samorzad_gov_pl: drop 144 third-party rows, relabel licence basis (v0.2.6, v2#49)

artifacts/source_findings.md

@@ -1290,3 +1290,71 @@ reserved" in their text (publication years 2010-2021); regex count only, not rev
 
 **Lesson** — a platform's "public domain" label can be the PD Mark. Check the URL behind
 the label, and check the date: a 2019 work cannot be public domain by term in Poland.
+
+## govpl — PAP reprints dropped (2026-10-06)
+
+**Verdict** — `kept`, minus 188 PAP reprints (slayerlabs/polish-dynaword-v2#51).
+88,190 → 88,002 documents, 80,666,129 → 80,445,758 tokens (−220,371). `release` stays
+`0.2.2`. `src/clean_govpl.py` filters the pinned v0.2.5 Parquet, so ids are unchanged,
+and writes `govpl.attribution.jsonl` with each row's gov.pl URL, subsite, title and date.
+`fetch_govpl.normalize()` applies the same rule (`pap_rule`) to future crawls.
+
+**Reason** — gov.pl's copyright page grants CC BY-SA 4.0 for its text but carves out
+works whose copyright belongs to third parties; their use needs the owners' consent.
+Articles that credit PAP as their source are such works.
+- The issue's rule (`Źródło:` line naming PAP) matches 154 rows. A hand-check of 20
+  (seed 51) found 20 text source lines and no photo credits; 14 credit PAP jointly with
+  the ministry, so the text may be partly the ministry's. PAP's archive is not public, so
+  verbatim PAP authorship was not confirmed.
+- A scan of the other rows added 34 verified text attributions: misspelt or colon-less
+  source lines (9), `Na podst. … PAP` (22, all uw-podlaski), a lone `(PAP)` sign-off line
+  (2) and one PAP interview transcript (prokuratura-krajowa). Photo credits (`fot. PAP`)
+  and rows that quote or mention PAP stay.
+- Left in, flagged: a mid-text `(PAP)` after one paragraph (govpl_30227, govpl_58869), a
+  footnote crediting PAP and the health ministry (govpl_81037), a KRRiT report quoting two
+  PAP questions (govpl_86914).
+- Not reviewed: 710 kept rows carry a `Źródło:` line naming some other source (mostly
+  government bodies; some name WHO, universities, journals).
+
+References: `rights_audit_2026-10/reports/govweb.md` (govpl block);
+https://www.gov.pl/web/gov/prawa-autorskie (read 2026-10-06).
+
+**Lesson** — an open-licence footer covers the publisher's own text. On a news-style
+site, check source lines for agency copy before taking the footer at face value.
+
+## samorzad_gov_pl — third-party slices dropped (2026-10-06)
+
+**Verdict** — kept, relabelled, 144 rows dropped: 72,472 → 72,328 documents,
+40,011,099 → 39,890,244 tokens (baseline: main after #46, not the v0.2.5 tag's
+72,771). The basis is now called an operator-declared platform licence. A question
+to KPRM is pending (polish-dynaword-v2 #49); its answer may still change the
+source's status. The drops are in `src/clean_samorzad_gov_pl.py`
+(`third_party_drop`, applied by `--refilter-release` to the pinned release):
+- 69 "Piątek z poezją" posts (all Powiat Sokołowski): poems credited to their
+  authors. All 69 hand-checked.
+- 9 IMGW-PIB notices carrying "Opracowanie niniejsze i jego format…", which forbids
+  changes and requires naming IMGW-PIB. All 9 hand-checked.
+- 66 articles whose `Źródło:` line credits a non-governmental source (Wikipedia,
+  media, associations including volunteer fire brigades, foundations, clubs, rural
+  housewives' circles, companies, named books or brochures), unless a photo or film
+  marker precedes it on the line. All 66 hand-checked as third-party text. The rule
+  was revised on the same matches (first pass 61 of 67 correct), so precision is
+  in-sample and recall is unknown; one known miss is a Fundacja Santander credit.
+  Part of the regex lists work titles found in that review.
+- Not dropped: contest entries. The issue's 112 is not reproducible; the closest
+  regex (awarded/winning/distinguished works) gives 111 documents, and a 40-document
+  sample (seed 49) contains no entry text: they are results announcements. Only 3
+  documents publishing entries were found; most entries are images, not text.
+
+**Reason** — gov.pl's copyright page says the service also carries third-party works
+whose use requires the owners' consent; the footer licence cannot cover them.
+Credits matter here: 810 documents have an `Źródło:` line, but in 405 every credit is
+only a de-identified `<URL>`, so their sources cannot be classified; 231 documents
+mention IMGW without the verbatim notice and were not reviewed.
+
+References: polish-dynaword-v2 #49;
+`rights_audit_2026-10/reports/govweb.md` L8-15; `reports/citations/samorzad_nonsa.md`
+S1, S8, C4; <https://www.gov.pl/web/gov/prawa-autorskie> (read 2026-10-05).
+
+**Lesson** — de-identification that replaces URLs also removes the provenance needed
+to screen reposts. Record the credited domain before masking URLs.

data/govpl/govpl.attribution.jsonl

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:ff137d2042ff32198ec7eb28347631dbf21735ab27afc2d823f8c27d0904aa33
+size 26318408

data/govpl/govpl.md

@@ -8,34 +8,46 @@ gov.pl — Polish government press releases
 - **Language:** Polish (pl)
 - **License:** `CC-BY-SA-4.0`
 - **Created (range):** 2005-09-26, 2026-07-19
-- **Added:** 2026-07-19
+- **Added:** 2026-07-19; PAP reprints removed 2026-10-06
 
 ## Licensing — traceable basis
-All textual content on gov.pl is published under CC-BY-SA 4.0 (site footer: "Treści tekstowe publikowane w serwisie … są udostępniane na licencji CC BY-SA 4.0"). Authored and hosted by the Polish government, which holds the right to license it — same open family as Polish Wikipedia.
+gov.pl publishes its text under CC BY-SA 4.0 (site footer and https://www.gov.pl/web/gov/prawa-autorskie: "Treści tekstowe publikowane w serwisie … są udostępniane na licencji CC BY-SA 4.0"). The same page excludes works whose copyright belongs to third parties; using those needs the owners' consent. Articles crediting PAP are therefore dropped (legal note).
 
 ## Provenance
-Ministry/agency press releases fetched directly from gov.pl (editor-content per article, HTML stripped) by src/fetch_govpl.py. Subsites + their non-uniform news-section paths are enumerated by src/discover_govpl.py into the committed manifest src/govpl_subsites.json (gov.pl has no sitemap/API). Each record's meta.url resolves on gov.pl.
+Ministry/agency press releases fetched directly from gov.pl (editor-content per article, HTML stripped) by src/fetch_govpl.py. Subsites + their non-uniform news-section paths are enumerated by src/discover_govpl.py into the committed manifest src/govpl_subsites.json (gov.pl has no sitemap/API). v0.2.6 removes PAP reprints from the v0.2.5 Parquet with src/clean_govpl.py, which also writes govpl.attribution.jsonl: each row's gov.pl URL, subsite, title and date.
+
+## Attribution
+CC BY-SA 4.0 requires credit and a link to the source. `govpl.attribution.jsonl` has one
+line per row, in Parquet order: `id`, `url` (the gov.pl article), `subsite` (the publishing
+ministry or agency), `title`, `date` (as listed on gov.pl) and `license`. Rows were matched
+to the 2026-07-19 crawl on SHA-1 of the stripped text; for 1,189 rows the same text was
+fetched from more than one URL, and the first in fetch order is given.
 
 ## Statistics
 | documents | characters | tokens (tiktoken proxy) |
 |---:|---:|---:|
-| 88,190 | 218,099,000 | 80,666,129 |
-
+| 88,002 | 217,515,240 | 80,445,758 |
 
 ## Per-document license metadata
 | license | documents |
 |---|---:|
-| `CC-BY-SA-4.0` | 88,190 |
+| `CC-BY-SA-4.0` | 88,002 |
 
 Author metadata present for **0** documents. Empty values mean the upstream record did not expose a machine-readable author field.
 
-## Filters applied (build_dynaword.py)
-Minimal, per Dynaword guidelines (heavy filtering left to downstream use):
+## Legal scope note
+gov.pl's copyright page carves out "utwory, co do których prawa autorskie przysługują osobom trzecim" from its CC BY-SA 4.0 grant. Articles crediting PAP (Polish Press Agency) as their source are such works, so 188 are dropped (rule fetch_govpl.pap_rule, applied by src/clean_govpl.py): a "Źródło:" line naming PAP (163), "Na podst. … PAP" (22), a "(PAP)" sign-off line (2) and one PAP interview transcript (1). PAP photo credits and articles that only quote or mention PAP stay. Other third-party text is not screened: 710 kept articles carry a "Źródło:" line naming some other source; they were not reviewed.
+
+## Filters applied
+`build_dynaword.py` (v0.2.2), minimal per Dynaword guidelines:
 - drop documents < 200 chars: **0**
 - drop non-Polish (diacritic ratio): **24**
 - exact cross-source dedup (sha1): **1,926**
-- OCR alpha-ratio < 0.70 (OCR sources only): **0**
 - read 90,140 → kept 88,190
 
+`src/clean_govpl.py` (v0.2.6), third-party carve-out:
+- drop PAP reprints: **188** (`Źródło:` line 163, `Na podst.` 22, `(PAP)` sign-off 2, interview 1)
+- read 88,190 → kept 88,002
+
 Token counts are a fast tiktoken (cl100k) proxy (~1% off Llama-3); the canonical
 Llama-3 count is computed at release.

data/govpl/govpl.parquet

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:9c0310fcdf45804e6547bd197cba55f58f12df618cd3d8b92b8e9077a3720b9f
-size 82940575
+oid sha256:5e97d5b1ec04aebadaddd9ecb9fc6835324885606f59d4e8569f5516528f839b
+size 82308595

data/govpl/govpl.stats.json

@@ -1,16 +1,25 @@
 {
   "read": 90140,
-  "kept": 88190,
   "drop_short": 0,
   "drop_lang": 24,
   "drop_dup": 1926,
   "drop_ocr": 0,
-  "chars": 218099000,
-  "tokens": 80666129,
+  "drop_pap": 188,
+  "drop_pap_by_rule": {
+    "source_line": 163,
+    "sign_off": 2,
+    "based_on": 22,
+    "interview": 1
+  },
+  "kept": 88002,
+  "chars": 217515240,
+  "tokens": 80445758,
   "licenses": {
-    "CC-BY-SA-4.0": 88190
+    "CC-BY-SA-4.0": 88002
   },
   "authors_with_value": 0,
-  "secs": 2.9,
-  "license": "CC-BY-SA-4.0"
-}
+  "license": "CC-BY-SA-4.0",
+  "input_rows": 88190,
+  "input_sha256": "9c0310fcdf45804e6547bd197cba55f58f12df618cd3d8b92b8e9077a3720b9f",
+  "fetched_sha256": "368259d0ea8522eced71b426ae238556d5330a410aa994bdcbdc7427478d783a"
+}

data/samorzad_gov_pl/NOTICE.md

@@ -4,11 +4,25 @@ The `samorzad_gov_pl` source contains textual articles collected from public
 pages hosted on the shared [samorzad.gov.pl](https://samorzad.gov.pl/)
 platform.
 
-The platform footer states that textual content published in the service,
-excluding audiovisual content, is available under the [Creative Commons
+The licence basis is an operator-declared platform licence: the platform
+footer states that textual content published in the service, excluding
+audiovisual content, is available under the [Creative Commons
 Attribution-ShareAlike 4.0 International](https://creativecommons.org/licenses/by-sa/4.0/)
 license (`CC-BY-SA-4.0`). Every page in the published v0.2 source dataset belongs
-to the v0.1 source set for which the shared footer reference was verified.
+to the v0.1 source set for which the shared footer reference was verified. The
+agreements between the platform operator and the publishing institutions are
+not published, and no per-institution grant was found. A question to the
+operator (KPRM) is pending
+([slayerlabs/polish-dynaword-v2#49](https://github.com/slayerlabs/polish-dynaword-v2/issues/49));
+its answer may change this source's status.
+
+gov.pl's copyright page states that the service also carries works whose
+copyright belongs to third parties and that their use requires the owners'
+consent (<https://www.gov.pl/web/gov/prawa-autorskie>, read 2026-10-05). The
+platform licence does not cover such works. Rows identified as third-party
+works (Friday poetry posts, IMGW-PIB notices, text credited to a
+non-governmental source) are excluded; third-party text that is not detectable
+from the article text may remain.
 
 For every released DynaWord document:
 
@@ -17,8 +31,8 @@ For every released DynaWord document:
   fetched `final_url`, recommended `attribution_url`, article title,
   institution, tenant identifier, license URL and modification notice;
 - article text was extracted from source HTML and de-identified upstream;
-- this contribution normalizes metadata and excludes short, non-Polish or
-  duplicate records.
+- this contribution normalizes metadata and excludes short, non-Polish,
+  duplicate or third-party records.
 
 When redistributing or adapting these texts, retain the source attribution,
 indicate changes and distribute adaptations under compatible

data/samorzad_gov_pl/samorzad_gov_pl.attribution.jsonl

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:2def7488c34e749204842bcf21d33149738b34de2515115e3f79401f7eb271c4
-size 72438302
+oid sha256:a9c7dc3d049a26fc17fab86ab96b271e94a29159f38cb57ace373dcfef1fab80
+size 73230755

data/samorzad_gov_pl/samorzad_gov_pl.md

@@ -13,7 +13,8 @@ institutions and other public bodies.
 - **Original platform:** https://samorzad.gov.pl/
 - **Domain:** local government, education, social services and culture
 - **Language:** Polish (`pl`)
-- **License:** `CC-BY-SA-4.0`
+- **License:** `CC-BY-SA-4.0`, an operator-declared platform licence (see
+  [Licensing](#licensing--operator-declared-platform-licence))
 - **Created:** 2020–2026 for dated records; most source records do not expose a date
 - **Added:** 2026-08-14
 
@@ -21,7 +22,7 @@ institutions and other public bodies.
 
 | read | kept | characters | tokens (`cl100k_base` proxy) | publishers |
 |---:|---:|---:|---:|---:|
-| 81,418 | 72,472 | 107,130,266 | 40,011,099 | 247 |
+| 81,418 | 72,328 | 106,816,031 | 39,890,244 | 247 |
 
 The released Parquet uses the canonical DynaWord schema:
 
@@ -56,14 +57,24 @@ release gates passed.
 
 `src/clean_samorzad_gov_pl.py` verifies and converts the v0.2 Parquets into the
 DynaWord schema, applies the common minimum-length and Polish-language checks,
-removes exact duplicate text and counts tokens with `cl100k_base`. Retained text
-is otherwise unchanged after upstream de-identification.
-
-## Licensing — traceable basis and attribution
-
-The platform's shared footer states that textual content published in the
-service, excluding audiovisual content, is available under CC BY-SA 4.0. A
-release-time audit verified the live shared footer script for all 81,422 pages
+removes exact duplicate text, drops three third-party slices (see
+[Filters applied](#filters-applied)) and counts tokens with `cl100k_base`.
+Retained text is otherwise unchanged after upstream de-identification.
+
+## Licensing — operator-declared platform licence
+
+The licence is declared by the platform operator, not by the authors. The
+shared footer of the platform, injected by a template script of the gov.pl
+platform operator (KPRM; formerly Ministry of Digital Affairs), states that
+textual content published in the service, excluding audiovisual content, is
+available under CC BY-SA 4.0. The articles are written by the 247 publishing
+institutions. The agreements those institutions sign with the operator are not
+published, and no document granting the operator a licence from them was found.
+The maintainer has asked KPRM whether the institutions authorise the footer
+licence ([slayerlabs/polish-dynaword-v2#49](https://github.com/slayerlabs/polish-dynaword-v2/issues/49)).
+That answer is pending and may change this source's status.
+
+A release-time audit verified the live shared footer script for all 81,422 pages
 in the v0.1 source set. The four records omitted from v0.2 were a subset of that
 audited set, so every v0.2 article remains covered by the same evidence. The
 evidence URL, byte count and SHA-256 digest are recorded in
@@ -76,6 +87,15 @@ publishing institution, platform tenant, license URL and modification notice.
 For 688 retained records the discovered and final URLs differ because the
 platform redirected the request; `attribution_url` uses the fetched final URL.
 
+**Third-party works.** gov.pl's copyright page states that the service also
+carries works whose copyright belongs to third parties and that their use
+requires the owners' consent (https://www.gov.pl/web/gov/prawa-autorskie, read
+2026-10-05). The platform licence does not reach those works. This release drops
+the third-party slices that a committed rule can identify (see
+[Filters applied](#filters-applied)). Third-party text can remain where nothing
+in the record marks it, e.g. contest entries by pupils or reposts whose source
+line was reduced to a `<URL>` placeholder by de-identification.
+
 Recommended attribution:
 
 > Publishing institution listed in the `author` column; article title and
@@ -96,12 +116,37 @@ not redistribute those files.
 | shorter than 200 characters | 6,932 |
 | below the Polish-language heuristic | 34 |
 | exact duplicate text | 1,980 |
-| retained | 72,472 |
+| third party: "Piątek z poezją" poems | 69 |
+| third party: IMGW-PIB notices | 9 |
+| third party: non-governmental "Źródło:" credit | 66 |
+| retained | 72,328 |
 
 Exact duplicate removal is performed after short-text and language checks.
 Existing DynaWord sources take priority when the same text is already present
 elsewhere in the corpus.
 
+The third-party drops run after deduplication and are defined by
+`third_party_drop()` in `src/clean_samorzad_gov_pl.py`. Every dropped document
+was read by hand before release.
+
+- **"Piątek z poezją" poems:** the title contains `Piątek z poezją`. All 69 are
+  from Powiat Sokołowski and reprint poems credited to named authors.
+- **IMGW-PIB notices:** the text contains `Opracowanie niniejsze i jego format`.
+  These meteorological warnings allow further sharing only verbatim with
+  IMGW-PIB named, which CC BY-SA's permission to adapt cannot honour
+  (9 documents from Gmina Bodzechów, Gmina Ciężkowice and Gmina Leżajsk).
+- **Non-governmental "Źródło:" credit:** a line's first `Źródło:` ("Source:")
+  credits a non-public author: Wikipedia, newspapers, radio and television,
+  associations (including volunteer fire brigades), foundations, clubs, rural
+  housewives' circles, companies, or a named book, memoir or brochure. A credit
+  preceded on the same line by a photo, illustration or film marker (`Fot.`,
+  `Ilustracja`, `film`) credits the image, not the text, and is not matched.
+  Credits to public bodies, including associations of local governments, are
+  not matched. 810 documents have an `Źródło:` line: in 405 every credit is
+  only a de-identified `<URL>` placeholder, which cannot be classified; 66 of
+  the 405 with a named credit are dropped. The 339 kept named credits were not
+  individually reviewed.
+
 ## Limitations and recommended use
 
 - Automated de-identification reduces but cannot guarantee removal of every
@@ -109,8 +154,11 @@ elsewhere in the corpus.
   their use case.
 - The collection is not a complete historical archive of samorzad.gov.pl. It
   contains the pages found and successfully extracted during one crawl.
-- After filtering, 51,964 retained records have an empty `created` value; no
+- After filtering, 51,834 retained records have an empty `created` value; no
   date is inferred for them.
+- Third-party text that carries no title, notice or source line naming its
+  author is not detected. This includes most contest entries and reposts whose
+  `Źródło:` line holds only a `<URL>` placeholder.
 - The source text can contain spelling errors, extraction remnants or captions.
 - Very short source records are excluded by DynaWord's 200-character minimum.
 - Institution names describe the publishing body, not necessarily the person
@@ -152,6 +200,19 @@ uv run --with pyarrow --with tiktoken python src/clean_samorzad_gov_pl.py \
   --added 2026-08-14
 ```
 
+The third-party drops were applied to the release above in place with:
+
+```bash
+uv run --with pyarrow python src/clean_samorzad_gov_pl.py --refilter-release
+```
+
+`--refilter-release` refuses to run unless the Parquet and sidecar in
+`data/samorzad_gov_pl/` match the SHA-256 values of the pre-carve-out release
+(`PRE_CARVE_OUT_SHA256`). It drops the matching rows, recomputes the row-derived
+statistics and writes through the same staged, validated replacement as a full
+build. A full build applies the same `third_party_drop()` at the same point,
+after deduplication.
+
 Use `--refresh-publishers` only when intentionally updating the frozen
 institution-name manifest from live tenant homepages. Review any changed name
 before publishing a new release.

data/samorzad_gov_pl/samorzad_gov_pl.parquet

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:3a683b4220b9b70698d7c524637d9ea5c1d72630b8f92ea7a871642a719cf2c0
-size 49612214
+oid sha256:a9a43a5d6f2407033e95af132ef4d6996ae98ab73f553b9261b6f18a734ae4db
+size 49498053

data/samorzad_gov_pl/samorzad_gov_pl.stats.json

@@ -1,22 +1,27 @@
 {
-  "authors_with_value": 72472,
+  "authors_with_value": 72328,
   "central_tenant_count": 1,
-  "chars": 107130266,
+  "chars": 106816031,
   "cross_source_hashes_loaded": 4255011,
   "drop_dup": 1980,
   "drop_lang": 34,
   "drop_ocr": 0,
   "drop_short": 6932,
-  "kept": 72472,
+  "drop_third_party": {
+    "imgw_notice": 9,
+    "nongov_source": 66,
+    "poetry_friday": 69
+  },
+  "kept": 72328,
   "license": "CC-BY-SA-4.0",
   "licenses": {
-    "CC-BY-SA-4.0": 72472
+    "CC-BY-SA-4.0": 72328
   },
   "local_public_tenant_count": 246,
   "publisher_count": 247,
   "read": 81418,
   "redirected_urls": 688,
-  "tokens": 40011099,
+  "tokens": 39890244,
   "upstream_artifact_count": 10,
   "upstream_dataset": "dawidmajewski/samorzad-gov-pl-articles",
   "upstream_deidentification_model": "ArkadiuszPawlak/fastpdn-ner-polish-pii",

src/clean_govpl.py

@@ -0,0 +1,126 @@
+#!/usr/bin/env python3
+"""Drop PAP reprints from `govpl` and write a per-row URL sidecar.
+
+gov.pl's copyright page (https://www.gov.pl/web/gov/prawa-autorskie) licenses the
+site's text under CC BY-SA 4.0 but excludes works whose copyright belongs to third
+parties. Articles that credit PAP (Polish Press Agency) as their source are such
+works (slayerlabs/polish-dynaword-v2#51). `fetch_govpl.pap_rule` names them; this
+script removes them from the released Parquet, so ids and the other rows stay as they
+are (a rebuild through build_dynaword.py would renumber ids).
+
+`govpl.attribution.jsonl` gives each kept row the gov.pl URL, subsite, title and date
+of the fetched article, matched on build_dynaword's dedup key (SHA-1 of the stripped
+text) in the fetch output `govpl.jsonl.zst`. Where the same text was fetched from
+several URLs, the first one in fetch order is used.
+
+Both inputs are pinned by SHA-256. Restore the Parquet before a rerun:
+  git checkout v0.2.5 -- data/govpl/govpl.parquet
+  python3 src/clean_govpl.py --fetched ~/speakleash/govpl.jsonl.zst
+"""
+from __future__ import annotations
+
+import argparse
+import hashlib
+import io
+import json
+import os
+import sys
+from collections import Counter
+from pathlib import Path
+
+import pyarrow.parquet as pq
+import zstandard
+
+sys.path.insert(0, str(Path(__file__).resolve().parent))
+from fetch_govpl import KEY as SOURCE, pap_rule  # one rule for fetch and release
+
+ROOT = Path(__file__).resolve().parents[1]
+DATA = ROOT / "data" / SOURCE
+INPUT_SHA256 = "9c0310fcdf45804e6547bd197cba55f58f12df618cd3d8b92b8e9077a3720b9f"  # v0.2.5
+FETCHED_SHA256 = "368259d0ea8522eced71b426ae238556d5330a410aa994bdcbdc7427478d783a"  # 2026-07-19 crawl
+BUILD_STATS = {  # build_dynaword.py counts for the v0.2.5 Parquet, kept for the datasheet
+    "read": 90140, "drop_short": 0, "drop_lang": 24, "drop_dup": 1926, "drop_ocr": 0,
+}
+
+
+def text_key(text: str) -> bytes:
+    """build_dynaword's dedup key."""
+    return hashlib.sha1(text.strip().encode("utf-8")).digest()
+
+
+def fetched_meta(path: Path) -> dict[bytes, dict]:
+    """text_key -> meta of the first fetched record with that text."""
+    meta: dict[bytes, dict] = {}
+    with path.open("rb") as compressed:
+        lines = io.TextIOWrapper(zstandard.ZstdDecompressor().stream_reader(compressed), encoding="utf-8")
+        for line in lines:
+            if line.strip():
+                record = json.loads(line)
+                meta.setdefault(text_key(record["text"]), record["meta"])
+    return meta
+
+
+def check_pin(path: Path, expected: str, restore: str) -> None:
+    with path.open("rb") as stream:
+        actual = hashlib.file_digest(stream, "sha256").hexdigest()
+    if actual != expected:
+        raise SystemExit(f"{path} does not match the pinned SHA-256; {restore}")
+
+
+def build(input_path: Path, fetched_path: Path, out_dir: Path) -> dict:
+    check_pin(input_path, INPUT_SHA256, "restore it with `git checkout v0.2.5 -- data/govpl/govpl.parquet`")
+    check_pin(fetched_path, FETCHED_SHA256, "it is not the 2026-07-19 gov.pl crawl")
+    table = pq.read_table(input_path)
+    meta = fetched_meta(fetched_path)
+
+    kept, attribution, dropped = [], [], Counter()
+    for row in table.to_pylist():
+        if rule := pap_rule(row["text"]):
+            dropped[rule] += 1
+            continue
+        found = meta.get(text_key(row["text"]))
+        if found is None:
+            raise SystemExit(f"{row['id']}: text not found in {fetched_path}")
+        kept.append(row)
+        attribution.append({
+            "id": row["id"], "url": found["url"], "subsite": found["subsite"],
+            "title": found["title"], "date": found["date"], "license": row["license"],
+        })
+
+    stats = {
+        **BUILD_STATS, "drop_pap": sum(dropped.values()), "drop_pap_by_rule": dict(dropped),
+        "kept": len(kept),
+        "chars": sum(len(r["text"]) for r in kept), "tokens": sum(r["token_count"] for r in kept),
+        "licenses": dict(Counter(r["license"] for r in kept)),
+        "authors_with_value": sum(bool(r["author"]) for r in kept), "license": "CC-BY-SA-4.0",
+        "input_rows": table.num_rows, "input_sha256": INPUT_SHA256, "fetched_sha256": FETCHED_SHA256,
+    }
+    build_drops = sum(v for k, v in BUILD_STATS.items() if k.startswith("drop_"))
+    if stats["read"] - build_drops != table.num_rows:
+        raise SystemExit("BUILD_STATS does not add up to the input row count")
+    staged = {
+        out_dir / f"{SOURCE}.parquet.tmp": lambda p: pq.write_table(
+            table.from_pylist(kept, schema=table.schema), p, compression="zstd"),
+        out_dir / f"{SOURCE}.attribution.jsonl.tmp": lambda p: p.write_text(
+            "".join(json.dumps(a, ensure_ascii=False) + "\n" for a in attribution), encoding="utf-8"),
+        out_dir / f"{SOURCE}.stats.json.tmp": lambda p: p.write_text(
+            json.dumps(stats, indent=2, ensure_ascii=False) + "\n", encoding="utf-8"),
+    }
+    for path, write in staged.items():
+        write(path)
+    for path in staged:
+        os.replace(path, path.with_suffix(""))
+    return stats
+
+
+def main() -> None:
+    parser = argparse.ArgumentParser(description=__doc__.split("\n")[0])
+    parser.add_argument("--input", type=Path, default=DATA / f"{SOURCE}.parquet")
+    parser.add_argument("--fetched", type=Path, required=True, help="govpl.jsonl.zst from fetch_govpl.py")
+    parser.add_argument("--out", type=Path, default=DATA)
+    args = parser.parse_args()
+    print(json.dumps(build(args.input, args.fetched, args.out), indent=2, ensure_ascii=False))
+
+
+if __name__ == "__main__":
+    main()

src/clean_samorzad_gov_pl.py

@@ -48,6 +48,58 @@ _SITE_NAME_RE = re.compile(
     re.IGNORECASE | re.DOTALL,
 )
 _TITLE_RE = re.compile(r"<title\b[^>]*>(.*?)</title>", re.IGNORECASE | re.DOTALL)
+
+# Third-party carve-out (slayerlabs/polish-dynaword-v2#49). gov.pl's copyright
+# page says the service also carries works whose rights belong to third parties
+# and whose use needs the owners' consent, so the operator-declared platform
+# licence does not reach them. Each rule below is one hand-checked slice.
+THIRD_PARTY_DROPS = ("poetry_friday", "imgw_notice", "nongov_source")
+# Powiat Sokołowski's "Piątek z poezją" series reprints poems by named authors.
+_POETRY_FRIDAY_TITLE_RE = re.compile(r"piątek z poezją", re.IGNORECASE)
+# IMGW-PIB warnings carry a notice allowing reuse only in verbatim form.
+_IMGW_NOTICE = "Opracowanie niniejsze i jego format"
+# First "Źródło:" on a line: the text before it and up to 200 credited chars.
+_SOURCE_LINE_RE = re.compile(
+    r"^(?P<before>[^\n]*?)\bźr[óo]d[łl]o\s*:\s*(?P<credit>[^\n]{0,200})",
+    re.IGNORECASE | re.MULTILINE,
+)
+# "Fot. X, źródło: Y" credits a photo, illustration or film, not the text.
+_MEDIA_CREDIT_RE = re.compile(
+    r"^\W*(?:fot|ilustracj|film)|\b(?:fot|ilustracj|film)\w*[^\n]{0,20}$",
+    re.IGNORECASE,
+)
+# Credits naming a non-governmental author: media, encyclopedias, NGOs (OSP
+# units are associations), companies, and authored works. Associations of
+# local governments (Stowarzyszenie Łódzki Obszar Metropolitalny) are public.
+_NONGOV_CREDIT_RE = re.compile(
+    r"""
+      wikipedi | wikimedia | telewizja | \btvp\b | \bradio\s
+    | serwis\s+samorządowy\s+pap | redakcj[ai]\s+(?:życia|tygodnika)
+    | podlasie\s?24 | jasło112 | tymbark\.in | polskiobserwator | it\s+w\s+administracji
+    | stowarzyszeni[ea](?!\s+łódzki\s+obszar) | fundacj[ia] | \bkgw\b | \bskgw\b
+    | ^koło\s+gospodyń | \bklub\b | \bmlks\b | junis | amazonki | izba\s+(?:lotnicza|gospodarcza)
+    | \bosp\b | \bzosp\b | związku\s+ochotniczych\s+straży | polskiego\s+związku\s+łowieckiego
+    | lokalna\s+grupa\s+działania
+    | sp\.\s*z\s*o\.\s*o | \bpge\b | edp\s+renewables | grupa\s+nowy\s+szpital | carrefour
+    | \bautor(?:\s+tekstu)?\s*: | broszura | wspomnienia | monografia
+    | rodzinne\s+gniazdo | historia\s+i\s+teraźniejszość | historia\s+włókniarza
+    | jubiläumsschrift | bio\s+artystki | domowe\s+historie
+    """,
+    re.IGNORECASE | re.VERBOSE,
+)
+MODIFICATION_NOTICE = (
+    "Article text was extracted from the source HTML and de-identified "
+    "upstream. This DynaWord release normalizes metadata and may exclude "
+    "short, non-Polish, duplicate or third-party records; retained text is "
+    "otherwise unchanged after upstream de-identification."
+)
+# The release on main 92b36dc before the third-party drops. refilter_release()
+# starts from it because a full build() also needs every other source's
+# Parquet for cross-source dedup.
+PRE_CARVE_OUT_SHA256 = {
+    f"{SOURCE}.parquet": "3a683b4220b9b70698d7c524637d9ea5c1d72630b8f92ea7a871642a719cf2c0",
+    f"{SOURCE}.attribution.jsonl": "2def7488c34e749204842bcf21d33149738b34de2515115e3f79401f7eb271c4",
+}
 DYNAWORD_SCHEMA = pa.schema(
     [
         ("id", pa.string()),
@@ -250,16 +302,25 @@ def normalize_record(
         "license": LICENSE,
         "license_url": LICENSE_URL,
         "modified": True,
-        "modification_notice": (
-            "Article text was extracted from the source HTML and de-identified "
-            "upstream. This DynaWord release normalizes metadata and may exclude "
-            "short, non-Polish or duplicate records; retained text is otherwise "
-            "unchanged after upstream de-identification."
-        ),
+        "modification_notice": MODIFICATION_NOTICE,
     }
     return row, attribution
 
 
+def third_party_drop(title: str, text: str) -> str | None:
+    """Return the THIRD_PARTY_DROPS slice a record belongs to, or None."""
+    if _POETRY_FRIDAY_TITLE_RE.search(title):
+        return "poetry_friday"
+    if _IMGW_NOTICE in text:
+        return "imgw_notice"
+    for line in _SOURCE_LINE_RE.finditer(text):
+        if _NONGOV_CREDIT_RE.search(line["credit"]) and not _MEDIA_CREDIT_RE.search(
+            line["before"]
+        ):
+            return "nongov_source"
+    return None
+
+
 def _polish_ratio(text: str) -> float:
     letters = ALPHA_RE.findall(text)
     return len(POLISH_RE.findall(text)) / len(letters) if letters else 0.0
@@ -403,10 +464,7 @@ def build(
         "drop_lang": 0,
         "drop_dup": 0,
         "drop_ocr": 0,
-        "chars": 0,
-        "tokens": 0,
-        "licenses": {LICENSE: 0},
-        "authors_with_value": 0,
+        "drop_third_party": dict.fromkeys(THIRD_PARTY_DROPS, 0),
         "publisher_count": len(publishers),
         "local_public_tenant_count": sum(
             publisher_type != "central" for publisher_type in tenant_types.values()
@@ -414,7 +472,6 @@ def build(
         "central_tenant_count": sum(
             publisher_type == "central" for publisher_type in tenant_types.values()
         ),
-        "redirected_urls": 0,
         "cross_source_hashes_loaded": len(prior_hashes),
         "license": LICENSE,
         "upstream_dataset": UPSTREAM_DATASET,
@@ -444,6 +501,12 @@ def build(
             stats["drop_dup"] += 1
             continue
         seen.add(digest)
+        # After dedup, so the dup set matches the pre-carve-out release and
+        # refilter_release() gives the same rows as a full build.
+        reason = third_party_drop(str(record.get("title") or ""), text)
+        if reason:
+            stats["drop_third_party"][reason] += 1
+            continue
         record["text"] = text
         candidates.append(record)
 
@@ -462,16 +525,26 @@ def build(
         row, attribution = normalize_record(source_row, publishers, token_count, added)
         rows.append(row)
         attribution_rows.append(attribution)
-        stats["kept"] += 1
-        stats["chars"] += len(row["text"])
-        stats["tokens"] += token_count
-        stats["licenses"][LICENSE] += 1
-        stats["authors_with_value"] += int(bool(row["author"]))
-        stats["redirected_urls"] += int(
-            attribution["source_url"] != attribution["final_url"]
-        )
 
     table = pa.Table.from_pylist(rows, schema=DYNAWORD_SCHEMA)
+    _write_release(output_dir, table, attribution_rows, stats)
+    return stats
+
+
+def _write_release(
+    output_dir: Path, table: pa.Table, attribution_rows: list[dict], stats: dict
+) -> None:
+    """Fill the row-derived stats, then stage, validate and swap in the release."""
+    stats.update(
+        kept=table.num_rows,
+        chars=sum(map(len, table["text"].to_pylist())),
+        tokens=sum(table["token_count"].to_pylist()),
+        licenses={LICENSE: table.num_rows},
+        authors_with_value=sum(map(bool, table["author"].to_pylist())),
+        redirected_urls=sum(
+            row["source_url"] != row["final_url"] for row in attribution_rows
+        ),
+    )
     with tempfile.TemporaryDirectory(prefix=f".{SOURCE}-", dir=output_dir) as temp:
         staged_dir = Path(temp)
         staged_parquet = staged_dir / f"{SOURCE}.parquet"
@@ -492,12 +565,38 @@ def build(
         )
         _validate_staged_release(staged_dir)
         _replace_release(staged_dir, output_dir)
+
+
+def refilter_release(
+    output_dir: Path, expected: dict[str, str] = PRE_CARVE_OUT_SHA256
+) -> dict:
+    """Apply third_party_drop() to the pinned pre-carve-out release in place."""
+    for name, digest in expected.items():
+        if _file_sha256(output_dir / name) != digest:
+            raise ValueError(f"{name} is not the pinned pre-carve-out release")
+    table = pq.read_table(output_dir / f"{SOURCE}.parquet")
+    with (output_dir / f"{SOURCE}.attribution.jsonl").open(encoding="utf-8") as handle:
+        attributions = [json.loads(line) for line in handle]
+    stats = json.loads((output_dir / f"{SOURCE}.stats.json").read_text(encoding="utf-8"))
+    stats["drop_third_party"] = dict.fromkeys(THIRD_PARTY_DROPS, 0)
+    keep: list[bool] = []
+    for attribution, text in zip(attributions, table["text"].to_pylist(), strict=True):
+        reason = third_party_drop(attribution["title"], text)
+        if reason:
+            stats["drop_third_party"][reason] += 1
+        keep.append(reason is None)
+    kept_attributions = [
+        {**attribution, "modification_notice": MODIFICATION_NOTICE}
+        for attribution, kept in zip(attributions, keep)
+        if kept
+    ]
+    _write_release(output_dir, table.filter(pa.array(keep)), kept_attributions, stats)
     return stats
 
 
 def main() -> None:
     parser = argparse.ArgumentParser()
-    parser.add_argument("--input", type=Path, required=True, help="directory containing source Parquet shards")
+    parser.add_argument("--input", type=Path, help="directory containing source Parquet shards")
     parser.add_argument(
         "--manifest",
         type=Path,
@@ -505,17 +604,27 @@ def main() -> None:
     )
     parser.add_argument("--repo-root", type=Path, default=Path(__file__).resolve().parents[1])
     parser.add_argument("--publishers", type=Path, default=Path(__file__).with_name("samorzad_gov_pl_publishers.json"))
-    parser.add_argument("--added", required=True)
+    parser.add_argument("--added")
     parser.add_argument("--refresh-publishers", action="store_true")
-    args = parser.parse_args()
-    stats = build(
-        input_path=args.input,
-        manifest_path=args.manifest or args.input.parent / "manifest.json",
-        repo_root=args.repo_root.resolve(),
-        publishers_path=args.publishers,
-        added=args.added,
-        refresh=args.refresh_publishers,
+    parser.add_argument(
+        "--refilter-release",
+        action="store_true",
+        help="apply the third-party drops to the pinned pre-carve-out release in data/",
     )
+    args = parser.parse_args()
+    if args.refilter_release:
+        stats = refilter_release(args.repo_root.resolve() / "data" / SOURCE)
+    elif args.input is None or args.added is None:
+        parser.error("--input and --added are required unless --refilter-release")
+    else:
+        stats = build(
+            input_path=args.input,
+            manifest_path=args.manifest or args.input.parent / "manifest.json",
+            repo_root=args.repo_root.resolve(),
+            publishers_path=args.publishers,
+            added=args.added,
+            refresh=args.refresh_publishers,
+        )
     print(json.dumps(stats, ensure_ascii=False, indent=2))
 
 

src/fetch_govpl.py

@@ -1,10 +1,11 @@
 #!/usr/bin/env python3
 """Fetch gov.pl ministry/agency press releases into a SpeakLeash-style
 .jsonl.zst, so build_dynaword.py consumes it unchanged (native source, not via
-SpeakLeash). All textual content on gov.pl is published under CC-BY-SA 4.0 (site
-footer: "Treści tekstowe publikowane w serwisie … CC BY-SA 4.0"), authored and
-hosted by the Polish government — so the redistributor genuinely holds the right
-to license it.
+SpeakLeash). gov.pl publishes its text under CC-BY-SA 4.0 (site footer: "Treści
+tekstowe publikowane w serwisie … CC BY-SA 4.0"), but its copyright page excludes
+works whose copyright belongs to third parties. Articles that credit PAP (Polish
+Press Agency) as their source are such works, so normalize() drops them
+(pap_rule below; the same rule filters the released Parquet in clean_govpl.py).
 
 Enumeration is driven by the committed manifest src/govpl_subsites.json (built
 by discover_govpl.py): {slug, news_path, pages} per subsite. This keeps the
@@ -30,6 +31,35 @@ from html_text import html_to_text
 KEY = "govpl"  # build_dynaword reads <KEY>.jsonl.zst
 MIN_CHARS = 200  # same floor as build_dynaword; skip stubs early
 
+# PAP attribution forms seen in the fetched articles (hand-checked, issue
+# slayerlabs/polish-dynaword-v2#51). A photo credit ("fot. PAP") is not one of them.
+_PAP = r"(?:\bPAP\b|\bPolsk\w*\s+Agencj\w*\s+Prasow\w*)"
+PAP_RULES = {
+    # "Źródło: /PAP, MKiDN/", "Źródła: PAP, Culture.pl", "Źródło informacji: PAP/MKIDN",
+    # "Żródło: Polska Agencja Prasowa". Covers the issue's rule
+    # (?i)\bźr[óo]d[łl]o\s*:\s*/?[^\n]{0,60}\bPAP\b and its misspellings.
+    "source_line": re.compile(
+        r"(?i)\b[źżz]r[óo]d[łl][oa](?:\s+informacji)?\s*:?\s*/?[^\n]{0,60}" + _PAP),
+    # "Na podst. PAP", "Na podst. PKP S.A. i PAP" (uw-podlaski)
+    "based_on": re.compile(r"(?i)\bna\s+podst\.[^\n]{0,60}\bPAP\b"),
+    # a line holding only "(PAP)" or "/PAP/", the agency's own sign-off
+    "sign_off": re.compile(r"(?m)^[ \t]*[/(*]?[ \t]*PAP[ \t]*[)/]?\.?[ \t\r]*$"),
+}
+# A PAP interview transcript: the agency's questions start lines with "PAP:".
+# ponytail: three such lines; one KRRiT report quoting two PAP questions stays in.
+_PAP_QUESTION = re.compile(r"(?m)^PAP\s*:")
+PAP_INTERVIEW_MIN_QUESTIONS = 3
+
+
+def pap_rule(text: str) -> str | None:
+    """Name of the first PAP rule that matches `text`, or None."""
+    for name, rx in PAP_RULES.items():
+        if rx.search(text):
+            return name
+    if len(_PAP_QUESTION.findall(text)) >= PAP_INTERVIEW_MIN_QUESTIONS:
+        return "interview"
+    return None
+
 # Body ends where the editor-content region gives way to gallery/attachments/tags.
 _END_MARKER = re.compile(
     r'<[^>]+class="[^"]*(?:attachments|art-tags|tags|social|share|gallery|files)[^"]*"'
@@ -54,12 +84,15 @@ def article_body(html: str) -> str:
 
 
 def normalize(art: dict, html: str, slug: str) -> dict | None:
-    """Listing entry {url,title,date} + article HTML -> record, or None if thin."""
+    """Listing entry {url,title,date} + article HTML -> record, or None if thin
+    or a PAP reprint."""
     body = article_body(html)
     if len(body) < MIN_CHARS:
         return None
     title = art.get("title") or ""
     text = f"{title}\n\n{body}" if title else body
+    if pap_rule(text):
+        return None
     return {
         "text": text,
         "meta": {

src/sources.py

@@ -471,22 +471,38 @@ SOURCES = {
         "pretty": "gov.pl — Polish government press releases",
         "license": "CC-BY-SA-4.0",
         "license_spdx": "CC-BY-SA-4.0",
-        "traceable": "All textual content on gov.pl is published under CC-BY-SA "
-                     "4.0 (site footer: \"Treści tekstowe publikowane w serwisie "
-                     "… są udostępniane na licencji CC BY-SA 4.0\"). Authored and "
-                     "hosted by the Polish government, which holds the right to "
-                     "license it — same open family as Polish Wikipedia.",
+        "traceable": "gov.pl publishes its text under CC BY-SA 4.0 (site footer and "
+                     "https://www.gov.pl/web/gov/prawa-autorskie: \"Treści tekstowe "
+                     "publikowane w serwisie … są udostępniane na licencji CC BY-SA "
+                     "4.0\"). The same page excludes works whose copyright belongs to "
+                     "third parties; using those needs the owners' consent. "
+                     "Articles crediting PAP are therefore dropped (legal note).",
         "upstream": "https://www.gov.pl/",
         "provenance": "Ministry/agency press releases fetched directly from gov.pl "
                       "(editor-content per article, HTML stripped) by "
                       "src/fetch_govpl.py. Subsites + their non-uniform news-section "
                       "paths are enumerated by src/discover_govpl.py into the "
                       "committed manifest src/govpl_subsites.json (gov.pl has no "
-                      "sitemap/API). Each record's meta.url resolves on gov.pl.",
+                      "sitemap/API). v0.2.6 removes PAP reprints from the v0.2.5 "
+                      "Parquet with src/clean_govpl.py, which also writes "
+                      "govpl.attribution.jsonl: each row's gov.pl URL, subsite, "
+                      "title and date.",
+        "legal_note": "gov.pl's copyright page carves out \"utwory, co do których "
+                      "prawa autorskie przysługują osobom trzecim\" from its CC BY-SA "
+                      "4.0 grant. Articles crediting PAP (Polish Press Agency) as their "
+                      "source are such works, so 188 are dropped (rule "
+                      "fetch_govpl.pap_rule, applied by src/clean_govpl.py): a "
+                      "\"Źródło:\" line naming PAP (163), \"Na podst. … PAP\" (22), a "
+                      "\"(PAP)\" sign-off line (2) and one PAP interview transcript (1). "
+                      "PAP photo credits and articles that only quote or mention PAP "
+                      "stay. Other third-party text is not screened: 710 kept articles "
+                      "carry a \"Źródło:\" line naming some other source; they were "
+                      "not reviewed.",
         "domain": "government",
         # actual min/max of article dates in the fetched corpus (90,124 dated).
         "created": "2005-09-26, 2026-07-19",
         "is_ocr": False,
+        "custom_datasheet": True,
     },
     "samorzad_gov_pl": {
         "added": "2026-08-14",
@@ -495,8 +511,9 @@ SOURCES = {
         "pretty": "samorzad.gov.pl — Polish public-sector institutions",
         "license": "CC-BY-SA-4.0",
         "license_spdx": "CC-BY-SA-4.0",
-        "traceable": "The shared samorzad.gov.pl platform footer publishes textual "
-                     "content under CC BY-SA 4.0. The v0.2 source release is a "
+        "traceable": "Operator-declared platform licence: the shared samorzad.gov.pl "
+                     "platform footer publishes textual content under CC BY-SA 4.0. "
+                     "The v0.2 source release is a "
                      "de-identified subset of the audited v0.1 pages, all of which "
                      "referenced the shared footer script; audiovisual media and "
                      "attachment bytes are not redistributed.",
@@ -509,6 +526,18 @@ SOURCES = {
                       "from each tenant homepage's og:site_name metadata and stored "
                       "as DynaWord author values; discovered, fetched and attribution "
                       "URLs plus license details are preserved in the sidecar.",
+        "legal_note": "The licence is declared by the platform operator (KPRM) in the "
+                      "shared footer; the operator's agreements with the publishing "
+                      "institutions are unpublished and no per-institution grant was "
+                      "found. A question to KPRM is pending (slayerlabs/polish-dynaword-v2#49) "
+                      "and its answer may change this source's status. gov.pl's copyright "
+                      "page states that the service also carries third-party works whose "
+                      "use requires the owners' consent; the platform licence does not "
+                      "cover them. 144 such rows are dropped by src/clean_samorzad_gov_pl.py: "
+                      "69 Friday poetry posts, 9 IMGW-PIB notices and 66 articles credited "
+                      "to a non-governmental source. Contest entries, reposts credited "
+                      "only by a de-identified URL and other undetected third-party text "
+                      "may remain.",
         "domain": "local government, education, social services and culture",
         "created": "2020-2026 for 21,892 dated records; 59,530 source records undated",
         "is_ocr": False,

src/test_govpl_contract.py

@@ -1,12 +1,20 @@
 """TDD contract for the direct gov.pl (press releases) ingestion path."""
+import json
+import re
 import sys
 import unittest
 from pathlib import Path
 
+import pyarrow.parquet as pq
+
 sys.path.insert(0, str(Path(__file__).resolve().parent))
 
 from discover_govpl import listing_articles, news_candidates, max_page
-from fetch_govpl import article_body, normalize, BASE
+from fetch_govpl import article_body, normalize, pap_rule, BASE
+
+DATA = Path(__file__).resolve().parents[1] / "data" / "govpl"
+ISSUE_51_RULE = re.compile(r"(?i)\bźr[óo]d[łl]o\s*:\s*/?[^\n]{0,60}\bPAP\b")  # polish-dynaword-v2#51
+RELEASED_ROWS = 88002  # 88,190 in v0.2.5 minus 188 PAP reprints
 
 LISTING = """
 <div class="art-prev">
@@ -79,6 +87,83 @@ class GovplContractTest(unittest.TestCase):
         self.assertEqual(max_page('<a href="?page=2">2</a><a href="?page=294">294</a>'), 294)
         self.assertEqual(max_page("<html>no pager</html>"), 1)  # single-page listing
 
+    def test_pap_reprints_are_named(self):
+        cases = {
+            "Tekst.\n\nŹródło: /PAP, MKiDN/\n\nZdjęcia (3)": "source_line",
+            "Tekst.\n\nŹródło: PAP": "source_line",
+            "Tekst. Źródło PAP, MKiDN": "source_line",
+            "Źródło informacji: PAP/MKIDN": "source_line",
+            "żródło: /PAP,MKiDN/": "source_line",
+            "Źródła: dzieje.pl, Mariusz Jarosiński (PAP), Muzeum Historii Polski": "source_line",
+            "Żródło: Polska Agencja Prasowa": "source_line",
+            "Tekst.\n\nNa podst. PKP S.A. i PAP\n\r\nZdjęcia (13)": "based_on",
+            "Tekst.\n\n(PAP)\n\n\r\nZdjęcia (1)": "sign_off",
+            "Wywiad\n\nPAP: Pytanie?\nOdpowiedź.\nPAP: Drugie?\nTak.\nPAP: Trzecie?\nNie.": "interview",
+        }
+        for text, rule in cases.items():
+            self.assertEqual(pap_rule(text), rule, text)
+
+    def test_pap_mentions_and_photo_credits_are_kept(self):
+        for text in (
+            "Minister na konferencji. Fot. PAP/Leszek Szymański",
+            "fot. PAP\n\nZdjęcia (2)",
+            "– powiedziała minister w rozmowie z PAP.",
+            "Wyniki ogłoszono w Centrum Prasowym PAP.",
+            "Akapit kończy się tak (PAP). Dalej tekst ministerstwa.",
+            "Źródło: MKiDN\n\nfot. PAP",
+            "Raport KRRiT\n\nPAP: Pytanie?\nOdpowiedź.\nPAP: Drugie?\nTak.",
+        ):
+            self.assertIsNone(pap_rule(text), text)
+
+    def test_normalize_drops_pap_reprint(self):
+        art = {"url": "/web/kultura/x", "title": "Nagłówek", "date": "14.07.2026"}
+        reprint = ARTICLE.replace("</p>", "</p><p>Źródło: /PAP, MKiDN/</p>", 1)
+        self.assertIsNotNone(normalize(art, ARTICLE, "kultura"))
+        self.assertIsNone(normalize(art, reprint, "kultura"))
+
+
+class GovplReleaseContractTest(unittest.TestCase):
+    """The released files: PAP reprints gone, stats and URL sidecar match the Parquet."""
+
+    @classmethod
+    def setUpClass(cls):
+        cls.table = pq.read_table(DATA / "govpl.parquet")
+        cls.stats = json.loads((DATA / "govpl.stats.json").read_text(encoding="utf-8"))
+        with (DATA / "govpl.attribution.jsonl").open(encoding="utf-8") as stream:
+            cls.sidecar = [json.loads(line) for line in stream]
+
+    def test_no_pap_reprints(self):
+        texts = self.table.column("text").to_pylist()
+        self.assertEqual([t for t in texts if ISSUE_51_RULE.search(t)], [])
+        self.assertEqual([t for t in texts if pap_rule(t)], [])
+
+    def test_row_count_and_stats(self):
+        st = self.stats
+        self.assertEqual(self.table.num_rows, RELEASED_ROWS)
+        self.assertEqual(st["kept"], RELEASED_ROWS)
+        self.assertEqual(st["drop_pap"], sum(st["drop_pap_by_rule"].values()))
+        self.assertEqual(st["input_rows"] - st["drop_pap"], st["kept"])
+        drops = sum(v for k, v in st.items() if k.startswith("drop_") and isinstance(v, int))
+        self.assertEqual(st["read"] - drops, st["kept"])
+        self.assertEqual(st["tokens"], sum(self.table.column("token_count").to_pylist()))
+        self.assertEqual(st["chars"], sum(map(len, self.table.column("text").to_pylist())))
+        self.assertEqual(st["licenses"], {"CC-BY-SA-4.0": RELEASED_ROWS})
+        ids = self.table.column("id").to_pylist()
+        self.assertEqual(len(set(ids)), len(ids))
+
+    def test_datasheet_matches_stats(self):
+        st, md = self.stats, (DATA / "govpl.md").read_text(encoding="utf-8")
+        self.assertIn(f"| {st['kept']:,} | {st['chars']:,} | {st['tokens']:,} |", md)
+        self.assertIn(f"drop PAP reprints: **{st['drop_pap']:,}**", md)
+        self.assertIn(f"read {st['input_rows']:,} → kept {st['kept']:,}", md)
+
+    def test_sidecar_gives_every_row_a_govpl_url(self):
+        self.assertEqual([a["id"] for a in self.sidecar], self.table.column("id").to_pylist())
+        for a in self.sidecar:
+            self.assertEqual(set(a), {"id", "url", "subsite", "title", "date", "license"})
+            self.assertTrue(a["url"].startswith(BASE + "/web/" + a["subsite"] + "/"), a)
+            self.assertEqual(a["license"], "CC-BY-SA-4.0")
+
 
 if __name__ == "__main__":
     unittest.main()

src/test_samorzad_gov_pl_contract.py

@@ -18,6 +18,8 @@ from clean_samorzad_gov_pl import (
     _replace_release,
     normalize_record,
     publisher_name_from_html,
+    refilter_release,
+    third_party_drop,
     verify_upstream_release,
 )
 
@@ -76,6 +78,86 @@ def test_normalize_record_maps_publisher_to_author_and_preserves_provenance():
     assert "de-identified upstream" in attribution["modification_notice"]
 
 
+@pytest.mark.parametrize(
+    ("title", "text", "expected"),
+    [
+        ("Piątek z poezją - wiersz", "Treść.", "poetry_friday"),
+        ("Ostrzeżenie", "Opracowanie niniejsze i jego format, jako ...", "imgw_notice"),
+        ("Historia", "Tekst.\nŹródło: Wikipedia, wolna encyklopedia", "nongov_source"),
+        ("Zebranie", "Tekst.\nŹródło: OSP Dąbrówka", "nongov_source"),
+        ("Pomoc", "Tekst.\nzdjęcie i źródło: Radio Kielce", "nongov_source"),
+        ("Koncert", "singiel promujący film „Miłość na pierwszą stronę” (źródło: oficjalne BIO Artystki)", "nongov_source"),
+        # photo, illustration and film credits are not text credits
+        ("Szlak", "Fot. Roland Mielnicki, źródło: tymbark.in", None),
+        ("Praga", "Ilustracja przedstawia obraz Canaletta, źródło: Wikimedia Commons.", None),
+        ("Dożynki", "Poniżej film z wydarzenia (źródło: Małopolskie Stowarzyszenie Sołtysów).", None),
+        # public-sector credits and article titles that only mention an NGO
+        ("Komunikat", "Tekst.\nŹródło: Urząd Gminy Raba Wyżna", None),
+        ("ŁOM", "Źródło: Stowarzyszenie Łódzki Obszar Metropolitalny", None),
+        ("Lasy", "Źródło: Koła Gospodyń Wiejskich promują zdrową żywność — Lasy Państwowe", None),
+        ("Konkurs", "Wyniki konkursu poezji. Nagrodzone prace zostaną wystawione.", None),
+    ],
+)
+def test_third_party_drop_rules(title, text, expected):
+    assert third_party_drop(title, text) == expected
+
+
+def test_refilter_release_drops_third_party_rows_and_recomputes_stats(tmp_path):
+    texts = ["Zwykły komunikat gminy.", "Źródło: Fundacja Przykład", "Wiersz."]
+    titles = ["Komunikat", "Artykuł", "Piątek z poezją"]
+    table = pa.Table.from_pylist(
+        [
+            {
+                "id": f"samorzad_gov_pl_{index}", "text": text, "source": "samorzad_gov_pl",
+                "added": "2026-08-14", "created": "", "token_count": index + 1,
+                "license": "CC-BY-SA-4.0", "author": "Gmina",
+            }
+            for index, text in enumerate(texts)
+        ],
+        schema=DYNAWORD_SCHEMA,
+    )
+    pq.write_table(table, tmp_path / "samorzad_gov_pl.parquet")
+    (tmp_path / "samorzad_gov_pl.attribution.jsonl").write_text(
+        "".join(
+            json.dumps(
+                {
+                    "id": f"samorzad_gov_pl_{index}", "title": title, "publisher": "Gmina",
+                    "source_url": "u", "final_url": "u" if index else "v",
+                    "attribution_url": "u" if index else "v", "modification_notice": "old",
+                }
+            )
+            + "\n"
+            for index, title in enumerate(titles)
+        ),
+        encoding="utf-8",
+    )
+    (tmp_path / "samorzad_gov_pl.stats.json").write_text(
+        json.dumps({"read": 3, "kept": 3, "drop_dup": 0}), encoding="utf-8"
+    )
+    pins = {
+        name: hashlib.sha256((tmp_path / name).read_bytes()).hexdigest()
+        for name in ("samorzad_gov_pl.parquet", "samorzad_gov_pl.attribution.jsonl")
+    }
+
+    stats = refilter_release(tmp_path, pins)
+
+    out = pq.read_table(tmp_path / "samorzad_gov_pl.parquet")
+    sidecar = [
+        json.loads(line)
+        for line in (tmp_path / "samorzad_gov_pl.attribution.jsonl").open(encoding="utf-8")
+    ]
+    assert out["id"].to_pylist() == ["samorzad_gov_pl_0"]
+    assert [row["id"] for row in sidecar] == ["samorzad_gov_pl_0"]
+    assert "third-party" in sidecar[0]["modification_notice"]
+    assert stats["drop_third_party"] == {
+        "poetry_friday": 1, "imgw_notice": 0, "nongov_source": 1,
+    }
+    assert (stats["read"], stats["kept"], stats["tokens"], stats["redirected_urls"]) == (3, 1, 1, 1)
+    assert stats["chars"] == len(texts[0])
+    with pytest.raises(ValueError, match="not the pinned pre-carve-out release"):
+        refilter_release(tmp_path, pins)
+
+
 def test_verify_upstream_release_rejects_modified_shard(tmp_path):
     data_dir = tmp_path / "data"
     data_dir.mkdir()
@@ -185,7 +267,14 @@ def test_release_artifact_matches_stats_and_has_institution_authors():
     assert table.schema == DYNAWORD_SCHEMA
     assert table.num_rows == stats["kept"] == len(attributions)
     assert stats["read"] == 81_418
-    assert stats["kept"] == 72_472
+    assert stats["kept"] == 72_328
+    assert stats["drop_third_party"] == {
+        "poetry_friday": 69,
+        "imgw_notice": 9,
+        "nongov_source": 66,
+    }
+    assert stats["chars"] == 106_816_031
+    assert stats["tokens"] == 39_890_244
     assert stats["authors_with_value"] == stats["kept"]
     assert stats["publisher_count"] == 247
     assert stats["local_public_tenant_count"] == 246
@@ -218,3 +307,9 @@ def test_release_artifact_matches_stats_and_has_institution_authors():
     assert sum(
         item["source_url"] != item["final_url"] for item in attributions
     ) == stats["redirected_urls"]
+    # slayerlabs/polish-dynaword-v2#49: no third-party slice survives.
+    assert [
+        item["id"]
+        for item, text in zip(attributions, texts, strict=True)
+        if third_party_drop(item["title"], text)
+    ] == []