Datasets:

costadev00
/

wikipedia-pt-br-extract

Modalities:

Formats:

Languages:

Size:

Libraries:

License:

Dataset card Data Studio Files Files and versions

You need to agree to share your contact information to access this dataset

This repository is publicly accessible, but you have to accept the conditions to access its files and content.

Log in or Sign Up to review the conditions and access this dataset content.

Portuguese Wikipedia Extract Feb 2026

This dataset contains cleaned text extracted from the latest Portuguese Wikipedia dump.

Schema

Each row has:

text: cleaned plain text (no templates/tags)
title: page title
page_id: page id from the XML dump
ns: page namespace id
section_texts: list of cleaned section texts (lead included)

Notes

Empty text pages are skipped.
Redirect pages can be skipped depending on extraction flags.
Non-main namespaces are skipped by default.
Disambiguation pages are skipped by default.
List items are removed by default (use --keep-lists to keep them).
The dataset is provided in JSONL and Parquet formats.

Downloads last month: 13

Total file size: