qf-data 0.1.0

Fast, minimal data utilities for trading: schemas, resampling, alignment, gaps, calendars.

qf-data

Fast, minimal data utilities for trading. Asset-agnostic, crypto-friendly.

Features

• Canonical Arrow schemas for trades, quotes, bars.
• UTC nanosecond time utilities.
• Trades→bars resampling. Bars→higher TF.
• Alignment and as-of joins.
• Missing-data handling and quality flags.
• Optional adapters for Pandas/Polars.

Install

pip install qf-data

Quick start

import numpy as np
from qf_data import step_from_str, trades_to_bars

ts = np.array([0, 500_000_000, 1_100_000_000], dtype=np.int64)  # ns
px = np.array([100.0, 101.0, 99.5])
sz = np.array([0.4, 0.1, 0.2])

step = step_from_str("1s")
bars = trades_to_bars(ts, px, sz, step, start_ns=0, end_ns=2_000_000_000)
print(bars.keys())  # dict of arrays

Design

• Pure Python + NumPy/Arrow. No pandas in core.
• Pure functions. No global state. UTC only.
• 24/7 by default. Calendars are pluggable.

License

MIT