Refactored Graph-based Cellular Automata (GCA)

This is a refactor of the original Graph-based_Cellular_Automata repository to make it:

• runnable via a single CLI (gca ...)
• modular (package code under gca/, thin wrappers instead of large scripts)
• reproducible (centralized outputs + deterministic seeding)

The scientific logic is preserved where possible; the refactor mainly changes structure.

Install

pip install -r requirements.txt
pip install -e .

Expected inputs

Place the original data bundle under:

raw_data/
  CCLE_expression.csv
  CCLE_mutations.csv
  sample_info.csv
  ...
Reference.csv
estimator.json
Effective_mutations.csv
In_complexes.csv
In_pathways.csv
sorted_drugs.csv

Pipeline (cell line → modules → drug modules)

gca init
gca baseline
gca simulate-cell-line              # or: --depmap-id ACH-000001
gca extract-modules
gca build-drug-modules

Outputs are written under:

outputs/
  baselines/<lineage>.json
  simulations/cell_line/<DepMap_ID>.json
  modules/cell_line/<DepMap_ID>.txt
  modules/drug/<DRUG_ID>{sensitive,resistant}.txt

Legacy code

Original source/*.py files are kept (unmodified) in legacy_source/ for comparison.