ros2_logger

A ROS 2 Python library for structured experiment data logging. Automatically discovers log variables from your node at shutdown, writes structured CSVs, and generates Jupyter analysis notebooks.

Key Features

• Custom log data types — ColumnarLog provides append-like list behavior for both scalar and vector sequences
• Automatic shutdown hooks — on node exit (normal or error), collects all log-type variables and writes them to CSV
• Column ordering — log columns are sorted first by a user-provided ordering integer, then alphabetically
• Auto directory creation — creates a data_analysis/ folder alongside the ROS 2 node that invoked the logger
• Built-in analysis tools — populates data_analysis/ with:
  • DataAnalysis.ipynb — full analysis and plotting notebook
  • SingleLogAnalysis.ipynb — single log file analysis
  • ControllerComparison.ipynb — compare runs across controllers
  • utilities.py — RMSE computation, plotting utilities, PDF export for Overleaf/LaTeX
  • log_files/ — subdirectory for CSV log files

Note: By default the logger requires time, x, y, z, and yaw columns to be present, ensuring minimum dataset completeness for quadrotor experiments. This can be modified in the source.

Quick Start

1. Clone into your ROS 2 workspace src/ directory:

<<<<<<< HEAD <<<<<<< Updated upstream git clone git@github.com:evannsmc/ROS2Logger.git ros2_logger

git clone git@github.com:evannsmc/ROS2Logger.git

Stashed changes ======= git clone git@github.com:evannsmc/ROS2Logger.git 82534d9 (update GitHub username from evannsm to evannsmc)

2. **Build** the workspace:
```bash
colcon build --symlink-install

<<<<<<< HEAD 3. Try the example — use the test_logger package for a working example.

Usage

from ros2_logger.logger import Logger
from ros2_logger.logtypes import ColumnarLog

class MyNode(Node):
    def __init__(self):
        super().__init__('my_node')
        self.logger = Logger(self)

        # Define log columns (name, ordering_int)
        self.time_log = ColumnarLog("time", 0)
        self.x_log    = ColumnarLog("x", 1)
        self.y_log    = ColumnarLog("y", 2)

    def control_loop(self):
        self.time_log.append(self.get_clock().now())
        self.x_log.append(state.x)
        self.y_log.append(state.y)

On shutdown, the logger automatically discovers all ColumnarLog attributes, writes them as columns to a CSV, and generates analysis notebooks in data_analysis/.

Package Structure

ros2_logger/
├── ros2_logger/
│   ├── logger.py                    # Main Logger class
│   ├── logtypes.py                  # ColumnarLog and related types
│   ├── shutdown_helpers.py          # Shutdown hook management
│   └── analysis_helpers_move/
│       ├── utilities.py             # RMSE, plotting, PDF export
│       └── generate_test_data.py    # Test data generation
├── package.xml
└── setup.py

License

MIT

4. Use my test_logger package for a quick and easy example of this system in action, and learn to operate this Logger package!

82534d9 (update GitHub username from evannsm to evannsmc)