# Changelog

## Baseline

- First survey — no prior run to compare against. CAI 55.9.

## Lenses

- Code Health 74.5
- Architecture 97.6
- Maturity 56.5
- Readiness 45.8
- Security 57.8
- Domain Modelling 100.0

## Changes since last survey

- 300 commits — 124 feature/other, 176 fixes

## By area

- src/deepmol — 137 commits
- (root) — 68 commits
- (repo) — 28 commits
- tests/integration_tests — 22 commits
- tests/unit_tests — 21 commits
- docs/deepmol_docs — 14 commits
- .github/workflows — 2 commits
- examples/notebooks — 2 commits
- tests/heavy_test — 2 commits
- Docker/Dockerfile — 1 commit
- docs/imgs — 1 commit
- docs/requirements_rtd.txt — 1 commit
- tests/data — 1 commit

## Notable commits

- fix: [ADD] colabs + [FIX] documentation
- fix: [ADD] label encoder + [FIX] mode checks
- fix: [ADD] np bert tokenizer + [FIX] llm as featurizer
- fix: [ADD] np classifier fingerprint + [FIX] keras model for multilabel
- fix: [ADD] option of chirality to npclassifierfp + [FIX] multitask splitter has no option of seed working
- fix: [ADD] single dense model for single and multitask and for regression and classification
- fix: [ADD] tests for voting ensemble for classification and regression
- fix: [ADD] voting ensemble for classification and regression
- fix: [BUMP] version +  [FIX] biosynfoni
- fix: [FIX] 3d generation + vocab
- fix: [FIX] CI
- fix: [FIX] Dockerfile
- fix: [FIX] Hyperparameter optimization with cv to be in accordance with scikeras engine
- fix: [FIX] KerasModel when it is not possible to use KerasClassifier and KerasRegressor
- fix: [FIX] LLMs feature names
- fix: [FIX] README.md
- fix: [FIX] README.md
- fix: [FIX] SHAP tests - force to access CPU
- fix: [FIX] avoid duplicated featurizers in mixed_featurizer
- fix: [FIX] base featurizer + [FIX] requirements to include right version of tensorflow probability
- …and 280 more
