fe

What's Changed

Summary

  1. 📝 Docs: Minor changes to documentation. Added examples for molecules
  2. 🧹 Tidy code: Improved progress bars for similarity calculation
  3. 🛠️ Code-dev: Change distance in SimilarityArguments for sim_function in the case of 'embeddings' calculation to harmonize with the rest of similarity functions
  4. 🛠️ Code-dev: Introduction of Canberra distance
  5. 🛠️ Code-dev: Introduction of Lipinski vector to molecular_similarity
  6. 🛠️ Code-dev: Substituted gzip for pickle for saving the precalculated partitions and additional metadata
  7. 🪲 Bug fix: Multiple bug fixes with if-else statements
  8. 🪲 Bug fix: Bug fix with metadata for embeddings trying to savequery_embds as well.
  9. 🪲 Bug fix: Small bug fixes with similarity calculations
  10. 🪲 Bug fix: Avoid the situation where fp='ecfp' and sim_function='jaccard' and derive it to tanimoto by default.
  11. 🪲 Bug fix: Apply distance to similarity transformation
  12. 🏎️ Optim: Optimize bulk similarity functions using vectorized numpy operations, almost 25x speedup with target_df of 20,000, probably scales better than original implementation
  13. 🏎️ Optim: Small optimization of the ccpart algorithm

Merge trail

Full Changelog: v.0.0.34...v.0.0.35