ManikKatha — Manik Bandopadhyay Bengali Literary Corpus

Bengali literary prose by Manik Bandopadhyay (1908-1956).

Seven novels: Padma Nadir Majhi (1936), Dibaratrir Kabya (1936), Chatushkon (1942), Majhir Chhele (1959), and Sahartali, Ahingsa and Darpan, the three collected in the 1965 selection Sera Manik. Novels are split into their printed chapters.

Sixty-one short stories: all fifty-eight of Uttarkaler Galpa-Sangraha (2nd edition, 1964) in their published order, plus Atmahatyar Adhikar, Pragaitihasik and Sarisrip from Sera Manik. Each story is one whole record, and no story appears twice.

Every record carries its work, chapter or story title, the collection it was gathered from, the year of first publication where the source volumes record it, and the page-level OCR confidence scores that produced the text, so consumers can filter to the quality level their use case needs.

At a glance

records	105
works	68 (7 long-form, 61 short stories)
characters	2,762,155
words	468,623
tokens (est.)	814,076
language	bn (Beng)
author	মানিক বন্দ্যোপাধ্যায়
first published	1936–1963
editions scanned	1936–1965
mean OCR confidence	84.0 / 100
licence	cc0-1.0

Each record is one chapter of a novel or one complete short story, so records carry whole narrative arcs rather than fragments.

Text is uncorrected OCR. Rather than silently shipping the errors, every record carries the OCR confidence scores that produced it, so you can filter to whatever quality level your use case needs — see Filtering by quality below.

Contents

Long-form works

work	title (en)	first published	edition	chapters	chars	tokens	mean conf
দর্পণ	Darpan	—	1965	7	382,949	108,990	86.8
সহরতলী	Sahartali	—	1965	14	378,740	107,401	88.7
অহিংসা	Ahingsa	—	1965	19	331,341	92,944	88.5
পদ্মানদীর মাঝি	Padma Nadir Majhi	1936	1954	1	214,917	70,731	78.4
দিবারাত্রির কাব্য	Dibaratrir Kabya	1936	1936	1	203,296	58,576	84.2
চতুষ্কোণ	Chatushkon	1942	1942	1	186,120	50,186	91.2
মাঝির ছেলে	Majhir Chhele	1959	1959	1	135,704	37,668	90.1

Short stories

One record each; collection_bn / collection_en name the volume they were printed in.

work	title (en)	first published	edition	records	chars	tokens	mean conf
সরীসৃপ	Sarisrip	—	1965	1	47,788	14,154	88.3
কোন দিকে	Kon Dike	1963	1964	1	42,618	13,222	80.4
বাস	Bas	1944	1964	1	35,100	10,818	80.5
কুষ্ঠরোগীর বৌ	Kushtharogir Bou	1943	1964	1	25,594	8,055	78.1
গুপ্তধন	Guptadhan	1939	1964	1	25,513	8,109	78.6
প্রাগৈতিহাসিক	Pragaitihasik	—	1965	1	24,637	7,529	87.5
চক্রান্ত	Chakranta	1947	1964	1	24,355	7,252	81.5
আজ কাল পরশুর গল্প	Aj Kal Parshur Galpa	1943	1964	1	23,190	7,490	79.4
ছাঁটাই রহস্য	Chhantai Rahasya	1947	1964	1	22,508	6,710	82.0
প্রাণাধিক	Pranadhik	1949	1964	1	20,000	6,064	81.1
আত্মহত্যার অধিকার	Atmahatyar Adhikar	—	1965	1	19,587	5,893	89.0
চিকিৎসা	Chikitsa	1954	1964	1	17,872	5,219	82.7
একান্নবর্তী	Ekannabarti	1947	1964	1	17,572	5,474	80.1
সংঘাত	Sanghat	1953	1964	1	17,462	5,618	80.7
ছিনিয়ে খায়নি কেন	Chhiniye Khayni Keno	1947	1964	1	17,312	5,430	81.9
কংক্রীট	Concrete	1943	1964	1	17,282	5,647	81.3
হারাণের নাতজামাই	Haraner Natjamai	1948	1964	1	17,171	5,795	81.3
বিচার	Bichar	1950	1964	1	16,453	5,055	80.3
উপায়	Upay	1963	1964	1	16,389	5,177	80.5
ছেলেমানুষি	Chhelemanushi	1948	1964	1	16,054	4,994	80.0
ধান	Dhan	1948	1964	1	15,118	4,640	81.5
মাসীপিসী	Masipisi	1943	1964	1	14,787	4,759	80.8
প্রাণের গুদাম	Praner Gudam	1943	1964	1	14,577	4,313	82.8
ছেঁড়া	Chhenra	1943	1964	1	14,289	4,707	81.8
সখী	Sakhi	1949	1964	1	14,204	4,349	80.7
গায়েন	Gayen	1948	1964	1	14,167	4,459	82.6
দুঃশাসনীয়	Duhshasaniya	1943	1964	1	14,066	4,463	80.3
যে বাঁচায়	Je Banchay	1944	1964	1	14,025	4,141	79.9
টিচার	Teacher	1947	1964	1	13,925	4,254	80.6
নমুনা	Namuna	1943	1964	1	13,810	4,366	80.5
প্যাক	Pyak	1939	1964	1	13,780	4,367	78.3
পেট ব্যথা	Pet Byatha	1943	1964	1	13,640	4,267	82.0
এক বাড়িতে	Ek Barite	1953	1964	1	13,321	4,008	80.1
ছোট বকুলপুরের যাত্রী	Chhoto Bakulpurer Jatri	1949	1964	1	13,117	4,123	79.5
চালক	Chalak	1947	1964	1	13,096	4,115	80.1
শিল্পী	Shilpi	1943	1964	1	13,070	4,223	81.2
সুবালা	Subala	1954	1964	1	12,816	4,037	81.4
পারিবারিক	Paribarik	1948	1964	1	12,789	3,897	82.4
এদিক ওদিক	Edik Odik	1954	1964	1	12,725	3,935	79.2
ফেরিওলা	Feriwala	1953	1964	1	12,498	3,966	79.0
কলহান্তরিত	Kalahantarita	1954	1964	1	12,423	3,759	81.4
রাঘব মালাকর	Raghab Malakar	1943	1964	1	12,007	3,945	78.7
ঢেউ	Dheu	1957	1964	1	11,720	3,642	80.7
মেজাজ	Mejaj	1949	1964	1	11,142	3,376	82.0
বাগদীপাড়া দিয়ে	Bagdipara Diye	1948	1964	1	10,999	3,418	78.9
পাষণ্ড	Pashanda	1954	1964	1	10,994	3,238	82.0
মহাকর্কট বটিকা	Mahakarkat Batika	1953	1964	1	10,411	3,169	80.2
দীঘি	Dighi	1948	1964	1	10,339	3,290	80.9
একটি বখাটে ছেলের কাহিনী	Ekti Bakhate Chheler Kahini	1963	1964	1	10,060	3,108	80.0
খতিয়ান	Khatian	1947	1964	1	9,904	3,038	81.3
ধর্ম	Dharma	1948	1964	1	9,449	2,896	79.2
মীমাংসা	Mimangsa	1954	1964	1	8,821	2,658	84.1
নিরুদ্দেশ	Niruddesh	1954	1964	1	8,735	2,615	82.1
কালোবাজারের প্রেমের দর	Kalobajarer Premer Dar	1957	1964	1	8,437	2,485	80.5
শত্রুমিত্র	Shatrumitra	1943	1964	1	7,990	2,479	81.1
অসহযোগী	Asahajogi	1954	1964	1	7,463	2,313	81.0
আপদ	Apad	1948	1964	1	7,222	2,235	80.3
গোপাল শাসমল	Gopal Shasmal	1943	1964	1	6,479	2,078	81.3
উপদলীয়	Upadaliya	1954	1964	1	6,055	1,675	82.3
যাকে ঘুস দিতে হয়	Jake Ghus Dite Hay	1943	1964	1	6,044	1,827	79.4
রক্ত নোনতা	Rakta Nonta	1956	1964	1	4,117	1,242	80.0

Record size distribution

	chars
smallest	4,117
median	15,118
90th percentile	46,594
largest	214,917

Sources

Every record traces back to a scanned printed edition on the Internet Archive.

source PDF	Internet Archive item	pages	works	status
2015.453204.Padmanadir-Majhi_text.pdf	in.ernet.dli.2015.453204	173	1	ready
10689.22151_text.pdf	dli.bengal.10689.22151	214	1	ready
2015.455811.Chatushkon-Ed_text.pdf	in.ernet.dli.2015.455811	105	1	ready
2015.455159.Majhir-Chhele_text.pdf	in.ernet.dli.2015.455159	94	1	ready
2015.456423.Sera-Manik_text.pdf	in.ernet.dli.2015.456423	547	6	ready
2015.456488.Uttarkaler-Galposangraha_text.pdf	in.ernet.dli.2015.456488	459	58	ready

Schema

column	type	meaning
text	string	the training field: NFC Bengali prose, paragraphs split by \n\n
id	string	stable record id, <dataset>/<work>/<unit>
work_bn / work_en / work_id	string	the work this record belongs to
collection_bn / collection_en	string	set for stories in a collection, else null
unit_type	string	chapter or story
unit_index	int32	position within the work
n_chars / n_words / n_tokens_est	int32	size; n_tokens_est is null if no tokenizer was configured
first_published / edition_year	int32	year of the work vs. of the scanned printing
ia_identifier / source_url / source_pdf	string	provenance
ocr_engine / ocr_params / ocr_source	string	how the text was produced
conf_mean	float32	mean OCR word confidence (0-100), null if unknown
conf_p10	float32	10th-percentile page confidence — catches local damage a mean hides
low_conf_word_frac	float32	fraction of words scored below 70
pages	list<struct>	per source page: page_index, printed_page, char_start, char_end, conf_mean, low_conf_word_frac

conf_* columns are null, never 0.0, when confidence is genuinely unknown — a fabricated zero would read as worst-possible quality and silently drop good text.

Filtering by quality

from datasets import load_dataset

ds = load_dataset("protibimbo/ManikKatha", split="train")

# 1. drop whole records that are mostly noise
good = ds.filter(lambda r: (r["conf_mean"] or 0) >= 75.0)

# 2. or excise only the bad pages, keeping the rest of the chapter
def drop_bad_pages(rec, floor=70.0):
    keep = [rec["text"][p["char_start"]:p["char_end"]]
            for p in rec["pages"] if (p["conf_mean"] or 0) >= floor]
    return {"text": "".join(keep)}

clean = ds.map(drop_bad_pages)

The pages spans are contiguous and tile text exactly, so slicing and rejoining them never loses or duplicates characters.

Intended use

Continued pretraining and LoRA/PEFT style tuning on next-token prediction — the records are raw prose, not instruction pairs. Text is not pre-chunked to a token window: packing to 2048/4096 with an EOS at document boundaries is a training-time decision, and pre-chunking would sever sentences and take that choice away. n_tokens_est is provided so you can plan packing without tokenizing first.

There is a single train split. For style evaluation, hold out a work yourself (filtering on work_id) so the held-out text is genuinely unseen.

# hold out one novel for evaluation
train = ds.filter(lambda r: r["work_id"] != "aranyak")
eval_ = ds.filter(lambda r: r["work_id"] == "aranyak")

How the text was produced

1. Extract — text and per-word confidence are read from the Internet Archive hOCR derivative of each scan. The PDF's own invisible text layer carries the same characters but no confidence, so hOCR is preferred wherever an item exists.
2. Normalise — Unicode NFC. Zero-width joiners are preserved: in Bengali they carry the reph / ya-phala distinction, so stripping them would change words.
3. Strip page furniture — running headers and page numbers are removed. The same running title is OCR'd differently on nearly every page, so headers are detected by fuzzy-clustering candidate lines rather than by exact match.
4. Reflow — printed line breaks are rejoined into paragraphs, including across page boundaries. Paragraphs are separated by \n\n.
5. Segment — split into works by verified page ranges, then into chapters by printed headings.

Limitations

• OCR is uncorrected (tesseract 5.0.0-alpha-20201231-10-g1236, tesseract 5.3.0-6-g76ae). Mean record confidence is 84.0/100, and 0 of 105 records fall below 75. Expect wrong conjuncts, stray Latin characters and mangled punctuation. The text is good enough to learn style and syntax from, and not a reliable edition to quote from.
• Front matter, editorial prefaces and critical appendices were excluded by page range, but header and page-number removal is heuristic and a few survive mid-text.
• Paragraph boundaries are reconstructed from printed line breaks and are approximate; dialogue segments more reliably than continuous narration.
• Chapter splits come from printed headings. Books whose headings did not survive OCR appear as a single long record rather than being split arbitrarily.
• Scans are of mid-century reprints, not first editions, so spelling reflects the printing house's conventions of the day.
• Single author and single genre: this is literary prose from one writer, and is not a general-purpose Bengali corpus.

Provenance and rights

Scans and OCR come from the Internet Archive; see ia_identifier and source_url on each record, and the Sources table above.

Cataloguing note: Internet Archive/DLI author fields for this collection are not always reliable. Attribution here was checked against each scan's own title page, not taken from the metadata record.

Public domain in India: the author died in 1956, so copyright expired in 2016 under life + 60.

US status is unresolved. Manik's career began in 1928 and every work here was first published in 1929 or later, so the URAA may have restored US copyright in all of it — there is no date filter that separates a clear subset from a risky one, as there is for this project's Tagore corpus. This is published as a known and accepted risk, not as a cleared position; confirm before relying on it.

first_published is set on 62 of 105 records. For the four standalone novels it is taken from the edition's own imprint. For the 58 Uttarkaler Galpa-Sangraha stories it is the year the story first appeared in book form, taken from the গল্প পরিচয় in that volume; a story may have appeared in a periodical earlier. It is null for Sahartali, Ahingsa, Darpan and the three stories taken from Sera Manik, whose scan records no dates.