AIStk

stats_streaming.py

Purpose

Implements streaming-friendly statistics on Polars LazyFrames, avoiding materialization into NumPy arrays.
Designed for large AIS datasets processed with collect(engine="streaming").

Responsibilities

• Compute distances, straight-line displacement, tortuosity, turn index, and SOG metrics.
• Implement all metrics as Polars expressions (no Python loops).
• Enable incremental execution with Polars streaming engine.

Usage Example

import polars as pl
from aistk.stats_streaming import compute_stats_lazy
from aistk.core import AISDataset

lf = AISDataset("data/ais")._build()
out = compute_stats_lazy(lf, level="mmsi").collect(engine="streaming")
print(out)

This site is open source. Improve this page.