Quick Start

Install, secrets, and a minimal first run

Getting started

Get a working local install and run a minimal pipeline path.

TipFull reference

Usage · Commands cheat sheet · Pipeline hub

1. Install

Python 3.11+ required.

pip install -e ".[dev]"

Optional extras you may want soon:

pip install -e ".[notebooks]"   # Jupyter walkthroughs
pip install -e ".[ocr]"         # pytesseract for page images
pip install -e ".[serve]"       # FastAPI DICIE REST server

2. Secrets

python scripts/setup_env.py          # create gitignored .env from template
# edit .env — paste keys (never commit)
python scripts/setup_env.py --status # which secrets are set (values never printed)

Useful variables: OPENROUTER_API_KEY, WANDB_API_KEY, optional HF_TOKEN. See Usage § Secrets and .env.example.

3. Minimal DICIE demo (no training)

python -m src.docie \
  --application salvage_claims \
  --text "LETTER OF GUARANTEE
Claim Number: CLM-2024-100200
VIN: 1HGCM82633A004352
Year: 2018
Make: Honda
Model: Accord" \
  --response-only

4. Sample corpus (optional)

python -m src.storage seed --seed 42 --also-export

Details: Sample Document Corpus.

5. Notebooks

pip install -e ".[notebooks]"
# browse via this Quarto site, or:
jupyter notebook notebooks/

Portal: Notebooks.

6. Documentation website

# local preview — install Quarto CLI from https://quarto.org/docs/get-started/
cd docs && quarto preview
# or: ./scripts/preview_docs_site.sh

# publish to Posit Connect Cloud (no GitHub Actions)
./scripts/publish_docs_site.sh

Full steps: How-to: launch the Quarto docs site.

Next steps

  • Train classifier / extractor smoke paths — Usage
  • Memo chain (src.pipeline.orchestrator) — Architecture
  • Discord bot — Usage
  • Browse rendered notebooks — Notebooks

See also

About · Data Provenance · DICIE · Live Posit site

Back to top