PiotrSty/tatoeba-czech

Tatoeba Czech sentences (tatoeba_cs) - community example sentences

Czech sentences from the weekly Tatoeba sentences_detailed export (downloaded 2026-10-05, ISO 639-3 ces rows). Sentence-level corpus: single community-contributed example sentences with per-record author attribution (username + dates).

• Source rows (ces): 89,961 of 13,597,683 export rows
• Retained records: 61,122 (67.9%)
• Measured cl100k_base proxy tokens: 984,041
• Characters: 2,044,211
• Time span (date added): 2009-11-23 - 2026-10-02
• Source snapshot: sentences_detailed.tar.bz2 (sha256 0af302eebda1d7c5773c505b7b6fa5bb41690eb8bcb31cf213a07caf04a88ba3, weekly export current through 2026-10-11; Tatoeba exports refresh Saturdays 06:30 UTC)

License - evidence at the origin

• Verbatim Tatoeba downloads page (https://tatoeba.org/en/downloads): "These files are released under CC BY 2.0 FR. A part of our sentences are also available under CC0 1.0."
• Per-record license field: CC-BY-2.0-FR; per-record attribution in artifacts/attribution.jsonl.gz: sentence id, contributing username, date added, date modified, source URL (https://tatoeba.org/en/sentences/view/<id>) and text_sha256. Contributor usernames are preserved as attribution.

Samples

Verbatim records from artifacts/sample.jsonl.

Sample 1

id: tatoeba_cs_1625298
source: tatoeba_cs
created: 2012-06-13
token_count: 14

Moje navigace nefunguje v zahraničí.

Sample 2

id: tatoeba_cs_1824136
source: tatoeba_cs
created: 2012-09-07
token_count: 15

Žije v Indii v nějaké vesnici.

Sample 3

id: tatoeba_cs_2465965
source: tatoeba_cs
created: 2013-05-30
token_count: 50

Podle této zprávy z Číny je základní problém v rozdílu mezi ekonomickým vývojem a vývojem společenslým.

Filtering

Sentence-level record model (one record per sentence). Rejections (28,839 records, one final decision per record in artifacts/decisions.jsonl.gz):

{
  "non_czech_text": 7839,
  "too_little_text": 21000
}

Minimum final text is 20 characters (short fragments and non-sentence rows drop); language gate langid == cs after the upstream ISO 639-3 ces filter (7,839 rows re-classified as non-Czech - short or ambiguous sentences). PII redaction (labelled patterns: e-mail, labelled phone, labelled IP ("IP adresa"), labelled national identifier ("rodné číslo"), CZ IBAN candidates): 0 pattern matches, 0 residual documents across all patterns. Exact dedup on 128-bit truncated sha256; near dedup (Jaccard >= 0.90 on capped 25-shingle sets) is inert for sentence-length records and documented as such. No boilerplate lines exceeded the 1% document-frequency threshold (max bin df 2).

Known limitations

• Sentence-level corpus: single example sentences, often translationese; not documents or conversations.
• Community-contributed content of uneven quality and mixed register.
• Contributor usernames remain as attribution (per the license).
• Near-dedup does not fire on sentence-length rows.

See artifacts/ for attribution, per-record decisions, QA counts, samples and checksums. verify() passes: 61,122 rows, unique IDs, token counts recomputed and text_sha256 re-checked against the emitted text. Tests: 10 (pytest).