Legalkit Pipeline: French Legal Codes from LegiFrance in Parquet Format

This pipeline automates the extraction and publication of French legal codes from the official LegiFrance PISTE API to Hugging Face Datasets. Developed by louisbrulenaudet and last updated in November 2024, it converts legislative text into Parquet format for high-performance data processing. The system handles concurrent uploads and dynamic documentation generation for the resulting legal corpora.

Use Cases

• Training Large Language Models (LLMs) on French legal terminology
• Building semantic search engines for French statutory law
• Legal text summarization of specific French codes

Strengths

• Sourced from official LegiFrance PISTE API
• Apache-2.0 open-source license
• Automated Parquet conversion for Hugging Face compatibility

Limitations

• Unknown total record count and specific code coverage
• Requires PISTE API access for pipeline reproduction
• Geographically restricted to French jurisdiction

Provenance

Source

LegiFrance (PISTE API)

Collection Method

Automated pipeline extraction via API

Freshness

Last updated November 2024; designed for dynamic updates via the PISTE API.

Geography

France