BIDS Query Language (BIQL)

A powerful SQL-like query language for BIDS neuroimaging datasets

Python 3.8+ License: MIT Tests

What is BIQL?

BIDS Query Language (BIQL) is a domain-specific query language designed for the Brain Imaging Data Structure (BIDS). It provides a familiar SQL-like syntax for finding, filtering, and analyzing neuroimaging datasets with unprecedented ease and power.

-- Find all functional scans from specific subjects
SELECT sub, ses, task, run, filepath
WHERE datatype=func AND sub IN [01, 02, 03]
ORDER BY sub, run

-- Get QSM reconstruction groups with file lists
SELECT filename, sub, ses, acq, COUNT(*) as total_files
WHERE (part=mag OR part=phase) AND suffix=MEGRE
GROUP BY sub, ses, acq

Key Features

🔍 Intuitive Syntax - SQL-like queries that feel natural to researchers and developers

📊 Powerful Aggregation - GROUP BY with auto-aggregated file lists for reconstruction workflows

🎯 Smart Pattern Matching - Wildcard and regex support for flexible file discovery

🧠 BIDS Native - Deep understanding of BIDS entities, metadata, and participant data

🔧 Multiple Output Formats - JSON, CSV, TSV, table, and paths output

Quick Example 

# Install BIQL
pip install biql

# Query your BIDS dataset
biql --dataset /path/to/bids "SELECT DISTINCT task WHERE datatype=func"

# Get QSM reconstruction groups
biql "SELECT sub, acq, COUNT(*) WHERE part=mag GROUP BY sub, acq" \
     --format table --dataset /data/qsm_study

Use Cases

  • Dataset Exploration - Quickly understand what data you have
  • Processing Pipelines - Generate file lists for batch processing
  • Metadata Analysis - Query acquisition parameters and demographics
  • Quality Control - Find missing files or inconsistent naming

Community

License

BIQL is open source software released under the MIT License.