Skip to content

CHANGELOG.md

Latest commit

 

History

History
77 lines (59 loc) · 4.03 KB

CHANGELOG.md

File metadata and controls

77 lines (59 loc) · 4.03 KB

Changelog

All notable changes to this project will be documented in this file.

The format is based on Keep a Changelog, and this project adheres to Semantic Versioning.

[Unreleased]

Fixed

  • OTDD Distance Returning All inf: Replaced removed torch.symeig calls with torch.linalg.eigh in vendored OTDD code, fixing compatibility with PyTorch 2.0+
  • OTDD Error Reporting: Upgraded silent warning to error level logging, added exception type to messages, and added a post-computation summary of failed pairs
  • NaN Feature Handling in OTDD: Added validation that detects and replaces NaN values in feature arrays before OTDD computation, preventing silent numerical failures

Removed

  • run_pipeline.py and run_pipeline.sh — redundant wrapper scripts superseded by the themap CLI

[v0.3.0] - 2025-08-19

Added

  • Pipeline Infrastructure: Complete configuration-driven pipeline system for distance computation workflows

    • themap.pipeline module with CLI, configuration management, and execution engine
    • Support for both directory-based dataset discovery and explicit dataset specification
    • YAML/JSON configuration files with validation and comprehensive examples

  • Distance Computation Fixes: Corrected dataset-level distance computation methodology

    • Fixed Euclidean and Cosine distance implementations to work properly with variable-sized datasets
    • Implemented proper pairwise distance computation between individual molecules across datasets
    • Fixed method naming inconsistencies (compute_distance() → get_distance())

  • Utility Scripts: New data processing and conversion tools

    • scripts/csv_to_jsonl.py - Convert CSV files to THEMAP's native JSONL.GZ format with SMILES validation
    • scripts/clean_smiles.py - SMILES validation and cleanup utility for datasets

  • Examples Reorganization: Structured example system by complexity level

    • examples/basic/ - Introductory examples for new users
    • examples/distance/ - Distance computation workflows
    • examples/advanced/ - Complex research applications
    • Comprehensive example configurations in configs/examples/

  • Enhanced Documentation: Pipeline usage guides and workflow documentation

    • docs/PIPELINE_GUIDE.md - Complete pipeline usage documentation
    • docs/README_PIPELINE_SCRIPTS.md - Utility scripts documentation

  • New Test Coverage: Comprehensive test suite for new functionality

    • Unit tests for distance computation modules
    • Pipeline component tests
    • Configuration validation tests

Fixed

  • Critical Distance Computation Bug: Resolved incorrect Euclidean/Cosine distance implementation

    • Previous implementation incorrectly flattened entire feature matrices from different-sized datasets
    • Now correctly computes pairwise distances between individual molecules across datasets
    • Uses mean of pairwise distances as dataset-level distance metric

  • Pipeline Method Calls: Fixed incorrect method names in pipeline execution

    • Updated all distance computation calls to use correct get_distance() method
    • Fixed TaskDistance instantiation and usage patterns

  • Dataset Loading: Fixed explicit dataset specification mode

    • Individual dataset loading now creates proper Task objects for distance computation
    • Both directory-based and explicit dataset modes now work correctly

Changed

  • Examples Structure: Moved examples from scripts/ to organized examples/ directory
  • Distance Module: Refactored distance computation classes for better consistency and correctness
  • Configuration System: Enhanced pipeline configuration with better validation and error handling

Removed

  • Deprecated example scripts from scripts/ directory (moved to examples/)
  • Legacy tasks_distance.py module (functionality integrated into new distance classes)

[v0.2.0] - YYYY-MM-DD

Added

  • Initial changelog entry. Describe new features, changes, and fixes here.

[v0.1.0] - YYYY-MM-DD

Added

  • Initial release.