Prune Codebase

Prune Codebase

Goal

Reduce a repo to only what is required for the application flow, without destructive deletion.

Workflow

1. Build an execution schematic of the current app flow.
2. Save the schematic artifact to codex_docs/app_execution_schematics.md.
3. Identify required files from that schematic.
4. Ignore docs/ and codex_docs/ during pruning decisions.
5. For each prune candidate:
   • Move it to pruned_objects/<original-path> (never delete).
   • Run a validation pass immediately after the move.
   • If validation fails, restore the object and mark it as required.
6. Repeat until all candidates are processed.
7. Reassess requirements.txt and README.md against the final pruned structure.
8. Move old requirements.txt and README.md into pruned_objects/ before writing updated versions.

Guardrails

• Prefer one-object-at-a-time pruning for traceability.
• Always keep path structure under pruned_objects/.
• Treat unclear dependencies as required unless validation proves otherwise.
• Do not skip validation between moves.

Example command patterns

mkdir -p pruned_objects
mkdir -p "pruned_objects/$(dirname path/to/object)"
mv "path/to/object" "pruned_objects/path/to/object"

# restore if needed
mkdir -p "$(dirname path/to/object)"
mv "pruned_objects/path/to/object" "path/to/object"