{"url":"/dataset/french-legal-cases","name":"french-legal-cases","full_name":null,"description_markdown":"---\r\nannotations_creators:\r\n- no-annotation\r\nlanguage:\r\n- fr\r\nlanguage_creators:\r\n- found\r\nlicense:\r\n- cc-by-4.0\r\nmultilinguality:\r\n- monolingual\r\npretty_name: French Legal Cases Dataset\r\nsize_categories:\r\n- n>1M\r\nsource_datasets:\r\n- la-mousse/INCA-17-01-2025\r\n- la-mousse/JADE-17-01-2025\r\n- la-mousse/CASS-17-01-2025\r\n- la-mousse/CAPP-17-01-2025\r\ntask_categories:\r\n- text-generation\r\n- text-classification\r\ntask_ids:\r\n- language-modeling\r\n- entity-linking-classification\r\n- fact-checking\r\n- intent-classification\r\n- multi-label-classification\r\n- multi-input-text-classification\r\n- natural-language-inference\r\n- semantic-similarity-classification\r\n- sentiment-classification\r\n- topic-classification \r\n- sentiment-analysis\r\n- named-entity-recognition\r\n- parsing\r\n- extractive-qa\r\n- open-domain-qa\r\n- closed-domain-qa\r\n- news-articles-summarization\r\n- news-articles-headline-generation\r\n- dialogue-modeling \r\n- dialogue-generation\r\n- abstractive-qa\r\n- closed-domain-qa\r\n- keyword-spotting\r\n- semantic-segmentation\r\n- tabular-multi-class-classification\r\n- tabular-multi-label-classification\r\n- document-retrieval\r\n- document-question-answering\r\npaperswithcode_id: french-legal-cases\r\ndataset_info:\r\n  features:\r\n  - name: id\r\n    dtype: string\r\n  - name: url\r\n    dtype: string\r\n  - name: case_number\r\n    dtype: string\r\n  - name: contenu\r\n    dtype: string\r\n  - name: sommaire\r\n    dtype: string\r\n  - name: sommaire_bis\r\n    dtype: string\r\n  - name: metadata\r\n    dtype: string\r\n  - name: dataset_source\r\n    dtype: string\r\n  - name: previous_text\r\n    dtype: string\r\n  - name: current_text\r\n    dtype: string\r\n  - name: next_text\r\n    dtype: string\r\n  - name: triplet_index\r\n    dtype: int64\r\nconfigs:\r\n- config_name: default\r\n  data_files:\r\n  - split: train\r\n    path: data/train-*\r\n---\r\n\r\n# Dataset Card for French Legal Cases Dataset\r\n\r\n## Dataset Description\r\n\r\n- **Homepage:** [https://huggingface.co/datasets/la-mousse/combined-fe-caselaw](https://huggingface.co/datasets/la-mousse/combined-fe-caselaw)\r\n- **Repository:** [https://huggingface.co/datasets/la-mousse/combined-fe-caselaw](https://huggingface.co/datasets/la-mousse/combined-fe-caselaw)\r\n- **Paper:** N/A\r\n- **Point of Contact:** [Your Contact Information]\r\n\r\n### Dataset Summary\r\n\r\nThis dataset combines French legal cases from multiple sources (INCA, JADE, CASS, CAPP) into a unified format with overlapping text triplets. It includes decisions from various French courts, processed to facilitate natural language processing tasks.\r\n\r\n### Supported Tasks and Leaderboards\r\n\r\n- **Tasks:**\r\n  - Text Generation\r\n  - Legal Document Analysis\r\n  - Text Classification\r\n  - Language Modeling\r\n\r\n### Languages\r\n\r\nThe dataset is monolingual (French).\r\n\r\n## Dataset Structure\r\n\r\n### Data Instances\r\n\r\nEach instance contains:\r\n- Document identifiers (id, url, case_number)\r\n- Original text content (contenu)\r\n- Summaries (sommaire, sommaire_bis)\r\n- Metadata (JSON formatted string)\r\n- Text triplets (previous_text, current_text, next_text)\r\n- Position tracking (triplet_index, window_index)\r\n\r\nExample:\r\n```python\r\n{\r\n    'id': 'CASS12345',\r\n    'url': 'https://...',\r\n    'case_number': '12-34567',\r\n    'contenu': 'Full text...',\r\n    'sommaire': 'Summary...',\r\n    'sommaire_bis': 'Additional summary...',\r\n    'metadata': '{\"date_decision\": \"2023-01-01\", ...}',\r\n    'dataset_source': 'CASS',\r\n    'previous_text': 'Previous chunk...',\r\n    'current_text': 'Current chunk...',\r\n    'next_text': 'Next chunk...',\r\n    'triplet_index': 0,\r\n    'window_index': 0\r\n}\r\n```\r\n\r\n### Data Fields\r\n\r\n- `id`: Unique identifier\r\n- `url`: Source URL\r\n- `case_number`: Case reference number\r\n- `contenu`: Full text content\r\n- `sommaire`: Primary summary\r\n- `sommaire_bis`: Secondary summary\r\n- `metadata`: JSON string containing additional metadata\r\n- `dataset_source`: Origin dataset (INCA/JADE/CASS/CAPP)\r\n- `previous_text`: Previous text chunk\r\n- `current_text`: Current text chunk\r\n- `next_text`: Next text chunk\r\n- `triplet_index`: Position in sequence of triplets\r\n- `window_index`: Window number for long texts\r\n\r\n### Data Splits\r\n\r\n- Training split only\r\n\r\n## Dataset Creation\r\n\r\n### Curation Rationale\r\n\r\nThis dataset was created to provide a standardized format for French legal texts, with overlapping text chunks suitable for various NLP tasks.\r\n\r\n### Source Data\r\n\r\n#### Initial Data Collection and Normalization\r\n\r\n- INCA: Court of Cassation decisions\r\n- JADE: Administrative court decisions\r\n- CASS: Court of Cassation decisions\r\n- CAPP: Court of Appeal decisions\r\n\r\n### Preprocessing\r\n\r\n- Text chunking with 230-token chunks and 30-token overlap\r\n- Sliding window approach for long texts\r\n- Metadata preservation and standardization\r\n- Token count verification\r\n- JSON formatting for metadata\r\n\r\n### Quality Control\r\n\r\n- Token length verification\r\n- Chunk coherence checks\r\n- Metadata validation\r\n- Error logging and handling\r\n\r\n## Considerations for Using the Data\r\n\r\n### Social Impact of Dataset\r\n\r\nThis dataset aims to improve access to and understanding of French legal decisions, potentially benefiting legal research and analysis.\r\n\r\n### Discussion of Biases\r\n\r\nThe dataset may reflect inherent biases in the French legal system and case selection/publication processes.\r\n\r\n### Other Known Limitations\r\n\r\n- Limited to published decisions\r\n- Varying detail levels across sources\r\n- Potential OCR errors in source texts\r\n\r\n## Additional Information\r\n\r\n### Dataset Curators\r\n\r\nLa-Mousse\r\n\r\n### Licensing Information\r\n\r\nCC-BY-4.0\r\n\r\n### Citation Information\r\n\r\n```\r\n@misc{french-legal-cases-2024,\r\n    title={French Legal Cases Dataset},\r\n    author={[Your Name]},\r\n    year={2024},\r\n    publisher={HuggingFace},\r\n    url={https://huggingface.co/datasets/la-mousse/combined-fr-caselaw}\r\n}\r\n```\r\n\r\n### Contributions\r\n\r\nThanks to [@huggingface](https://github.com/huggingface) for the dataset hosting and infrastructure.","description_withheld":null,"homepage":"https://huggingface.co/datasets/Tonic/combined-fr-caselaw","introduced_date":"2025-01-20","introduced_date_note":null,"introduced_by":null,"license":{"name":"mit","url":null},"modalities":[],"tasks":[{"name":"Text Retrieval","url":"/task/text-retrieval","datasets_with_task":"/datasets/task/text-retrieval"}],"languages":[{"name":"French","url":"/datasets/language/french"}],"variants":["french-legal-cases"],"data_loaders":[{"repo":"https://github.com/huggingface/datasets","url":"https://huggingface.co/datasets/Tonic/combined-fr-caselaw","frameworks":["tf","pytorch","jax"]},{"repo":"https://github.com/huggingface/datasets","url":"https://huggingface.co/datasets/La-Mousse/combined-fr-caselaw","frameworks":["tf","pytorch","jax"]}],"num_papers_in_archive":0,"source":{"archive":"pwc-archive (Hugging Face), CC BY-SA 4.0","snapshot":"2025-07-28"},"benchmarks":[],"papers_with_a_benchmark_row":[],"syntology_totals":{"read_at":"2026-09-24T18:15:14+00:00","papers_with_samples":0,"samples_harvested":0,"samples_ran":0,"samples_unverified":0,"pointer_only_for_licence":0,"papers_with_no_sample_that_ran":0,"note":"the per-paper counts above, summed; not a rate"},"papers_note":"The archive never published its papers-using-dataset list; these are papers with a leaderboard row on this dataset's benchmarks."}