Datasets:

PiotrSty
/

rcl-legislacja-consultations

Error code:   StreamingRowsError
Exception:    TypeError
Message:      Couldn't cast array of type
struct<application/msword: int64, application/pdf: int64, application/rtf: int64, application/vnd.oasis.opendocument.text: int64, application/vnd.openxmlformats-officedocument.spreadsheetml.sheet: int64, application/vnd.openxmlformats-officedocument.wordprocessingml.document: int64, application/zip: int64, text/html: int64, text/plain: int64>
to
{'application/msword': Value('int64'), 'application/pdf': Value('int64'), 'application/rtf': Value('int64'), 'application/vnd.oasis.opendocument.text': Value('int64'), 'application/vnd.openxmlformats-officedocument.spreadsheetml.sheet': Value('int64'), 'application/vnd.openxmlformats-officedocument.wordprocessingml.document': Value('int64'), 'application/zip': Value('int64'), 'image/jpeg': Value('int64'), 'image/png': Value('int64'), 'text/plain': Value('int64'), 'text/xml': Value('int64')}
Traceback:    Traceback (most recent call last):
                File "/src/services/worker/src/worker/utils.py", line 147, in get_rows_or_raise
                  return get_rows(
                      dataset=dataset,
                  ...<4 lines>...
                      column_names=column_names,
                  )
                File "/src/libs/libcommon/src/libcommon/utils.py", line 272, in decorator
                  return func(*args, **kwargs)
                File "/src/services/worker/src/worker/utils.py", line 127, in get_rows
                  rows_plus_one = list(itertools.islice(safe_iter(ds, dataset=dataset), rows_max_number + 1))
                File "/src/services/worker/src/worker/utils.py", line 483, in safe_iter
                  yield from ds.decode(False) if ds.features else ds
                File "/usr/local/lib/python3.14/site-packages/datasets/iterable_dataset.py", line 2840, in __iter__
                  for key, example in ex_iterable:
                                      ^^^^^^^^^^^
                File "/usr/local/lib/python3.14/site-packages/datasets/iterable_dataset.py", line 2373, in __iter__
                  for key, pa_table in self._iter_arrow():
                                       ~~~~~~~~~~~~~~~~^^
                File "/usr/local/lib/python3.14/site-packages/datasets/iterable_dataset.py", line 2398, in _iter_arrow
                  for key, pa_table in self.ex_iterable._iter_arrow():
                                       ~~~~~~~~~~~~~~~~~~~~~~~~~~~~^^
                File "/usr/local/lib/python3.14/site-packages/datasets/iterable_dataset.py", line 536, in _iter_arrow
                  for key, pa_table in iterator:
                                       ^^^^^^^^
                File "/usr/local/lib/python3.14/site-packages/datasets/iterable_dataset.py", line 419, in _iter_arrow
                  for key, pa_table in self.generate_tables_fn(**gen_kwags):
                                       ~~~~~~~~~~~~~~~~~~~~~~~^^^^^^^^^^^^^
                File "/usr/local/lib/python3.14/site-packages/datasets/packaged_modules/json/json.py", line 343, in _generate_tables
                  self._cast_table(pa_table, json_field_paths=json_field_paths),
                  ~~~~~~~~~~~~~~~~^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
                File "/usr/local/lib/python3.14/site-packages/datasets/packaged_modules/json/json.py", line 132, in _cast_table
                  pa_table = table_cast(pa_table, self.info.features.arrow_schema)
                File "/usr/local/lib/python3.14/site-packages/datasets/table.py", line 2378, in table_cast
                  return cast_table_to_schema(table, schema)
                File "/usr/local/lib/python3.14/site-packages/datasets/table.py", line 2312, in cast_table_to_schema
                  cast_array_to_feature(
                  ~~~~~~~~~~~~~~~~~~~~~^
                      table[name] if name in table_column_names else pa.array([None] * len(table), type=schema.field(name).type),
                      ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
                      feature,
                      ^^^^^^^^
                  )
                  ^
                File "/usr/local/lib/python3.14/site-packages/datasets/table.py", line 1861, in wrapper
                  return pa.chunked_array([func(chunk, *args, **kwargs) for chunk in array.chunks])
                                           ~~~~^^^^^^^^^^^^^^^^^^^^^^^^
                File "/usr/local/lib/python3.14/site-packages/datasets/table.py", line 2068, in cast_array_to_feature
                  _c(array.field(name) if name in array_fields else null_array, subfeature)
                  ~~^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
                File "/usr/local/lib/python3.14/site-packages/datasets/table.py", line 1863, in wrapper
                  return func(array, *args, **kwargs)
                File "/usr/local/lib/python3.14/site-packages/datasets/table.py", line 2158, in cast_array_to_feature
                  raise TypeError(f"Couldn't cast array of type\n{_short_str(array.type)}\nto\n{_short_str(feature)}")
              TypeError: Couldn't cast array of type
              struct<application/msword: int64, application/pdf: int64, application/rtf: int64, application/vnd.oasis.opendocument.text: int64, application/vnd.openxmlformats-officedocument.spreadsheetml.sheet: int64, application/vnd.openxmlformats-officedocument.wordprocessingml.document: int64, application/zip: int64, text/html: int64, text/plain: int64>
              to
              {'application/msword': Value('int64'), 'application/pdf': Value('int64'), 'application/rtf': Value('int64'), 'application/vnd.oasis.opendocument.text': Value('int64'), 'application/vnd.openxmlformats-officedocument.spreadsheetml.sheet': Value('int64'), 'application/vnd.openxmlformats-officedocument.wordprocessingml.document': Value('int64'), 'application/zip': Value('int64'), 'image/jpeg': Value('int64'), 'image/png': Value('int64'), 'text/plain': Value('int64'), 'text/xml': Value('int64')}

Need help to make the dataset viewer work? Make sure to review how to configure the dataset viewer, and open a discussion for direct support.

RCL Legislacja Consultations Metadata

Staging dataset-source artifact for Polish Dynaword / Slayer dataset research.

This private repository contains metadata, raw HTML snapshots, restricted source documents and extracted-text payloads from public legislacja.gov.pl RCL legislative-process pages. It is intended for source inventory, provenance review, manual review queueing, legal/PII screening and future controlled data-mix decisions.

The current main revision includes unreviewed payloads under raw_documents/ and extracted_text/. See RESTRICTED_SOURCE_NOTICE.md. Their presence in this private repository does not assign a blanket open license or clear any row for training or unrestricted redistribution.

The repository-level license metadata does not assign a blanket open-content license to RCL documents. Rights and reuse status must be reviewed at document or source level before text is released or included in a training corpus.

rcl_inventory/ - full project-list inventory for law projects, 2606 projects.
rcl_2026_consultations/ - 2026 consultation metadata batch, 150 projects.
rcl_2025_consultations/ - complete 2025 project-of-law metadata crawl: 315 projects, 250 consultation catalogs, 5556 document records and 4166 selected-document review rows. Raw public HTML snapshots are retained for provenance; document binaries are not included.
rcl_2026_consultations/review_queue.csv - manual-review queue generated from selected consultation documents.
*/raw_html/ - raw public HTML snapshots used for auditability.
*/runs/ - run summaries with crawl arguments and evidence counters.
rcl_gold_pilot_v0_1/ - content-addressed manifest, checksums, run evidence, extraction metadata, machine-observations and annotation evidence for the bounded 40-document review pilot, with restricted raw and extracted payloads.
rcl_gold_pilot_v0_1/legal_review_pack/ - metadata-only legal review queue with document provenance, working legal premise, PII flags and empty decision fields for legal reviewers.
rcl_remaining_consultations_v0_1/ - source-processing batch with restricted payloads for the 42 review-queue rows not included in the accepted pilot; now accepted into that batch's annotations.csv and queued for legal review.
rcl_2026_selected_remaining_918_v0_1/ - source-processing batch with restricted payloads for the remaining 918 rows from the 1000 selected-document queue after the accepted 82 rows.
rcl_2025_selected_v0_1/ - source and extraction evidence for 4166 selected 2025 consultation documents. It contains content-addressed manifests, checksums, machine observations, validation reports and restricted raw and extracted payloads.
rcl_2026_delta_20260826_v0_1/ - incremental source and extraction evidence for 8 newly selected document URLs found in the 2026-08-26 source check, including restricted raw and extracted payloads.
rcl_dynaword_handoff_v0_1/ - metadata-only Polish DynaWord adapter package: 1000 candidate rows, schema mapping metadata, token-count proxy, row-level blockers and source datasheet. It intentionally has no text column and no raw or extracted text payload.

Current Evidence

Full inventory projects: 2606
2026 batch projects: 150
2026 projects with consultation catalog URL: 127
2026 consultation documents found: 1672
2026 documents selected by current keyword filter: 1000
Restricted source binaries included in current main: 5174
Restricted extracted-text files included in current main: 5174
Crawl errors: 0
Gold-pilot review decisions completed: 40 of 40
Gold-pilot text extraction: 38 extracted, 2 empty-text cases requiring OCR or manual inspection
Remaining consultation batch: 42 rows, 42 downloaded locally, 41 extracted, 1 empty-text case, 0 download failures, 42 accepted annotation rows
Full selected queue: 1000 rows
Remaining 918 batch: 918 downloaded locally, 0 download failures, 774 extracted, 144 empty-text cases, 547 machine PII hints
Polish DynaWord handoff: 1000 metadata-only candidate rows, 82 reviewed rows, 918 draft-only rows, token-count proxy 5,615,223, training-ready rows 0
2025 batch projects: 315; projects with consultation catalogs: 250
2025 consultation documents found: 5556; selected unique source URLs: 4166
2025 local source validation: 4166 of 4166 downloaded, 0 failures, 4042 unique artifact digests
2025 local extraction validation: 3323 extracted, 843 empty/unsupported, 2472 good, 375 usable, 476 poor and 843 not-extractable quality hints
2025 machine PII triage: 2434 yes hints and 1732 uncertain; reviewed rows: 0
2026-08-26 delta: 1 new project, 14 modified projects and 16 new document URLs; 8 URLs matched the selection rule
2026-08-26 local delta processing: 8 of 8 downloaded, 0 failures, 7 unique artifact digests, 6 extracted and 2 empty/unsupported
2026-08-26 delta quality hints: 5 good, 1 poor and 2 not extractable; machine PII hints: 3 yes and 5 uncertain; reviewed rows: 0

Intended Use

This is not yet a cleaned training corpus. Treat it as a dataset-source and provenance object. Before inclusion in any training mix, documents need source review, PII screening, legal/provenance decisions and extraction quality checks. The 82 accepted review rows are review evidence only; all keep legal_status=review_needed. In this release, review_needed means the source is a public official RCL/RPL document with a plausible public-sector information reuse path, but no blanket dataset license has been assigned and legal review must still verify attribution, reuse, copyright/database-right and PII constraints.

The 918-row batch contains machine observations and draft suggestions only. It has not been human-accepted and does not clear any row for release or training.

The 2025 batch is also unreviewed. Its extraction and PII fields are machine triage evidence only. The metadata-only publication does not clear any 2025 row for text release, redistribution or training.

The 2026-08-26 delta is unreviewed as well. Its eight source rows remain blocked on legal, PII, extraction-quality and human-review gates. One duplicate content digest across distinct source URLs is retained as provenance evidence.

The DynaWord handoff package is an adapter/review artifact for possible future source inclusion and controlled ablation design. It must not be loaded as a training split until legal review, PII scrubbing and human review gates are complete.

The current private release includes restricted source payloads and must not be interpreted as a claim that source documents are licensed under Polish Dynaword's dataset license. The immutable v0.3.0-rcl-2026-08-26-metadata-delta tag preserves the preceding metadata-only publication boundary.

Slayer Ontology Mapping

Object: RCL source inventory / consultation metadata source.
Version: this uploaded snapshot.
Protocol: scripts/rcl_downloader.py and scripts/rcl_review_queue.py in the companion code repository.
Evidence: run summaries, document/category counts, extraction metadata and accepted calibration annotations.
Claim status: preliminary; candidate claims require manual-review evidence.

Legal Review Note

See rcl_gold_pilot_v0_1/legal_review_note.md for the working legal-review premise and source links for RCL/RPL public-sector information reuse review. See rcl_gold_pilot_v0_1/legal_review_pack/ for the row-level queue to be completed by legal reviewers.

Downloads last month: 1,443

Total file size:

3.38 GB

RCL Legislacja Consultations Metadata

Contents

Current Evidence

Intended Use

Slayer Ontology Mapping

Legal Review Note