Skip to content

Get Ffield Opt Report Workflow

Generate a complete force-field optimization report from one run directory.

This composite workflow delegates to the existing focused commands instead of reimplementing their analysis logic. REPORT_COMMANDS is intentionally kept near the top of the module so developers can see the complete report pipeline without tracing the runner.

Command: get-ffield-opt-report

Generate a complete force-field optimization report by running the existing trainset-data, trainset-comment, optimization-result, plot, and bulk-modulus workflows in sequence. Run this command inside a force-field optimization directory to use fort.99, fort.74, trainset.in, and geo automatically.

By default, results are written under: reaxkit_workspace/analysis/get-ffield-opt-report//

Examples


  1. Build a report from the current optimization directory:
     reaxkit get-ffield-opt-report

  2. Build a report from another run and choose the output directory:
     reaxkit get-ffield-opt-report --run-dir run --output report

Arguments

Scientific choices

Flag Required Default Help Choices
--min-points No 6 Minimum EOS points required for each bulk-modulus fit (default: 6)
--flip-sign No False Flip energy signs during bulk-modulus fitting
--no-shift-min-to-zero No False Do not shift minimum energy to zero during bulk-modulus fitting

Input and file selection

Flag Required Default Help Choices
--engine No reaxff, ams, lammps
--input No Input path used for engine detection
--run-dir, --dir No . Force-field optimization run directory
--fort99 No fort.99 Path to fort.99
--fort74 No fort.74 Path to fort.74
--trainset No trainset.in Path to trainset file
--geo No geo Path to the geo file

Outputs and plots

Flag Required Default Help Choices
--entry-per-figure No 6 Maximum entries per grouped-bar figure (default: 6)
--output, --outdir No Optional report-directory override
--detail-format No Optional detail format (default: Parquet; legacy: CSV). parquet, csv

Execution

Flag Required Default Help Choices
--execution No auto Execution backend; unsupported backends fall back to serial with a logged reason. auto, serial, threads, processes
--workers No 0 Frame workers: auto or N (default: auto).
--chunk-size No 0 Maximum in-flight frames: auto or N.

Storage and cache

Flag Required Default Help Choices
--run-id No Run identifier for run-scoped layout. Example: --run-id run_91ac0e, which reuses that run identifier.
--project-root No reaxkit_workspace Project root that contains inputs/, data/, analysis/, etc. Example: --project-root ./workspace, which stores run artifacts there.
--analysis-id No Optional analysis artifact id; defaults to run id. Example: --analysis-id comparison-a, which names the analysis artifact explicitly.
--input-cache, --no-input-cache No True Reuse parsed input frames across commands (default: enabled; use --no-input-cache to force source reads for reproducibility checks or benchmarks). Example: --no-input-cache, which reloads frames from their source files.
--frame-cache-max-gb No 10.0 Maximum workspace frame-cache size in GiB (default: 10; use 0 for unlimited). Example: --frame-cache-max-gb 20, which caps cached frames at 20 GiB.
--output-profile No standard Select the shared artifact policy. Standard writes declared default outputs; minimal keeps core tables; full and legacy include optional details. Default: standard. Example: --output-profile full, which includes declared optional detail tables. minimal, standard, full, legacy

Diagnostics and compatibility

Flag Required Default Help Choices
-h, --help No show this help message and exit
--help-all, --all-flags No Show every option, grouped by purpose.
--log No verbose, quiet

Common Runtime and Presentation Arguments

These are shared workflow-level CLI flags added before command-specific options, covering runtime context (engine/input/storage) and output presentation/export behavior.

Each command table above includes its shared and inherited options.