Testing the Samples

How to run and validate the curated samples.

Pre-checks

• Verify dependencies: python scripts/verify_install.py
• Ensure each sample has .env (copy from .env.example where provided) and requirements.txt installed.

Run a single sample

cd samples/01-basic-receipt-processing
python process_receipt.py

Expect: completes without import errors; outputs parsed receipt data.

Run all samples (batch)

cd samples
python test_all_examples.py

What it does: discovers sample dirs, runs primary script with timeouts, reports success/failure summary.

Env/provider notes

• Basic/no-API samples: set OCR_MODE=mock when available (e.g., sample 01) to avoid external calls.
• LLM-required samples: set OPENAI_API_KEY or AZURE_OPENAI_KEY (e.g., 02, 10, 22).
• GitHub samples: set GITHUB_TOKEN (e.g., 03, 19, 20).
• Parallel/agent samples: watch async output and timing (e.g., 16, 17, 25).

Verification checklist (per sample)

• README present and purpose clear
• requirements installed
• .env configured
• Runs to completion (<60s for basics) with sensible output
• Links in README work

Troubleshooting

• Import error No module named orchestrator: run pip install -e " .[dev]" from repo root.
• Auth errors: populate .env with required keys.
• Hangs/timeouts: run with a timeout; check network/backends.