sayurio/bangla-kobita-scrape-bangla-literature

--- license: mit task_categories: - text-generation language: - bn tags: - bengali - poetry - literature - web-scraped - non-ai pretty_name: Bangla Kobita Poetry Archive size_categories: - 100K<n<1M --- # Bangla Kobita Poetry Archive ## Overview This repository contains a curated text dataset of Bengali poetry scraped from the web, primarily targeting comprehensive poetry platforms like [bangla-kobita.com](https://www.bangla-kobita.com/). The primary goal of this archive is to preserve a rich collection of purely human-written Bengali poems (Bangla Kobita), creating a distinct record of human artistic expression, emotion, and linguistic rhythm separate from AI-generated text. ## Purpose and Usage This dataset is published publicly and strictly for **educational, research, linguistic analysis, and archival purposes**. It is a highly specialized resource for Natural Language Processing (NLP) researchers, data scientists, and linguists looking to: * Pre-train or fine-tune Bengali Large Language Models (LLMs) on creative, metaphorical, and emotionally complex text. * Study poetic structures, rhyme schemes, and metrical patterns in modern and classic Bengali literature. * Perform sentiment analysis and emotion detection on deeply expressive artistic writing. ## Dataset Details * **Source:** bangla-kobita.com (and/or related Bengali poetry archives) * **Collection Method:** Web scraping * **Content Type:** Text (Bengali poems, including titles, poet names, and stanza-formatted text written by humans without the use of AI). * **Repository:** `sayurio/bangla-kobita-scrape-bangla-literature` ## Copyright and Fair Use Disclaimer This archive is created under the principles of **Fair Use** (under Section 107 of the Copyright Act) for purposes such as criticism, comment, teaching, scholarship, and research. * **No Ownership Claimed:** The creator of this repository does not claim any ownership, authorship, or copyright over the original poems. All rights, title, and interest in the original text remain entirely with their respective poets, authors, and the publishing platforms. * **Non-Commercial:** This dataset is provided completely free of charge and is strictly not intended for commercial gain, monetization, or profit. * **Transformative Use:** The data has been aggregated, extracted from its original web formatting, and compiled specifically for computational analysis, archiving, and educational study. This represents a transformative use of the original publicly available artistic material. **Takedown Requests:** If you are a copyright holder or a poet whose work is included in this dataset and wish for it to be removed from this archive, please open an issue or contact the repository owner directly. Please submit a removal request specifying the exact poem titles, URLs, or text snippets you wish to have taken down so they can be accurately located within the dataset and removed. ## How to Use You can load this dataset directly into your Python environment using the Hugging Face `datasets` library: ```python from datasets import load_dataset # Load the dataset dataset = load_dataset("sayurio/bangla-kobita-scrape-bangla-literature") # View the structure of the first poem entry print(dataset['train'][0]) ```--- 许可证：MIT许可证 任务类别： - 文本生成 语言： - 孟加拉语（bn） 标签： - 孟加拉语 - 诗歌 - 文学 - 网络爬取 - 非AI生成 美观名称：孟加拉语诗歌档案（Bangla Kobita Poetry Archive） 规模类别：10万<n<100万条数据 --- # 孟加拉语诗歌档案（Bangla Kobita Poetry Archive） ## 数据集概览 本仓库收录了经整理的孟加拉语诗歌文本数据集，数据主要从[bangla-kobita.com](https://www.bangla-kobita.com/)等综合性诗歌平台网络爬取所得。本档案的核心目标是留存一批纯人类创作的孟加拉语诗歌（Bangla Kobita）的丰富馆藏，构建一份独立于AI生成文本的、关于人类艺术表达、情感与语言韵律的专属记录。 ## 用途与使用规范 本数据集公开发布，且仅可用于**教育、研究、语言分析及档案留存**用途。它是面向自然语言处理（Natural Language Processing, NLP）研究者、数据科学家与语言学家的专业资源，可用于以下场景： * 在富有创造性、隐喻性与复杂情感的文本语料上，对孟加拉语大语言模型（Large Language Models, LLMs）进行预训练或微调； * 研究现当代孟加拉语文学中的诗歌结构、押韵规则与格律模式； * 对极具表现力的艺术文本开展情感分析与情绪检测任务。 ## 数据集详情 * **数据来源：** bangla-kobita.com（及/或相关孟加拉语诗歌档案库） * **采集方式：** 网络爬取 * **内容类型：** 文本（人类未借助AI创作的孟加拉语诗歌，包含诗歌标题、诗人姓名与分节格式的正文） * **仓库地址：** `sayurio/bangla-kobita-scrape-bangla-literature` ## 版权与合理使用声明 本档案依据**合理使用原则**（符合《版权法》第107条）创建，用于评论、评述、教学、学术研究等场景。 * **无版权主张：** 本仓库的创建者不对原诗歌主张任何所有权、著作权或版权。原始文本的全部权利、所有权与权益均归属于原诗人、作者及发布平台。 * **非商业性质：** 本数据集完全免费提供，且严格禁止用于商业获利、变现或牟利行为。 * **转型性使用：** 本数据已从原始网页格式中提取、聚合并整理，专门用于计算分析、档案留存与教育研究，属于对公开可获取的原创艺术素材的转型性使用。 **下架请求：** 若您是本数据集收录作品的版权持有者或原诗人，希望将其从本档案中移除，请提交Issue或直接联系仓库所有者。提交移除请求时，请明确注明需移除的诗歌标题、链接或文本片段，以便我们在数据集中准确定位并删除对应内容。 ## 使用方法 您可通过Hugging Face的`datasets`库，直接在Python环境中加载本数据集： python from datasets import load_dataset # Load the dataset dataset = load_dataset("sayurio/bangla-kobita-scrape-bangla-literature") # View the structure of the first poem entry print(dataset['train'][0])