Skip to content

quickstart.md

Latest commit

 

History

History
64 lines (44 loc) · 1.53 KB

quickstart.md

File metadata and controls

64 lines (44 loc) · 1.53 KB
Raw

Quickstart Guide

5-Minute Setup

Prerequisites

  • Python 3.10+
  • UV (recommended) or pip

Install

curl -LsSf https://astral.sh/uv/install.sh | sh
uv pip install -e .[dev]

First Evaluation

Option A: Use the API

uv run uvicorn nedc_bench.api.main:app --reload
# Submit via curl (note: enum values are lowercase)
curl -s -X POST "http://localhost:8000/api/v1/evaluate" \
  -F reference=@data/csv_bi_parity/csv_bi_export_clean/ref/aaaaaajy_s001_t000.csv_bi \
  -F hypothesis=@data/csv_bi_parity/csv_bi_export_clean/hyp/aaaaaajy_s001_t000.csv_bi \
  -F algorithms=all -F pipeline=dual | jq

# Watch progress
wscat -c ws://localhost:8000/ws/<job_id>

Option B: Use scripts (no CLI)

# Run all Beta algorithms and save SSOT
uv run python scripts/run_beta_batch.py

# Compare Alpha vs Beta parity
uv run python scripts/compare_parity.py

Option C: Docker Compose

docker-compose up -d
curl http://localhost:8000/api/v1/health

Understanding Output

  • TP/FP/FN and FA/24h are returned per algorithm. See docs/algorithms/metrics.md for definitions.
  • Epoch and IRA operate on fixed windows; TAES returns fractional counts; DP/Overlap return integer counts.

Example Dataset

  • Sample CSV_BI files live under data/csv_bi_parity/csv_bi_export_clean/ with ref/ and hyp/ subfolders and list files in lists/.

What’s Next