Add CC BY 4.0 Polish OpenStax academic textbooks

#23
Fabryka AI org
•
edited 11 days ago

Add Polish OpenStax academic prose

Adds openstax_pl: 1,422 documents and 4,632,374 measured cl100k_base proxy tokens from eight Polish textbook volumes (physics ×3, psychology, microeconomics, macroeconomics, marketing, nutrition). The complete discovered snapshot has 1,612 pages. Each page has an explicit retained or rejected decision.

Source: PiotrSty/openstax-pl-textbooks. Pinned payload: 5b31f74eb3733b330c6093dabc919fe304f6ef6f. License: CC BY 4.0, stated on each Polish edition's own foreword page (links below).

Translation status

These are Polish translations of the English OpenStax books (original US authors cited per book — e.g. Ling / Moebs / Sanny for physics), and the translations are human work by named academics, credited in the preserved forewords (and per record in attribution.jsonl: edition_contributors):

  • Mikroekonomia / Makroekonomia — "Tłumaczenia: dr Bożena Horbaczewska, Szkoła Główna Handlowa w Warszawie (rozdziały: …); dr Krystian Jaworski, SGH (rozdziały: …)" — per-chapter credits;
  • Marketing — scientific editing by Lidia Danik (dr hab., prof. SGH), named chapter translations and content additions, Adam Oleksiuk (dr, SGH) credited;
  • Psychologia — translators Anna Czechowska, Nika Fronc-Iniewicz, Magdalena Kroenke, Monika Olasek, Magdalena Skrzypek, Magdalena Szwedowska, Małgorzata Załoga (technical editing: Patryk Bojarski);
  • Fizyka t. 1–3 — the foreword describes the Polish translation of the OpenStax textbook made for the OpenStax Poland effort.

No machine translation is claimed or indicated anywhere in the source material.

License at the origin — per book

The CC BY 4.0 grant lives on each edition's foreword page (verified live; the pages and their excerpts are preserved in books.json):

Each foreword states the grant ("…jest udostępniana na licencji Creative Commons Uznanie autorstwa 4.0 Międzynarodowe", linking https://creativecommons.org/licenses/by/4.0/deed.pl) with the attribution condition "uznania autorstwa OpenStax i osób podpisanych pod oryginałem".

Sample (verbatim, from the built shard)

Physics chapter (TeX formulas preserved):

id: openstax_pl_eb84605e23aa69df43d1da80c0c64d4582500948814f4eaba3d688f5e83e9167
source: openstax_pl
added: 2026-09-05
created: 2017-12-05
token_count: 4708
license: CC-BY-4.0
author: William Moebs; Samuel J. Ling; Jeff Sanny; OpenStax Poland and the Polish edition contributors (see foreword)

10.8

Praca i energia kinetyczna w ruchu obrotowym

Cel dydaktyczny

W tym podrozdziale nauczysz się:

stosować twierdzenie o równoważności pracy i energii do analizy ruchu obrotowego i obliczania, jaką pracę należy wykonać nad układem, aby obrócić go o określony kąt wokół stałej osi;

wyznaczać, wykorzystując twierdzenie o równoważności pracy i energii, prędkość kątową obracającej się bryły sztywnej;

wyznaczać moc dostarczaną obracającej się bryle sztywnej, biorąc pod uwagę moment obrotowy siły i prędkość kątową;

weryfikować uzyskane wartości zmiennych kątowych i zastosowane równania;

wskazywać związek otrzymanych wyników i użytych praw z odpowiednimi zmiennymi i prawami w ruchu postępowym.

Do tej pory szeroko omówiliśmy zagadnienia kinematyki i dynamiki ruchu obrotowego brył […]

Marketing chapter:

id: openstax_pl_807a8d3897c384e9e42344d4b2697f718583f5dd85ca7c973b4de8e51a496a46
source: openstax_pl
added: 2026-09-05
created: 2024-11-13
token_count: 8078
license: CC-BY-4.0
author: Maria Gomez Albrecht; Mark Green; Linda Hoffman; OpenStax Poland and the Polish edition contributors (see foreword)

13.2

Miks promocyjny i jego elementy

Cel dydaktyczny

Po przeczytaniu tego podrozdziału będziesz w stanie:

1
zdefiniować pojęcie miksu promocyjnego i wyjaśnić jego znaczenie,

2
wymienić i opisać elementy miksu promocyjnego.

Czym jest miks promocyjny?

Wszyscy jesteśmy konsumentami, dlatego każdy z nas jest celem licznych kampanii promocyjnych. Aby nawiązać kontakt ze swoimi klientami, firmy przeprowadzają obszerne badania rynku, starając się lepiej zrozumieć ich zachowania. Dzięki temu mogą odpowiednio skonstruować komunikat, który przemówi do konsumenta, a także znaleźć najlepsze kanały dotarcia do niego.

Po określeniu i zdefiniowaniu rynku docelowego, stworzeniu dobrego produktu, wyborze strategii cenowej i optymalnej ceny oraz zdecydowaniu o metodzie dystrybucji marketer jest […]

Microeconomics chapter (math in TeX):

id: openstax_pl_a9c14b43eb96ae77e34d3095f7c02e062b305782ac067a2c1dfa3abf4f43285e
source: openstax_pl
added: 2026-09-05
created: 2022-09-07
token_count: 1554
license: CC-BY-4.0
author: Steven A. Greenlaw; David Shapiro; Waldemar Karpa; OpenStax Poland and the Polish edition contributors (see foreword)

6.4

Produkcja w długim okresie

Cel dydaktyczny

Po przeczytaniu tego podrozdziału będziesz w stanie:

Zrozumieć różnice między funkcją produkcji w krótkim a funkcją produkcji w długim okresie

W długim okresie wszystkie czynniki produkcji (łącznie z kapitałem) są zmienne, a więc nasza funkcja produkcji ma postać:
$Q = f \left[\right. L , K \left]\right.$
Rozważmy sytuację przedsiębiorstwa, które zajmuje się przygotowywaniem korespondencji handlowej w różnych językach na zlecenie. Firma zatrudnia maszynistki, które będziemy traktować jako czynnik produkcji – praca. Wykorzystują one komputery, które są kapitałem przedsiębiorstwa. Na początku firma ma niewiele zleceń, tak że tylko jedna maszynistka i jeden komputer są zajęci przez cały dzień. Powiedzmy, że w ciągu jednego dnia firma jest w […]

Normalization — steps and before/after

src/fetch_openstax_pl.py: extract() (paragraph-preserving extraction from the Rex/Connexions HTML containers — markup, frontmatter, indices, "Bibliografia" and answer keys excluded; media omitted; MathML → TeX via TeX annotations / alttext / mathml-to-latex, pages with conversion failures rejected) -> normalize() (NFKC; zero-width spaces and soft hyphens removed) -> pattern PII redaction -> gates (every page language-checked; exact dedup on normalized text; seeded MinHashLSH + exact 5-word-shingle Jaccard ≥ 0.9). Before/after (preserved container openstax_pl_4a74bcce…, 63,041 → 15,431 chars):

BEFORE (source HTML container, first lines)

<div class="chapter-content-module" data-book-content="true" data-type="page" id="page_c9ba6155-…" xmlns:cmlnle="http://katalysteducation.org/cmlnle/1.0"><h2 data-rex-keep="true" data-type="document-title" id="…"><span class="os-number">1.2</span>
<span class="os-divider"> </span>
<span class="os-text" data-type="" itemprop="">Układy jednostek miar</span>
</h2><div class="ui-has-child-title" data-type="abstract"><header><h3 data-type="title">Cel dydaktyczny</h3></header><section>W tym podrozdziale nauczysz się:
<ul>
<li>definiować podstawowe jednostki układu SI;</li>
<li>tworzyć jednostki pochodne na podstawie jednostek podstawowych;</li>
<li>korzystać z przedrostków jednostek miar układu SI.</li>
</ul></section></div>
<p id="fs-id1168329295259">Jak już wiemy, zakres zainteresowań fizyki obejmuje wiele przeróżnych obiektów i zjawisk. […]

AFTER (shipped record text)

1.2

Układy jednostek miar

Cel dydaktyczny

W tym podrozdziale nauczysz się:

definiować podstawowe jednostki układu SI;

tworzyć jednostki pochodne na podstawie jednostek podstawowych;

korzystać z przedrostków jednostek miar układu SI.

Jak już wiemy, zakres zainteresowań fizyki obejmuje wiele przeróżnych obiektów i zjawisk. […]

(Figure <img> blocks, os-* scaffolding and term-markup spans disappear — figure captions/alt text are not kept as prose; headings, lists, paragraphs and TeX survive.)

PII — which scrubber ran

Pattern redaction in src/fetch_openstax_pl.py (e-mail -> [REDACTED:EMAIL], labelled phone -> [REDACTED:PHONE]) — not SlayerLab/NERGAL. Spot check (60 of 1,422 records, seed 20260928): zero hits across all residual classes (placeholders, e-mail-like, digit runs, IBAN-like, street cues, phone words) — textbook chapters carry author/translator credits (published attribution, not PII). Per AGENTS.md: no release before a NERGAL pass; merge with release: None is fine meanwhile.

Registry and line endings (review items 1)

Rebased onto main @ 34e7ea3; the src/sources.py diff is exactly the registry entry — +21 / −0, no whole-file rewrite. The entry now carries added: "2026-09-05" and release: None. Line-ending note in the review thread (main's blob is CRLF today; matching it is what keeps the diff minimal — a repo-wide LF normalization is maintainer work).

Data sample links and validation (original)

24 complete deterministic sample documents, three per book. All retained text replays from preserved source HTML containers; checks cover schema, record counts, hashes, attribution IDs, token totals and book coverage. Every retained page has a language classification. Normalized exact and seeded MinHash candidate / exact Jaccard deduplication are applied within the source. The Slayer graph records content-addressed source and dataset versions, the processing protocol and actual run, typed lineage, Evidence, falsifiable Claims, Actors and evidence-backed status.

Integration limits

Corpus-wide exact/near dedup and benchmark overlap checks remain pending; source-key absence does not establish textual novelty. Equations/tables and exercises referencing omitted figures are documented limitations. Please run the target-wide gates and test a capped educational share before a stable release. This PR contributes the source and its review evidence, not a new stable release.

Test plan

  • python3 -m pytest src/test_openstax_contribution.py — passed on the rebased branch
  • Before/after extraction excerpt (above)
  • Residual-PII spot check (above — clean)
  • License verified per book at the origin (foreword pages, table above)
  • Reviewer: translation status statement accepted
Fabryka AI org

Review against AGENTS.md ("What the PR must show", "Normalize the text before you build"). Needed before merge:

  1. Line endings. src/sources.py is committed with CRLF, so it shows as a whole-file rewrite (the real change is 13 lines) and will conflict with every other source PR. Please convert to LF.
  2. License evidence is currently your own books.json. Link the OpenStax Polska license statement for each book.
  3. Translation. State whether the texts are translations of English OpenStax books, and by whom (human / MT).
  4. Normalization — list the steps and show a before/after excerpt.
  5. Sample inline. Paste 2–3 real documents into the description, one fenced block each: metadata as key: value lines, a blank line, then the text with its real line breaks, truncated with […] (format as in #24). A linked .sample.jsonl renders on the Hub as one escaped line per document and can't be reviewed at a glance.
  6. PII. AGENTS.md requires SlayerLab/NERGAL. The source can merge with release: None after another scrubber, but it won't be admitted to a release until the shard is re-run through NERGAL. State in the datasheet which scrubber ran.
Fabryka AI org

Review addressed — pushed as a rebase + review-fix commit on refs/pr/23, onto main @ 34e7ea3. Point by point:

  1. Line endings. The src/sources.py diff is now exactly the registry entry: +21 / −0 — no whole-file rewrite. Measurement note: git cat-file blob origin/main:src/sources.py is CRLF today (533/533 CR); the old PR file was an old-base CRLF copy, which produced the rewrite. Matching main's bytes minimizes the diff — converting this file to LF against a CRLF main would recreate the whole-file rewrite. A repo-wide LF normalization of sources.py is maintainer work (happy to send it as its own PR). The entry also now carries added: "2026-09-05" and release: None per AGENTS.md.
  2. License at the origin. Per-book license statements are now linked in the description table and datasheet — each edition's own foreword page on openstax.org (e.g. https://openstax.org/books/fizyka-dla-szkół-wyższych-tom-1/pages/przedmowa, verified live: "…jest udostępniana na licencji Creative Commons Uznanie autorstwa 4.0 Międzynarodowe", deed.pl link, attribution condition "uznania autorstwa OpenStax i osób podpisanych pod oryginałem") plus the openstax.pl catalogue page per book. books.json remains the preserved copy of those statements (excerpt + sha256), not the primary evidence.
  3. Translation. Stated in the description and datasheet: these are Polish translations of the English OpenStax originals (original US authors cited per book), made by named human academics — per-chapter credits preserved in the forewords and in attribution.jsonl: edition_contributors: dr Bożena Horbaczewska (SGH) and dr Krystian Jaworski (SGH) for the economics volumes (chapter lists in the forewords), scientific editing by Lidia Danik (prof. SGH) + named chapter translations/content additions (Adam Oleksiuk credited) for Marketing, and a 7-person translator team (Anna Czechowska, Nika Fronc-Iniewicz, Magdalena Kroenke, Monika Olasek, Magdalena Skrzypek, Magdalena Szwedowska, Małgorzata Załoga) for Psychologia. No MT is claimed or indicated anywhere in the source material.
  4. Normalization. Datasheet lists the steps: extract() (paragraph-preserving extraction from the preserved Rex/Connexions HTML containers — markup, frontmatter, indices, "Bibliografia" and answer keys excluded; media omitted; MathML → TeX via TeX annotations / alttext / mathml-to-latex, pages with conversion failures rejected) -> normalize() (NFKC; zero-width spaces and soft hyphens removed) -> pattern PII redaction -> gates (every page language-checked; exact dedup on normalized text; MinHashLSH + exact 5-word-shingle Jaccard ≥ 0.9). Before/after excerpt in the description (openstax_pl_4a74bcce…, 63,041 → 15,431 chars): os-* scaffolding, figure <img> blocks and term-markup spans are gone; headings, lists, paragraphs and TeX survive.
  5. Sample inline. 3 real documents in the description, one fenced block each (key: value metadata, blank line, real line breaks, […] truncation, format as in #24): physics chapter (TeX preserved), marketing chapter, microeconomics chapter (math in TeX).
  6. PII. Datasheet names the scrubber: pattern redaction in src/fetch_openstax_pl.py (e-mail -> [REDACTED:EMAIL], labelled phone -> [REDACTED:PHONE]) — not NERGAL. Spot check (60 of 1,422 records, seed 20260928): zero hits across all residual classes (placeholders, e-mail-like, digit runs, IBAN-like, street cues, phone words) — the only names are author/translator credits. Agreed: no release before a NERGAL pass; merge with release: None is fine meanwhile.

Verification: python3 -m pytest src/test_openstax_contribution.py — passed on the rebased branch.

ppuzio changed pull request status to merged
ppuzio deleted the refs/pr/23 ref

Sign up or log in to comment