Skip to content

GETTING_STARTED.md

Latest commit

 

History

History
72 lines (50 loc) · 1.95 KB

GETTING_STARTED.md

File metadata and controls

72 lines (50 loc) · 1.95 KB
Raw

Getting started — public data → Neo4j → API

Tutorial path (roadmap 5.2): clone to running API in ~15 minutes.

1. Clone and configure

git clone https://github.com/LordKay-sudo/bioinsight-graph.git
cd bioinsight-graph
cp .env.example .env

2. Full stack (recommended)

docker compose up --build
Service URL
Web UI http://localhost:8080
API docs http://localhost:8000/docs
Neo4j Browser http://localhost:7474

The seed job loads the Open Targets 24.06 frozen slice (CC0 public data) into Neo4j. See PROVENANCE.md.

3. Verify the graph

curl http://localhost:8000/api/v1/stats
curl "http://localhost:8000/api/v1/genes?q=BRCA1"
curl http://localhost:8000/api/v1/genes/ENSG00000012048/evidence

In the UI: search BRCA1 → open gene detail → switch to Compare or search a disease.

4. Local dev (no Docker UI)

docker compose up -d neo4j
py -3 scripts/build_frozen_slice.py
py -3 scripts/seed_neo4j.py
cd api && uvicorn app.main:app --reload --port 8000
cd web && npm install && npm run dev

5. Notebook walkthrough (roadmap 4.4)

With the stack running:

cd api && pip install -r requirements.txt -r ../notebooks/requirements.txt
jupyter notebook ../notebooks/one_gene_exploration.ipynb

See notebooks/README.md.

6. Agent / MCP layer (optional)

Point embabel-mcp at http://localhost:8000/api/v1. Use MCP prompt public-data-to-mcp-tutorial for a guided Cursor setup.

Strict production ingest

For real bulk Open Targets data with ontology validation:

py -3 scripts/download_opentargets_bulk.py --release 24.06 --max-genes 500
py -3 scripts/etl_opentargets.py --input data/raw/opentargets_bulk.json --strict
py -3 scripts/seed_neo4j.py --strict

See ONTOLOGY_SCHEMA.md and BENCHMARKS.md.