Laboratoire-de-Chemoinformatique/SynPlanner v1.4.2

on GitHub

Added

ORD (Open Reaction Database) Support

• New synplan/utils/ord/ package for reading ORD .pb Dataset files via protobuf
  (dataset_pb2.py, reaction_pb2.py) without depending on ord-schema
• iter_ord_reactions() iterator for lazy ORD .pb file parsing
• convert_ord_to_smiles() utility for batch ORD-to-SMILES conversion
• synplan ord_convert CLI command for converting ORD .pb files to .smi
• ReactionReader and RawReactionReader now accept .pb files natively
• _ORDReadAdapter for transparent ORD reading through the existing Reader protocol
• 367-line test suite (test_ord_reader.py) covering ORD parsing

Configuration

• ReactorConfig pydantic model for typed Reactor construction parameters
  (automorphism_filter, delete_atoms, one_shot, fix_aromatic_rings,
  fix_tautomers) with to_reactor_kwargs() serialization
• load_reaction_rules() now accepts optional reactor_config parameter

Changed

Rule Extraction

• Rule deduplication now uses CGR (condensed graph of reaction) instead of
  ReactionContainer hashing — correctly preserves query-level atom annotations
  (neighbors, hybridization) when rules contain QueryContainer molecules
• _update_rules_statistics() and sort_rules() updated to use cgr_to_rule
  mapping for CGR-based dedup
• process_completed_batch() receives cgr_to_rule dict

Docker

• Added .dockerignore to exclude .git, .venv, docs, tests, tutorials,
  build caches, and data directories from Docker build context

Dependencies

• Added protobuf>=4.21 to core dependencies (ORD .pb support)
• Added grpcio-tools>=1.78.0 to dev dependencies (protobuf code generation)

Fixes & Cleanup

• depict_settings() calls updated to module-level function (was
  MoleculeContainer.depict_settings())
• routes_clustering_report / routes_subclustering_report: safer target SMILES
  lookup with .get() fallback instead of direct key access
• Removed unused yaml imports from filtering.py and standardizing.py
• Removed unused os import from cli.py
• Removed unused Any import from mapping.py
• Import order cleanup (ruff/black formatting)