HDDFLYZER v0.1.0

HDDFLYZER v0.1.0

Initial private release of HDDFLYZER: High-Dimensional Descriptor-based Feature Space Analyzer.

HDDFLYZER is a CLI-first cheminformatics toolkit with a Python API for reproducible, traceable molecular descriptor-space workflows. It generates descriptor, similarity, projection, and visualization outputs from compound collections, stores them with metadata, and supports later reconstruction, inspection, and analysis of completed runs.

Highlights

• Initial project import.
• CLI entry point: hddflyzer.
• Dataset-first workflow outputs under results/<tag>/.
• Manifest-based run traceability through manifest.json and workflow_summary.md.
• Pipeline orchestration layer with run_pipeline(), run_workflow(), and execute_workflow().
• Results reconstruction API with WorkflowRun, ResultArtifact, LoadedArtifact, load_workflow_run(), and load_artifact().
• Scientific result views for descriptor, similarity, and projection spaces.
• Science helpers for alignment, descriptor/projection correlations, neighborhood preservation, and group comparisons.
• Visualization helpers connected to reconstructed results.
• MkDocs documentation scaffold.
• Initial test suite included.

Validation

• pytest tests -q passed locally: 288 passed, 1 skipped.
• Initial repository import was reviewed to exclude generated outputs, caches, build artifacts, site/, root results/, and Python cache files.

Notes

This is an initial pre-release research software version. Output contracts and public APIs may still evolve as the project matures.