Alkaloids-dataset

This dataset contains 5,813 open-access scientific articles related to alkaloids. Articles were sourced from PubMed Central and ChemRxiv, focusing on literature suitable for NLP and QA system development in the chemistry domain. All PMC articles were filtered using "free full text" and downloaded only if a PMCID and open PDF were available. ChemRxiv articles were accessed through the official API and include only those with downloadable open-access PDFs.

Original PDF files are not included in this dataset due to file size and licensing constraints. Instead, each article is provided in plain-text format. Texts were extracted using the PyMuPDF library and lightly cleaned for consistency. Metadata is included for each article in JSON format, listing title, authors, publication year, source, DOI, and file paths.
The dataset is structured as follows:
1) metadata.json: metadata for all articles
2) extracted_texts/ :plain-text versions extracted from PDFs
3) README.md
The dataset supports research in question answering, information retrieval, summarization, and chemical entity recognition. This dataset is for non-commercial academic research only.