README.md

rebake Python Bindings

Python bindings for the rebake-rs robotics data processing library. This package lets you convert ROS bag files to ML-ready datasets (like LeRobot format) using Python.

Installation

Requirements

• Python 3.9 or later
• Rust toolchain (for building from source)
• uv (Python package manager)

Install with uv

cd rebake-rs/python
uv sync

Optional extras

Use extras when you need rebake's development-only dependencies:

# Python test dependencies (pytest, Pillow, rosbags)
uv sync --extra test

# Documentation dependencies
uv sync --extra docs

Quick Start

Here is a simple example that converts a ROS bag file to a LeRobot dataset:

import rebake

# Step 1: Load a ROS bag file
config = rebake.Rosbag1IngestorConfig()
ingestor = config.build()
context = ingestor.ingest("/path/to/your/rosbag.bag")

# Step 2: Synchronize data to a fixed frame rate
sync_config = rebake.ZeroOrderHoldTimeSynchronizerConfig(fps=30)
synchronizer = sync_config.build()
context = synchronizer.run(context)

# Step 3: Convert to LeRobot format
transformer_config = rebake.LeRobotV21TransformerConfig(
    outdir="./output",
    robot_model="/path/to/robot_model.yaml",
)
transformer = transformer_config.build()
context = transformer.run(context)

print("Done! Output saved to ./output")

Documentation

Full API documentation is available locally via MkDocs.

Building Documentation

# Install docs dependencies
uv sync --extra docs

# Build documentation (output in site/ directory)
uv run mkdocs build

# Or serve locally with hot-reload at http://127.0.0.1:8000
uv run mkdocs serve

Viewing in VSCode

You can preview the built documentation directly in VSCode using the Live Preview extension:

1. Install the Live Preview extension
2. Build the documentation:

   uv run mkdocs build

3. Right-click the site/index.html file and select "Show Preview"

Alternatively, use uv run mkdocs serve and open http://127.0.0.1:8000 in your browser for live-reloading during development.