BIDCI

BIDCI is a Python project for loading, preprocessing, and visualizing EEG data for brain–computer interface (BCI) tasks. It currently focuses on BIDS-formatted motor imagery datasets and builds on top of MNE-Python.

The goal of BIDCI is to provide:

• A clear, configurable pipeline for basic EEG preprocessing (filtering, epoching, event handling).
• Reusable visualization functions for sanity checks and condition-wise inspection.
• A structure that can be extended to additional BCI paradigms (e.g., SSVEP, P300) and more advanced analysis.

At this stage, the main entry point is a configuration file (YAML) and a simple Python script (test.py) that:

1. Loads the dataset according to the config.
2. Runs a preprocessing pipeline.
3. Generates optional summary plots for each subject and run.

Configuration and typed model

This project now uses a Pydantic ConfigModel to validate and coerce configuration values read from YAML. The runtime flow is:

• config.yaml (YAML) -> read with yaml.safe_load
• validated/coerced into ConfigModel via ConfigModel.model_validate(raw)
• used in code as a typed object (e.g., config.preprocessing.tmin)

Notes:

• The YAML file is still required as the canonical configuration input for reproducible runs.
• ConfigModel provides defaults and clear validation errors; required fields (e.g., subjects, runs) must be present or the loader will raise a clear error.
• If you prefer programmatic configuration, you can construct ConfigModel directly in Python instead of using YAML.