Update splitting - Githubissues

plinder-org / plinder

Protein Ligand INteraction Dataset and Evaluation Resource

https://plinder.sh

Apache License 2.0

163 stars 9 forks source link

Update splitting #28

Closed tjduigna closed 2 months ago

tjduigna commented 2 months ago

This PR contains outstanding changes to the splitting code, on which gs://plinder/2024-06/v2/splits/split.parquet is dependent. Here is an all-too-abbreviated summary of the included changes:

Utilities to update and expand the plindex for usage in splitting
Correctly treat the ligand scores when clustering on tanimoto_similarity_max
Updated graph config defaults used in splitting
- additional knobs to control which systems are candidates for test/val
Various changes for how systems are removed from "train" to improve novel ligand count
Improved stratification and test set analysis

github-actions[bot] commented 2 months ago

Coverage report

Click to see where and how coverage changed

File	Statements	Missing	Coverage	Coverage (new stmts)	Lines missing
src/plinder/core/index
utils.py
src/plinder/core/scores
clusters.py
query.py
src/plinder/core/split
plot.py					17-18, 38, 44-56, 155, 157, 170, 175, 188-204, 257-260, 294-312, 328-329, 378-396, 399-430, 433-475, 478-605, 608-616, 627-628, 631-662, 665-712, 715-757, 764, 766-768, 771-775, 779-822, 834
utils.py					181-182
src/plinder/data
clusters.py					194-231, 273, 331-368, 379
smallmolecules.py					152-156, 163-166
splits.py					245-307, 324-333, 428-433, 564, 574-582, 599-613, 655-744, 751-866, 889-894, 903-905, 917-920, 962
src/plinder/data/utils/annotations
aggregate_annotations.py
ligand_utils.py
src/plinder/eval/docking
stratify_test_set.py					10-14, 35, 63-64, 73-110, 173, 184, 224-247, 282-298
Project Total

This report was generated by python-coverage-comment-action