Improve Logging System & Enhance FIRE Algorithm for Compound Molecules #2
Description
Overview
The project already includes a light ML automation pipeline. This issue will focus on two improvements:
-
Better Logging System
- Design and implement a more robust logging system for the pipeline and simulation steps.
- Aim: Improved traceability, easier debugging, and actionable logs for automation steps.
- Suggested: Consider integrating structured logging (e. g., JSON logs, timestamps, levels) and optional file-based/system log output.
-
Improving FIRE for Compound Molecules
- Strengthen the Fast Inertial Relaxation Engine (FIRE) algorithm particularly for handling compound (multi-component) molecules.
- Evaluate current limitations or edge cases where compound geometries and optimizations fail or are suboptimal.
- Possible approaches: enhanced convergence criteria, special-case handling for multi-center atoms, user-configurable parameters for compound types.
Acceptance Criteria
- Robust logging established for all automation steps and ML submodules.
- FIRE improvements validated with compound/multi-component molecular input, with before/after benchmarks or examples if possible.
- Documentation updated for new logging options and usage notes for enhanced FIRE support.
Further suggestions, benchmarks, or implementation notes are welcome in comments.