ohlcv-hub 0.1.0

OHLCV data fetching and processing tool

ohlcv-hub

A small dataset builder for US stocks and ETFs: fetch daily OHLCV bars from Alpaca, optionally resample to weekly, validate, and export to Parquet or CSV with a validation report (missing days, duplicates, OHLC sanity).

Features

• Daily bars (1d): Fetch historical daily bars from Alpaca Market Data API; normalize, validate, and export.
• Weekly bars (1w): Resample daily bars to weekly (Monday 00:00 UTC); same validation and export.
• Validation: Duplicate detection, monotonic timestamps, OHLC sanity, volume checks; missing trading days (NYSE calendar).
• Output: Parquet (default) or CSV; optional validation_report.json.
• Rate limiting: Proactive throttling and 429 retry with backoff; optional --verbose debug logging.

Quickstart

Install

From PyPI (once published):

pip install ohlcv-hub

For local development:

pip install -e ".[dev]"

Environment

Set your Alpaca credentials:

• ALPACA_API_KEY — required
• ALPACA_API_SECRET — required
• ALPACA_DATA_BASE_URL — optional (default: https://data.alpaca.markets)

Example commands

# Check configuration
ohlcv-hub doctor

# Fetch daily bars (default feed: iex)
ohlcv-hub fetch --symbols SPY,QQQ --start 2024-01-01 --end 2024-02-01 --tf 1d --out ./data

# Fetch weekly bars (resampled from daily)
ohlcv-hub fetch --symbols SPY --start 2024-01-01 --end 2024-02-01 --tf 1w --out ./data

# Use a different feed: --feed iex | sip | boats | otc
ohlcv-hub fetch --symbols SPY --start 2024-01-01 --end 2024-01-10 --tf 1d --out ./data --feed sip

Output

• Daily: <out>/ohlcv_1d_<YYYYMMDD>_<YYYYMMDD>.parquet (or .csv)
• Weekly: <out>/ohlcv_1w_<YYYYMMDD>_<YYYYMMDD>.parquet (or .csv)
• Report (if --report): <out>/validation_report.json (missing days, issues summary)

Schema

Column	Description
symbol	Ticker (e.g. SPY)
timeframe	1d or 1w
ts	Bar timestamp (UTC)
open, high, low, close	OHLC
volume	Volume
source	Data source
currency	Currency (e.g. USD)
adjustment	Adjustment (e.g. raw)

Notes / Limitations (MVP)

• No intraday: Daily and weekly bars only; no minute or realtime data.
• No strategies: Fetch, validate, and export only; no backtesting or signals.
• Alpaca: Requires an Alpaca account and data subscription. Free tier often includes IEX; SIP may require a paid plan. Use --feed iex (default) or --feed sip as appropriate.

Development

pip install -e ".[dev]"
pytest -q
bandit -r ohlcv_hub -x tests -q

For dependency auditing (Python 3.10+): pip install -e ".[audit]" then pip-audit.

Internal release and security notes live under docs/internal/.

License

MIT