README.md

metadata

title: Modular Detector V2
emoji: 🧭
colorFrom: indigo
colorTo: blue
sdk: docker
pinned: false
license: mit
short_description: Modular addition helper.

Local run:

python -m venv .venv
source .venv/bin/activate
pip install -r requirements.txt
uvicorn app.main:app --reload

Open http://127.0.0.1:8000

Default embedding model: Qwen/Qwen3-Embedding-0.6B Default dataset: Molbap/modular-detector-embeddings

Note: The embedding model and the index must match. If you change the model, you must rebuild and re-upload the index.

Rebuild method index (from repo root, expects transformers clone at ./transformers or ./transformers_repo):

python scripts/build_index.py

Quick inference (curl):

curl -s http://127.0.0.1:8000/api/analyze \
  -H "Content-Type: application/json" \
  -d '{
    "code": "class Foo:\n    def forward(self,x):\n        return x\n",
    "top_k": 5,
    "granularity": "method",
    "precision": "float32"
  }' | jq

Push app to Space:

hf upload --repo-type space Molbap/modular-detector-v2 . \
  --include "Dockerfile" \
  --include "requirements.txt" \
  --include "README.md" \
  --include "app/**" \
  --include "static/**" \
  --commit-message "Update app"

Push method index to dataset:

hf upload --repo-type dataset Molbap/modular-detector-embeddings . \
  --include "embeddings_methods.safetensors" \
  --include "code_index_map_methods.json" \
  --include "code_index_tokens_methods.json" \
  --commit-message "Update method index"