libibt

A Python library for reading iRacing IBT telemetry files. Rust core with PyO3 bindings for fast parsing; returns data as PyArrow tables.

Installation

pip install libibt

or with uv:

uv add libibt

From source

uv sync
just build

Requires a Rust toolchain (for maturin) and uv.

Usage

from libibt import ibt

# Load from file path, bytes, PathLike, or file-like object
log = ibt('session.ibt')

print(log)
# LogFile(file_name='session.ibt', channels=142, laps=12)

# Access a single channel (PyArrow table with 'timecodes' + value columns)
speed = log.channels['Speed']
print(speed.column_names)  # ['timecodes', 'Speed']

# Merge all channels into one table
table = log.get_channels_as_table()
df = table.to_pandas()

# Access laps (PyArrow table: num, start_time, end_time — all in ms)
for i in range(log.laps.num_rows):
    lap = log.laps.column("num")[i].as_py()
    start = log.laps.column("start_time")[i].as_py()
    end = log.laps.column("end_time")[i].as_py()
    print(f"Lap {lap}: {start} - {end}")

# Session metadata
print(log.metadata['track_name'])
print(log.metadata['session_info_yaml'])  # Full iRacing session YAML

Filtering and resampling

from libibt import ibt

log = ibt('session.ibt')

# Select specific channels
subset = log.select_channels(['Speed', 'Throttle', 'Brake'])

# Filter to a time range (ms, inclusive start, exclusive end)
segment = log.filter_by_time_range(60000, 120000)

# Filter to a specific lap
lap5 = log.filter_by_lap(5)

# Combine filtering and channel selection
lap5_speed = log.filter_by_lap(5, channel_names=['Speed', 'Throttle'])

# Resample all channels to match a reference channel's timebase
aligned = log.resample_to_channel('Speed')

# Resample to a custom timebase
import pyarrow as pa
target = pa.array(range(0, 100000, 100), type=pa.int64())  # 10 Hz
resampled = log.resample_to_timecodes(target)

# Chain operations
df = (log
    .filter_by_lap(5)
    .select_channels(['Speed', 'RPM', 'Gear'])
    .get_channels_as_table()
    .to_pandas())

All methods return new LogFile instances (immutable pattern).

Progress callback

def on_progress(current, total):
    print(f"Channel {current}/{total}")

log = ibt('session.ibt', progress=on_progress)

Channel metadata

Each channel carries typed metadata via the ChannelMetadata dataclass:

from libibt import ibt, ChannelMetadata

log = ibt('session.ibt')

# Extract from a channel table
meta = ChannelMetadata.from_channel_table(log.channels['Speed'])
print(meta.units)        # 'm/s'
print(meta.desc)         # 'GPS based speed'
print(meta.interpolate)  # True

# Or from a PyArrow field directly
field = log.channels['Speed'].schema.field('Speed')
meta = ChannelMetadata.from_field(field)

# Serialize back to PyArrow field metadata
raw = meta.to_field_metadata()  # dict[bytes, bytes]

ChannelMetadata is a frozen (immutable) dataclass with fields:

Field	Type	Description
units	str	Unit string (e.g., "m/s", "revs/min")
desc	str	Channel description
interpolate	bool	Whether to use linear interpolation when resampling

Metadata is preserved through all LogFile operations (filtering, resampling, merging).

Metadata fields

log.metadata contains fields from both the binary header and parsed session YAML:

Key	Description
session_info_yaml	Full iRacing session info YAML
session_info	Parsed YAML as a dict
track_name	Track internal name
track_display_name	Track display name
track_city	Track city
track_country	Track country
track_length	Track length
track_id	Track ID
track_type	Track type
series_id, season_id, session_id, sub_session_id	iRacing IDs
tick_rate	Sample rate (typically 60 Hz)
record_count	Total number of records
lap_count	Number of laps
session_start_date	Session start date
start_time, end_time	Session time bounds
event_type	Event type (e.g., "Race")
session_type, session_name	Current session type and name
driver_name	Recording driver's username
driver_user_id, driver_irating, driver_license	Driver details
car_name, car_id	Car screen name and ID
car_gear_count, car_redline_rpm, car_shift_rpm, car_idle_rpm	Car specs
num_drivers	Number of drivers (excluding pace car)
weather_temp, weather_surface_temp, weather_humidity	Weather conditions
weather_skies, weather_wind_speed, weather_wind_dir	Weather conditions
car_setup	Full car setup dict
sectors	Sector split definitions

Limitations

Array variables (count > 1, e.g. tire temperature arrays) are not yet supported. Only scalar variables are returned as channels.

Development

# Build Rust extension (release)
just build

# Build (debug, faster compile)
just build-debug

# Run tests
just test

# Run all checks (format, clippy, typecheck, tests)
just check

# Format code
just format

License

MIT