Upload README.md with huggingface_hub

README.md

@@ -1,13 +1,18 @@
 ---
+license: apache-2.0
 language:
 - fr
-license: apache-2.0
 multilinguality:
 - monolingual
-size_categories:
-- 1K<n<10K
+tags:
+- finetuning
+- legal
+- french law
+- droit français
+- Code général des impôts, annexe II
 source_datasets:
 - original
+pretty_name: Code général des impôts, annexe II
 task_categories:
 - text-generation
 - table-question-answering
@@ -15,222 +20,29 @@ task_categories:
 - text-retrieval
 - question-answering
 - text-classification
-pretty_name: Code général des impôts, annexe II
-tags:
-- finetuning
-- legal
-- french law
-- droit français
-- Code général des impôts, annexe II
-configs:
-- config_name: default
-  data_files:
-  - split: train
-    path: data/train-*
-dataset_info:
-  features:
-  - name: ref
-    dtype: string
-  - name: texte
-    dtype: string
-  - name: dateDebut
-    dtype: int64
-  - name: dateFin
-    dtype: int64
-  - name: num
-    dtype: string
-  - name: id
-    dtype: string
-  - name: cid
-    dtype: string
-  - name: type
-    dtype: string
-  - name: etat
-    dtype: string
-  - name: nota
-    dtype: string
-  - name: version_article
-    dtype: string
-  - name: ordre
-    dtype: int64
-  - name: conditionDiffere
-    dtype: 'null'
-  - name: infosComplementaires
-    dtype: 'null'
-  - name: surtitre
-    dtype: 'null'
-  - name: nature
-    dtype: string
-  - name: texteHtml
-    dtype: string
-  - name: dateFinExtension
-    dtype: int64
-  - name: versionPrecedente
-    dtype: string
-  - name: refInjection
-    dtype: string
-  - name: idTexte
-    dtype: 'null'
-  - name: idTechInjection
-    dtype: string
-  - name: origine
-    dtype: string
-  - name: dateDebutExtension
-    dtype: int64
-  - name: idEliAlias
-    dtype: 'null'
-  - name: cidTexte
-    dtype: 'null'
-  - name: sectionParentId
-    dtype: string
-  - name: multipleVersions
-    dtype: bool
-  - name: comporteLiensSP
-    dtype: bool
-  - name: sectionParentTitre
-    dtype: string
-  - name: infosRestructurationBranche
-    dtype: 'null'
-  - name: idEli
-    dtype: 'null'
-  - name: sectionParentCid
-    dtype: string
-  - name: numeroBo
-    dtype: 'null'
-  - name: infosRestructurationBrancheHtml
-    dtype: 'null'
-  - name: historique
-    dtype: 'null'
-  - name: infosComplementairesHtml
-    dtype: 'null'
-  - name: renvoi
-    dtype: 'null'
-  - name: fullSectionsTitre
-    dtype: string
-  - name: notaHtml
-    dtype: string
-  - name: inap
-    dtype: 'null'
-  splits:
-  - name: train
-    num_bytes: 2156110
-    num_examples: 672
-  download_size: 705389
-  dataset_size: 2156110
+size_categories:
+- 1K<n<10K
 ---
-# Code général des impôts, annexe II, non-instruct (2024-08-13)
+# Code général des impôts, annexe II, non-instruct (2024-08-14)
 
-This project focuses on fine-tuning pre-trained language models to create efficient and accurate models for legal practice. 
+The objective of this project is to provide researchers, professionals and law students with simplified, up-to-date access to all French legal texts, enriched with a wealth of data to facilitate their integration into Community and European projects.
 
-Fine-tuning is the process of adapting a pre-trained model to perform specific tasks or cater to particular domains. It involves adjusting the model's parameters through a further round of training on task-specific or domain-specific data. While conventional fine-tuning strategies involve supervised learning with labeled data, instruction-based fine-tuning introduces a more structured and interpretable approach.
+Normally, the data is refreshed daily on all legal codes, and aims to simplify the production of training sets and labeling pipelines for the development of free, open-source language models based on open data accessible to all.
 
-Instruction-based fine-tuning leverages the power of human-provided instructions to guide the model's behavior. These instructions can be in the form of text prompts, prompts with explicit task descriptions, or a combination of both. This approach allows for a more controlled and context-aware interaction with the LLM, making it adaptable to a multitude of specialized tasks.
+## Concurrent reading of the LegalKit
 
-Instruction-based fine-tuning significantly enhances the performance of LLMs in the following ways:
+[<img src="https://raw.githubusercontent.com/louisbrulenaudet/ragoon/main/assets/badge.svg" alt="Built with RAGoon" width="200" height="32"/>](https://github.com/louisbrulenaudet/ragoon)
 
-- Task-Specific Adaptation: LLMs, when fine-tuned with specific instructions, exhibit remarkable adaptability to diverse tasks. They can switch seamlessly between translation, summarization, and question-answering, guided by the provided instructions.
-- Reduced Ambiguity: Traditional LLMs might generate ambiguous or contextually inappropriate responses. Instruction-based fine-tuning allows for a clearer and more context-aware generation, reducing the likelihood of nonsensical outputs.
-- Efficient Knowledge Transfer: Instructions can encapsulate domain-specific knowledge, enabling LLMs to benefit from expert guidance. This knowledge transfer is particularly valuable in fields like tax practice, law, medicine, and more.
-- Interpretability: Instruction-based fine-tuning also makes LLM behavior more interpretable. Since the instructions are human-readable, it becomes easier to understand and control model outputs.
-- Adaptive Behavior: LLMs, post instruction-based fine-tuning, exhibit adaptive behavior that is responsive to both explicit task descriptions and implicit cues within the provided text.
+To use all the legal data published on LegalKit, you can use RAGoon:
+```bash
+pip3 install ragoon
+```
 
-## Concurrent reading of the LegalKit
+Then, you can load multiple datasets using this code snippet:
 
-To use all the legal data published on LegalKit, you can use this code snippet:
 ```python
 # -*- coding: utf-8 -*-
-import concurrent.futures
-import os
-
-import datasets
-from tqdm.notebook import tqdm
-
-def dataset_loader(
-    name:str,
-    streaming:bool=True
-) -> datasets.Dataset:
-    """
-    Helper function to load a single dataset in parallel.
-
-    Parameters
-    ----------
-    name : str
-        Name of the dataset to be loaded.
-
-    streaming : bool, optional
-        Determines if datasets are streamed. Default is True.
-
-    Returns
-    -------
-    dataset : datasets.Dataset
-        Loaded dataset object.
-
-    Raises
-    ------
-    Exception
-        If an error occurs during dataset loading.
-    """
-    try:
-        return datasets.load_dataset(
-            name,
-            split="train",
-            streaming=streaming
-        )
-
-    except Exception as exc:
-        logging.error(f"Error loading dataset {name}: {exc}")
-
-        return None
-
-
-def load_datasets(
-    req:list,
-    streaming:bool=True
-) -> list:
-    """
-    Downloads datasets specified in a list and creates a list of loaded datasets.
-
-    Parameters
-    ----------
-    req : list
-        A list containing the names of datasets to be downloaded.
-
-    streaming : bool, optional
-        Determines if datasets are streamed. Default is True.
-
-    Returns
-    -------
-    datasets_list : list
-        A list containing loaded datasets as per the requested names provided in 'req'.
-
-    Raises
-    ------
-    Exception
-        If an error occurs during dataset loading or processing.
-
-    Examples
-    --------
-    >>> datasets = load_datasets(["dataset1", "dataset2"], streaming=False)
-    """
-    datasets_list = []
-
-    with concurrent.futures.ThreadPoolExecutor() as executor:
-        future_to_dataset = {executor.submit(dataset_loader, name): name for name in req}
-
-        for future in tqdm(concurrent.futures.as_completed(future_to_dataset), total=len(req)):
-            name = future_to_dataset[future]
-
-            try:
-                dataset = future.result()
-
-                if dataset:
-                    datasets_list.append(dataset)
-
-            except Exception as exc:
-                logging.error(f"Error processing dataset {name}: {exc}")
-
-    return datasets_list
-
+from ragoon import load_datasets
 
 req = [
     "louisbrulenaudet/code-artisanat",
@@ -240,7 +52,7 @@ req = [
 
 datasets_list = load_datasets(
     req=req,
-    streaming=True
+    streaming=False
 )
 
 dataset = datasets.concatenate_datasets(
@@ -248,60 +60,64 @@ dataset = datasets.concatenate_datasets(
 )
 ```
 
-## Dataset generation
-
-This JSON file is a list of dictionaries, each dictionary contains the following fields:
-
-- `instruction`: `string`, presenting the instruction linked to the element.
-- `input`: `string`, signifying the input details for the element.
-- `output`: `string`, indicating the output information for the element.
-- `start`: `string`, the date of entry into force of the article. 
-- `expiration`: `string`, the date of expiration of the article. 
-- `num`: `string`, the id of the article.
-
-We used the following list of instructions for generating the dataset:
-```python
-instructions = [
-    "Compose l'intégralité de l'article sous forme écrite.",
-    "Écris la totalité du contenu de l'article.",
-    "Formule la totalité du texte présent dans l'article.",
-    "Produis l'intégralité de l'article en écriture.",
-    "Développe l'article dans son ensemble par écrit.",
-    "Génère l'ensemble du texte contenu dans l'article.",
-    "Formule le contenu intégral de l'article en entier.",
-    "Rédige la totalité du texte de l'article en entier.",
-    "Compose l'intégralité du contenu textuel de l'article.",
-    "Rédige l'ensemble du texte qui constitue l'article.",
-    "Formule l'article entier dans son contenu écrit.",
-    "Composez l'intégralité de l'article sous forme écrite.",
-    "Écrivez la totalité du contenu de l'article.",
-    "Formulez la totalité du texte présent dans l'article.",
-    "Développez l'article dans son ensemble par écrit.",
-    "Générez l'ensemble du texte contenu dans l'article.",
-    "Formulez le contenu intégral de l'article en entier.",
-    "Rédigez la totalité du texte de l'article en entier.",
-    "Composez l'intégralité du contenu textuel de l'article.",
-    "Écrivez l'article dans son intégralité en termes de texte.",
-    "Rédigez l'ensemble du texte qui constitue l'article.",
-    "Formulez l'article entier dans son contenu écrit.",
-    "Composer l'intégralité de l'article sous forme écrite.",
-    "Écrire la totalité du contenu de l'article.",
-    "Formuler la totalité du texte présent dans l'article.",
-    "Produire l'intégralité de l'article en écriture.",
-    "Développer l'article dans son ensemble par écrit.",
-    "Générer l'ensemble du texte contenu dans l'article.",
-    "Formuler le contenu intégral de l'article en entier.",
-    "Rédiger la totalité du texte de l'article en entier.",
-    "Composer l'intégralité du contenu textuel de l'article.",
-    "Rédiger l'ensemble du texte qui constitue l'article.",
-    "Formuler l'article entier dans son contenu écrit.",
-    "Quelles sont les dispositions de l'article ?",
-    "Quelles dispositions sont incluses dans l'article ?",
-    "Quelles sont les dispositions énoncées dans l'article ?",
-    "Quel est le texte intégral de l'article ?",
-    "Quelle est la lettre de l'article ?"
-]
-```
+### Data Structure for Article Information
+
+This section provides a detailed overview of the elements contained within the `item` dictionary. Each key represents a specific attribute of the legal article, with its associated value providing detailed information.
+
+1. **Basic Information**
+   - `ref` (string): **Reference** - A reference to the article, combining the title_main and the article `number` (e.g., "Code Général des Impôts, art. 123").
+   - `texte` (string): **Text Content** - The textual content of the article.
+   - `dateDebut` (string): **Start Date** - The date when the article came into effect.
+   - `dateFin` (string): **End Date** - The date when the article was terminated or superseded.
+   - `num` (string): **Article Number** - The number assigned to the article.
+   - `id` (string): **Article ID** - Unique identifier for the article.
+   - `cid` (string): **Chronical ID** - Chronical identifier for the article.
+   - `type` (string): **Type** - The type or classification of the document (e.g., "AUTONOME").
+   - `etat` (string): **Legal Status** - The current legal status of the article (e.g., "MODIFIE_MORT_NE").
+
+2. **Content and Notes**
+   - `nota` (string): **Notes** - Additional notes or remarks associated with the article.
+   - `version_article` (string): **Article Version** - The version number of the article.
+   - `ordre` (integer): **Order Number** - A numerical value used to sort articles within their parent section.
+
+3. **Additional Metadata**
+   - `conditionDiffere` (string): **Deferred Condition** - Specific conditions related to collective agreements.
+   - `infosComplementaires` (string): **Additional Information** - Extra information pertinent to the article.
+   - `surtitre` (string): **Subtitle** - A subtitle or additional title information related to collective agreements.
+   - `nature` (string): **Nature** - The nature or category of the document (e.g., "Article").
+   - `texteHtml` (string): **HTML Content** - The article's content in HTML format.
+
+4. **Versioning and Extensions**
+   - `dateFinExtension` (string): **End Date of Extension** - The end date if the article has an extension.
+   - `versionPrecedente` (string): **Previous Version** - Identifier for the previous version of the article.
+   - `refInjection` (string): **Injection Reference** - Technical reference to identify the date of injection.
+   - `idTexte` (string): **Text ID** - Identifier for the legal text to which the article belongs.
+   - `idTechInjection` (string): **Technical Injection ID** - Technical identifier for the injected element.
+
+5. **Origin and Relationships**
+   - `origine` (string): **Origin** - The origin of the document (e.g., "LEGI").
+   - `dateDebutExtension` (string): **Start Date of Extension** - The start date if the article has an extension.
+   - `idEliAlias` (string): **ELI Alias** - Alias for the European Legislation Identifier (ELI).
+   - `cidTexte` (string): **Text Chronical ID** - Chronical identifier of the text.
+
+6. **Hierarchical Relationships**
+   - `sectionParentId` (string): **Parent Section ID** - Technical identifier of the parent section.
+   - `multipleVersions` (boolean): **Multiple Versions** - Indicates if the article has multiple versions.
+   - `comporteLiensSP` (boolean): **Contains Public Service Links** - Indicates if the article contains links to public services.
+   - `sectionParentTitre` (string): **Parent Section Title** - Title of the parent section (e.g., "I : Revenu imposable").
+   - `infosRestructurationBranche` (string): **Branch Restructuring Information** - Information about branch restructuring.
+   - `idEli` (string): **ELI ID** - European Legislation Identifier (ELI) for the article.
+   - `sectionParentCid` (string): **Parent Section Chronical ID** - Chronical identifier of the parent section.
+
+7. **Additional Content and History**
+   - `numeroBo` (string): **Official Bulletin Number** - Number of the official bulletin where the article was published.
+   - `infosRestructurationBrancheHtml` (string): **Branch Restructuring Information (HTML)** - Branch restructuring information in HTML format.
+   - `historique` (string): **History** - Historical context or changes specific to collective agreements.
+   - `infosComplementairesHtml` (string): **Additional Information (HTML)** - Additional information in HTML format.
+   - `renvoi` (string): **Reference** - References to content within the article (e.g., "(1)").
+   - `fullSectionsTitre` (string): **Full Section Titles** - Concatenation of all titles in the parent chain.
+   - `notaHtml` (string): **Notes (HTML)** - Additional notes or remarks in HTML format.
+   - `inap` (string): **INAP** - A placeholder for INAP-specific information.
 
 ## Feedback