KabLiterary

A 28,907-paragraph, 657,713-word corpus of classical world literature in canonical Kabyle Latin orthography, from the AƔBALU project.

Eleven works — translated or originally composed in Kabyle — spanning epic poetry, gothic fiction, philosophical prose, folktales, military strategy, and 19th-century novels. The corpus is designed for language modelling, vocabulary probing, fine-tuning, and literary translation benchmarking where long-form, high-register Kabyle text is required.

from datasets import load_dataset

ds = load_dataset("agbalu/KabLiterary")
# DatasetDict({
#     'train': Dataset(27619),
#     'dev': Dataset(352),
#     'test': Dataset(936)
# })

Splits

Split boundaries are at the book level. No paragraph from a dev or test book appears in train.

split	books	paragraphs	words	characters	purpose
train	9 books	27,619	630,454	3,837,504	Pre-training, instruction tuning, MLM
dev	The Prophet (Nnbi)	352	4,243	24,240	Hyperparameter tuning, dev loss
test	Alice in Wonderland (Tirmit n Alice)	936	23,016	138,756	Held-out perplexity, generation eval
total	11 books	28,907	657,713	4,000,500	Full literary corpus

Corpus Inventory

Book	Kabyle Title	Author	Year	Source	Genre	Split	Paragraphs	Words
The Count of Monte Cristo	Lkont n Monte-Cristo	Alexandre Dumas	1844	French	Adventure fiction	train	4,362	106,000
Les Misérables	Imeɣban	Victor Hugo	1862	French	Historical fiction	train	2,862	96,315
The Confessions	Isteɛṛufen n Uɣustin Aqeddis	Augustine of Hippo	397	English (Latin)	Theology / Philosophy	train	1,336	91,652
Dracula	Dracula, Amudir n Yiḍ	Bram Stoker	1897	English	Gothic horror	train	2,239	133,380
The Arabian Nights	Alf Yiḍ d Yiḍ	Anonymous / A. Lang	1898	English (Arabic)	Folklore tales	train	1,737	85,944
The Odyssey	Asikel n Ulises	Homer / S. Butler	-800	English (Greek)	Epic poetry	train	14,427	101,584
The Art of War	Taẓuri n Ṭṭraḍ	Sun Tzu / L. Giles	1910	English (Chinese)	Military philosophy	train	410	9,803
The Fables of Bidpai	Timucuha n Bidpai	Bidpai / Ibn al-Muqaffa	1884	English (Sanskrit)	Fables & folklore	train	106	5,013
Poems of Si Mohand	Isefra n Si Mohand u Mhand	Si Mohand u Mhand	1904	Kabyle (original)	Classical poetry	train	140	763
The Prophet	Nnbi	Kahlil Gibran	1923	English	Poetic philosophy	dev	352	4,243
Alice in Wonderland	Tirmit n Alice deg Tmurt n Lɛeǧayeb	Lewis Carroll	1865	English	Children's fiction	test	936	23,016

Isefra n Si Mohand u Mhand is the single original Kabyle work — oral poetry transcribed by Mouloud Mammeri and not derived from any source language. All other works are machine-translated from the source editions listed above.

Schema

Every record shares a fixed rectangular schema:

field	type	description
id	string	Unique record identifier (e.g. kabliterary_train_00001)
book_id	string	Canonical book slug (e.g. le-comte-de-monte-cristo, alice-in-wonderland)
book_title	string	Standard title in English or original language
book_title_kab	string	Title in Kabyle Latin orthography
author	string	Author name
year	int32	Original publication year (negative for BC dates)
genre	string	Literary genre classification
source_language	string	ISO 639-3 code of the source language (fra, eng, kab)
target_language	string	ISO 639-3 code of the target language (kab)
chapter	string	Chapter or section title
paragraph_id	int32	1-indexed paragraph counter within the work
text	string	Kabyle literary text in canonical Latin orthography
source_text	string	Aligned source text where 1-to-1 line parallel; empty string elsewhere
word_count	int32	Word count of the Kabyle text
char_count	int32	Character count of the Kabyle text

source_text is populated only for works where line-level alignment was preserved during translation — The Art of War, The Prophet, and Isefra n Si Mohand u Mhand. For the longer prose novels, paragraphs were batched during translation and source_text is empty.

Usage

Default splits

from datasets import load_dataset

dataset = load_dataset("agbalu/KabLiterary")
print(dataset["test"][4]["text"])
# 'Tella Alice tebda teɛya aṭas seg akken teqqim ɣer tama n uletma-s...'

Filtering by work or genre

from datasets import load_dataset

train = load_dataset("agbalu/KabLiterary", split="train")

# French 19th-century novels only
french_novels = train.filter(
    lambda row: row["book_id"] in ("le-comte-de-monte-cristo", "les-miserables")
)
print(f"{len(french_novels):,} paragraphs, {sum(french_novels['word_count']):,} words")

# Records with 1-to-1 source alignment
parallel = train.filter(lambda row: len(row["source_text"]) > 0)
print(f"{len(parallel):,} line-level parallel records")

Full plain-text books

All 11 works are included in the repository's books/ directory:

git clone https://huggingface.co/datasets/agbalu/KabLiterary
ls KabLiterary/books/
cat KabLiterary/books/the-prophet.txt

Curation

Translation

The ten translated works were produced by NLLB-based machine translation and subjected to a multi-stage quality pass before inclusion:

1. Repetition pruning. 276+ beam-search generation loops were detected and removed across all ten texts. No tail-degenerate paragraph survives in any split.
2. Heading localisation. Residual French chapter headings in Les Misérables and other works were translated into standard Kabyle Latin (e.g. L'évêque travaille → Aselway ixeddem, L'Alouette → Tiwiziwt).
3. Source-language artifact removal. Any paragraph retaining a title or quotation in English or French was re-translated or manually corrected to Kabyle.

Orthography

All text is normalised to AƔBALU canonical Kabyle Latin orthography (ɣ ɛ ḥ ḍ ṣ ṭ ẓ ṛ č ǧ). A census across all 28,907 records confirmed:

• zero Greek epsilon homoglyphs (ε U+03B5)
• zero soft hyphens (\u00ad)
• NFC normalisation throughout

HTML entities (" and variants) were repaired to standard punctuation during the quality pass.

Known Limitations

• Machine-translated syntax. Ten of eleven works are machine-translated. The text is fluent and orthographically correct, but phrase-level structure carries influence from the French or English source. No human post-editing was performed beyond the artifact and heading corrections described above.
• No 1-to-1 paragraph alignment in prose novels. Longer works (Monte-Cristo, Les Misérables, Dracula, The Arabian Nights) were translated with multi-sentence batching. Paragraph boundaries do not map to individual source sentences.
• Archaic and specialised vocabulary. Works such as Confessions of Saint Augustine and The Odyssey require theological and classical vocabulary that has no established Kabyle equivalent. The translations use loan-adapted morphological frames; the decisions are internally consistent but not reviewed against a Kabyle lexicographic standard.
• Single source edition per work. Each translation was produced from one source edition. Variant readings, abridgements, or multi-edition differences were not considered.

Reproduction

make release-kabliterary

The dataset is assembled by tools.build_kabliterary from the verified texts in artifacts/translations/ and data/documents/.

Citation

@misc{agbalu_kabliterary,
  title  = {KabLiterary: A 657K-Word Corpus of Classical Literary Translations in Kabyle},
  author = {AƔBALU},
  year   = {2026},
  url    = {https://huggingface.co/datasets/agbalu/KabLiterary}
}

Licence

CC-BY-SA-4.0. Source texts are public domain (Project Gutenberg and digitised archives). Translations and the assembled dataset are produced by the AƔBALU project under CC-BY-SA-4.0. Attribution is required; derivative datasets must carry the same licence.

Part of AƔBALU, a Kabyle NLP corpus and model suite.