DataFrames

Converts a ParsedMatch into a dictionary of pandas DataFrames, one per table. This is the primary interface for ML/data-science workflows.

Parquet dependency

to_parquet() requires pyarrow (recommended) or fastparquet:

bash

pip install pyarrow

Available tables

Key	Contents
`"players"`	Per-player snapshots at each sample tick
`"players_minute"`	Per-player stats aggregated by game minute
`"positions"`	Hero position heatmap grid (64-unit cells)
`"combat_log"`	Every damage, kill, heal, and modifier event
`"wards"`	Ward placements with exact map coordinates
`"objectives"`	Tower kills, barracks, Roshan kills
`"teamfights"`	Detected fight windows with participant stats
`"smoke_events"`	Smoke activations with grouped heroes and centroid
`"aegis_events"`	Aegis pickups, steals, and denies
`"draft"`	Pick and ban events in order
`"match"`	Single-row match metadata (id, duration, winner, …)
`"radiant_advantage"`	Per-minute gold and XP advantage curves
`"chat"`	All-chat and team-chat messages
`"courier_snapshots"`	Courier state sampled each tick

DataFrame conversion for :class:ParsedMatch output.

python

def build_dataframes(match: ParsedMatch) -> dict[str, pd.DataFrame]

Convert a :class:ParsedMatch into a dict of pandas DataFrames.