#!/usr/bin/env python3
# Description
###############################################################################
'''
Command-line interface for SHAP analysis runs.
Usage:
from OCDocker.OCScore.Analysis.SHAP.Cli import main
'''
# Imports
###############################################################################
from __future__ import annotations
import argparse
import json
from typing import Optional
from .Runner import run_shap_analysis
from .Studies import StudyHandles
# License
###############################################################################
'''
OCDocker
Authors: Rossi, A.D.; Monachesi, M.C.E.; Spelta, G.I.; Torres, P.H.M.
Federal University of Rio de Janeiro
Carlos Chagas Filho Institute of Biophysics
Laboratory for Molecular Modeling and Dynamics
This program is proprietary software owned by the Federal University of Rio de Janeiro (UFRJ),
developed by Rossi, A.D.; Monachesi, M.C.E.; Spelta, G.I.; Torres, P.H.M., and protected under Brazilian Law No. 9,609/1998.
All rights reserved. Use, reproduction, modification, and distribution are allowed under this UFRJ license,
provided this copyright notice is preserved. See the LICENSE file for details.
Contact: Artur Duque Rossi - arturossi10@gmail.com
'''
# Classes
###############################################################################
# Functions
###############################################################################
## Private ##
## Public ##
[docs]
def build_argparser() -> argparse.ArgumentParser:
'''Build command-line argument parser for SHAP analysis.
Returns
-------
argparse.ArgumentParser
Configured argument parser with all SHAP analysis command-line arguments.
'''
p = argparse.ArgumentParser(description="Run SHAP analysis for OCScore NeuralNet.")
p.add_argument("--storage", required=True)
p.add_argument("--ao_study", required=True)
p.add_argument("--nn_study", required=True)
p.add_argument("--seed_study", required=True)
p.add_argument("--mask_study", required=True)
p.add_argument("--df_path", required=True)
p.add_argument("--base_models", required=True)
p.add_argument("--study_number", type=int, required=True)
p.add_argument("--out_dir", required=True)
p.add_argument("--explainer", default="deep", choices=["deep","kernel"])
p.add_argument("--background_size", type=int)
p.add_argument("--eval_size", type=int)
p.add_argument("--stratify_by", nargs="*")
p.add_argument("--seed", type=int, default=0)
p.add_argument("--no_csv", action="store_true")
return p
[docs]
def main(argv: Optional[list[str]] = None) -> int:
'''Main entry point for SHAP analysis command-line interface.
Parameters
----------
argv : Optional[list[str]], optional
Command-line arguments to parse. If None, uses sys.argv. Default is None.
'''
args = build_argparser().parse_args(argv)
studies = StudyHandles(
ao_study_name=args.ao_study,
nn_study_name=args.nn_study,
seed_study_name=args.seed_study,
mask_study_name=args.mask_study,
storage=args.storage,
)
out = run_shap_analysis(
studies=studies,
df_path=args.df_path,
base_models_folder=args.base_models,
study_number=args.study_number,
out_dir=args.out_dir,
background_size=args.background_size,
eval_size=args.eval_size,
explainer=args.explainer,
stratify_by=args.stratify_by,
seed=args.seed,
save_csv=not args.no_csv,
)
print(json.dumps(out.__dict__, indent=2))
return 0
if __name__ == "__main__":
raise SystemExit(main())