Til-Books-Adebiportal

Kazakh book texts from adebiportal.kz (the Kazakh literary portal). Part of the TilQazyna Kazakh-LLM corpus program. Each book's PDF was fetched (via the portal viewer's viewer.php?file= endpoint) and its text extracted with PyMuPDF.

Schema

One row per book: slug, book_id, url, title, text, n_chars, source (source=adebiportal).

Usage

from datasets import load_dataset
ds = load_dataset("TilQazyna/Til-Books-Adebiportal", split="train")

~2,255 books, mostly Kazakh literature. Digital text extraction (no OCR). For training, combine with the other Til-Books-* sources; quality scoring/tiering is applied in the unified Til-Books / Til-Corpus flagships. Non-commercial Kazakh-language research; respect adebiportal.kz and author rights.