[CODE QUALITY] Q1 — 5.8 GB of bloat in repo (venvs + tensorflow source)

[cluster2600]

Summary

5.8 GB of binary/generated files committed directly into the repo, making it unusable for normal git operations (clone, push, fetch).

Bloat Found

tensorflow/        ~4.5 GB  — TensorFlow source code (should be a pip dependency)
venv_new/          ~800 MB  — Python virtualenv (should be in .gitignore)
env-coreml/        ~500 MB  — Second virtualenv (duplicate)

Fix

1. Add to .gitignore:

venv*/
env*/
*.egg-info/
__pycache__/
tensorflow/

2. Remove from git history (BFG or git filter-repo):

git filter-repo --path tensorflow --invert-paths
git filter-repo --path venv_new --invert-paths  
git filter-repo --path env-coreml --invert-paths
git push origin --force --all

3. Add tensorflow as proper dependency in pyproject.toml:

[tool.poetry.dependencies]
tensorflow = "^2.15"

Impact

• Repo size will drop from ~6 GB to ~50 MB
• Clone time: from 20+ minutes to <30 seconds
• CI/CD becomes feasible

Sprint

Sprint 1 — Story 1.3 (Dependency cleanup)

Found by PM agent audit — Feb 17, 2026