How lamalab-org/polymetrix works
PolyMetriX is a polymer-informatics library for preparing datasets, representing polymer structures, calculating descriptors, comparing polymer–molecule mixtures, and supporting train/test splitting. A typical workflow loads curated glass-transition data or supplies pSMILES, computes polymer-level features, and uses the resulting features and labels for modeling. The graph includes the documented comparison and splitting capabilities. Some implementation details are unsampled; those capabilities are shown without guessing their internal call paths.
Main entry points: data_main.py Feature data preparation · sidechain_backbone_featurizer.py Polymer part features · curated_tg_dataset.py Curated Tg dataset · molecule.py Molecule featurizer
Components 14
Data and splitting
- Curated Tg dataset
src/polymetrix/datasets/curated_tg_dataset.py - Dataset interface
src/polymetrix/datasets/dataset.py - Dataset splitters
src/polymetrix/splitters/splitters.py - Feature data preparation
scripts/data_main.py
Chemical structures
- Polymer model
src/polymetrix/featurizers/polymer.py - Molecule model
src/polymetrix/featurizers/molecule.py
Feature calculation
- Feature calculator base
src/polymetrix/featurizers/base_featurizer.py - Chemical descriptors
src/polymetrix/featurizers/chemical_featurizer.py - Polymer part features
src/polymetrix/featurizers/sidechain_backbone_featurizer.py - Combined featurizer
src/polymetrix/featurizers/multiple_featurizer.py - Molecule featurizer
src/polymetrix/featurizers/molecule.py
Mixture comparison
- Polymer–molecule comparator
src/polymetrix/featurizers/comparator.py
Other components
- Researcher
- Dataset hosting
Connections 18
- Researcher → Curated Tg dataset: loads
README.md - Curated Tg dataset → Dataset hosting: downloads CSV
src/polymetrix/datasets/curated_tg_dataset.py - Curated Tg dataset → Dataset interface: implements
src/polymetrix/datasets/curated_tg_dataset.py - Dataset splitters → Dataset interface: reads labels
src/polymetrix/splitters/splitters.py - Researcher → Feature data preparation: runs
scripts/data_main.py - Feature data preparation → Polymer model: parses pSMILES
scripts/data_main.py - Feature data preparation → Chemical descriptors: selects descriptors
scripts/data_main.py - Feature data preparation → Polymer part features: selects polymer features
scripts/data_main.py - Feature data preparation → Combined featurizer: combines features
scripts/data_main.py - Polymer part features → Polymer model: extracts polymer parts
src/polymetrix/featurizers/sidechain_backbone_featurizer.py - Chemical descriptors → Feature calculator base: extends
src/polymetrix/featurizers/chemical_featurizer.py - Polymer part features → Feature calculator base: uses
src/polymetrix/featurizers/sidechain_backbone_featurizer.py - Molecule featurizer → Feature calculator base: uses
src/polymetrix/featurizers/molecule.py - Combined featurizer → Polymer part features: invokes
src/polymetrix/featurizers/multiple_featurizer.py - Researcher → Polymer–molecule comparator: compares mixture
README.md - Polymer–molecule comparator → Polymer part features: uses polymer features
README.md - Polymer–molecule comparator → Molecule featurizer: uses molecule features
README.md - Molecule featurizer → Molecule model: featurizes
src/polymetrix/featurizers/molecule.py