Development#

This guide provides information for developers who want to contribute to or extend AutoClean EEG.

Setting Up Development Environment#

Prerequisites:

  • Python 3.10 or higher

  • Git

  1. Clone the repository:

    git clone https://github.com/cincibrainlab/autoclean_pipeline
cd autoclean_pipeline

  2. Create a virtual environment:

    python -m venv .venv
source .venv/bin/activate  # On Windows: .venv\Scripts\activate

  3. Install development dependencies:

    pip install -e ".[dev]"

Project Structure#

The project is organized as follows:

  • configs/: Configuration templates

  • src/autoclean/core/: Core classes and functionality

    • pipeline.py: Main entry point for the API

    • task.py: Base class for all task implementations

  • src/autoclean/io/: Modular processing functions

    • export.py: Exporting functions

    • import.py: Importing functions

  • src/autoclean/mixins/signal_processing/: Signal processing related functions

  • src/autoclean/mixins/viz/: Visualization related functions

  • src/autoclean/step_functions/: Modular processing functions

    • continuous.py: Core preprocessing steps

    • reports.py: Post-task reports such as processing log

  • src/autoclean/plugins/: Import and event handling plugins

  • src/autoclean/tasks/: Task implementations

    • resting_eyes_open.py: Resting state task

    • assr_default.py: ASSR task

    • And others…

  • src/autoclean/utils/: Utility functions

    • config.py: Configuration handling

    • database.py: Database operations

    • logging.py: Logging functionality

  • src/autoclean/tools/: Additional features for the pipeline

    • autoclean_review.py: Review GUI

Architecture#

AutoClean follows a modular architecture with several key components:

  1. Pipeline Class: Central coordinator that manages configuration, processing, and output.

  2. Task Classes: Implementations for specific EEG paradigms (resting state, ASSR, etc.).

  3. Step Functions: Modular processing operations that can be combined into workflows.

  4. Database Tracking: Database-backed tracking of processing runs.

The architecture uses a combination of:

  • Abstract Base Classes: For extensibility and consistent interfaces

  • Mixins: For shared functionality across tasks

  • Asynchronous Processing: For parallel file processing

  • YAML Configuration: For reproducible processing parameters