moseplib

Core Python library for reading, processing, and analyzing sensor data from the MOSEP (Modular Hardware and Software System for Multi-Sensor Environment Perception) measurement setup. Provides tools to extract point cloud and weather station data from ROS2 bag files and convert them into analysis-ready formats.

Features

• ROS2 bag file reading — Extract point cloud (LiDAR/radar) and weather station data from .mcap bag files, including support for custom Lufft WSX weather station message types.
• Point cloud processing — Convert raw point clouds to pointcloudset datasets stored as Parquet files. Resample, subset, and aggregate by configurable time periods and spatial regions.
• Weather time series — Deserialize Lufft weather station messages into structured pandas DataFrames with multi-level columns (category × parameter). Resample and combine data from multiple weather stations.
• Spatial filtering — Limits utility for defining 3D bounding boxes and applying inclusion/exclusion filters to point cloud data.
• Statistics — Per-frame point cloud statistics (point count, mean/sum intensity).
• Metadata extraction — CLI tool to extract bag file metadata (topics, duration, precipitation stats) and generate Markdown reports with embedded precipitation plots.
• Image extraction — Decode CompressedImage messages from bag files to RGB images.

Installation

Requires Python ≥ 3.9, < 3.12.

pip install moseplib

Or as part of the mosep-analysis workspace (recommended):

# From the mosep-analysis root
uv sync

CLI

extract_metadata

Extract metadata and precipitation plots from ROS2 bag files:

extract_metadata <path_to_bagfile> [--output_dir <dir>]

Generates a Markdown file with YAML frontmatter containing bag metadata (topics, timestamps, duration, precipitation statistics) and an embedded precipitation plot.

Package Structure

src/moseplib/
├── cli/
│   ├── get_metadata_from_bagfiles.py   <- extract_metadata CLI entry point
│   └── sensor_setup.yaml              <- Sensor configuration (Ouster LiDAR, Smartmicro radar)
├── config/
│   ├── precipitation_codes.py          <- WMO Synop Code 4680 weather type definitions
│   └── custom_ros_msgs/
│       └── lufft_wsx_interfaces/msg/   <- Custom ROS2 message definitions for Lufft weather station
├── data/
│   ├── config.py                       <- Package-level config (topics, field names, paths)
│   ├── utils.py                        <- Limits (3D bounding box), spatial filtering utilities
│   ├── read_rosbag.py                  <- ROS2 bag file reading and message deserialization
│   ├── pointcloud_processing.py        <- Point cloud extraction, loading, resampling, aggregation
│   ├── timeseries_processing.py        <- Weather time series loading, deserialization, resampling
│   ├── pc_statistics.py                <- Point cloud statistics (n_points, mean/sum intensity)
│   └── extract_pcs_from_bagfile.py     <- Standalone script for batch point cloud extraction
└── tools/
    ├── fix_ros2_metadata_file.py       <- Fix timestamp ordering in ROS2 bag metadata files
    └── create_parquet_files_from_pc.py <- Batch convert bag files to parquet

Supported Sensors

Sensor	Type	Details
Ouster OS-1-64-U02	LiDAR	1024×10 mode, RNG19_RFL8_SIG16_NIR16_DUAL profile
Smartmicro UMRR-11 Type 132	Radar	—
Lufft WSX-series	Weather station	Temperature, humidity, wind, precipitation, radiation, air pressure

Dependencies

• pandas — DataFrames and time series
• pointcloudset — Point cloud dataset management
• rosbags — ROS2 bag file reading
• pyarrow — Parquet I/O
• rich — Terminal output formatting
• tqdm — Progress bars
• kaleido — Static image export for Plotly figures