Datasets:

CatholicCorpus
/

catholiccorpus

Error code:   FeaturesError
Exception:    UnicodeDecodeError
Message:      'utf-8' codec can't decode byte 0xac in position 8: invalid start byte
Traceback:    Traceback (most recent call last):
                File "/src/services/worker/src/worker/job_runners/split/first_rows.py", line 246, in compute_first_rows_from_streaming_response
                  iterable_dataset = iterable_dataset._resolve_features()
                                     ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
                File "/usr/local/lib/python3.12/site-packages/datasets/iterable_dataset.py", line 4196, in _resolve_features
                  features = _infer_features_from_batch(self.with_format(None)._head())
                                                        ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
                File "/usr/local/lib/python3.12/site-packages/datasets/iterable_dataset.py", line 2533, in _head
                  return next(iter(self.iter(batch_size=n)))
                         ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
                File "/usr/local/lib/python3.12/site-packages/datasets/iterable_dataset.py", line 2711, in iter
                  for key, pa_table in ex_iterable.iter_arrow():
                                       ^^^^^^^^^^^^^^^^^^^^^^^^
                File "/usr/local/lib/python3.12/site-packages/datasets/iterable_dataset.py", line 2249, in _iter_arrow
                  yield from self.ex_iterable._iter_arrow()
                File "/usr/local/lib/python3.12/site-packages/datasets/iterable_dataset.py", line 494, in _iter_arrow
                  for key, pa_table in iterator:
                                       ^^^^^^^^
                File "/usr/local/lib/python3.12/site-packages/datasets/iterable_dataset.py", line 384, in _iter_arrow
                  for key, pa_table in self.generate_tables_fn(**gen_kwags):
                                       ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
                File "/usr/local/lib/python3.12/site-packages/datasets/packaged_modules/text/text.py", line 98, in _generate_tables
                  batch = f.read(self.config.chunksize)
                          ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
                File "/usr/local/lib/python3.12/site-packages/datasets/utils/file_utils.py", line 844, in read_with_retries
                  out = read(*args, **kwargs)
                        ^^^^^^^^^^^^^^^^^^^^^
                File "<frozen codecs>", line 322, in decode
              UnicodeDecodeError: 'utf-8' codec can't decode byte 0xac in position 8: invalid start byte

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.

CatholicCorpus

An open-access, NLP-ready corpus of Catholic texts spanning 2,000 years of the Catholic intellectual tradition — from the Church Fathers to the 20th century.

67,772 content files | 16 collections | 35.9 GB | 2,000 years of coverage

What's In the Corpus

#	Collection	Content Files	Size	Format
01	Git Repos (CSEL, Aquinas Opera Omnia, Septuagint, Byzantine Text, eBible)	4,668	3.4 GB	TEI XML, TXT
02	Corpus Corporum (Patrologia Latina + 29 Latin/Greek corpora from mlat.uzh.ch)	29,820	4.2 GB	TEI XML
03	Gutenberg (Douay-Rheims, Imitation of Christ, Baltimore Catechism, etc.)	18	9 MB	TXT
04	Direct Downloads (Clementine Vulgate, SBLGNT, Canon Law 1917, Roman Catechism)	1,244	42 MB	TXT, PDF
05	CCEL (Ante-Nicene/Nicene Fathers, Summa Theologiae, devotional classics)	61	691 MB	PDF
06	Catholic Encyclopedia (complete 1913 edition, 15 volumes)	18	1.0 GB	PDF
07	Latin & Franciscan (Anselm, Peter Lombard, Bonaventure, Scotus, Ockham)	30	1.1 GB	PDF
08	Liturgical & Hymns (Divine Office, hymns, pre-1928 papal encyclicals)	31,141	202 MB	HTML, GABC
09	Archive & Misc (Council of Trent, Golden Legend, Butler's Lives, Albert the Great)	65	2.1 GB	PDF
11	Catholic Bible (Douay-Rheims, Haydock annotated Bible, Lapide Commentary)	12	673 MB	PDF
12	English Catholic Thinkers (Newman 33 vols, Chesterton 7, Belloc 5)	45	841 MB	PDF
13	Mystics (Therese, Julian of Norwich, Catherine of Siena, Ignatius, etc.)	13	182 MB	PDF
14	Liturgical Expansion (Liber Usualis, Missale Romanum, Rituale, Breviarium)	16	2.2 GB	PDF
15	Late Scholastics (Suarez 27 vols, Bellarmine, Cajetan, a Lapide, Banez, Molina)	42	3.9 GB	PDF
16	Patrologia Graeca (159 of 161 Migne PG volumes with OCR text and EPUB)	471	15.5 GB	PDF, TXT, EPUB

Task 17 (Modern Magisterium) contains a download script for post-1928 Vatican documents. The script is MIT-licensed, but the downloaded content is under copyright and is not included in the redistributable corpus.

Text Extraction

The text_extraction/ folder contains a ready-to-run pipeline that processes all content files from Tasks 01–17 (PDF, EPUB, HTML, XML) into unified plain .txt format.

Pre-extracted plain text (47,407 files, 5.7 GB, 2.64B GPT-2 tokens) is available at CatholicCorpus/catholiccorpus-text.

To run it yourself:

cd text_extraction
pip install pdfplumber lxml beautifulsoup4 ebooklib
caffeinate -i python3 extract_text.py

Provenance

Every downloaded file has a companion _source.json sidecar documenting its source URL, license, and download date.

Getting Started

Browse the Corpus

Each collection folder contains texts organized by author or collection, with _source.json sidecars for provenance.

Re-download Everything

Each task folder has an idempotent download_script.py:

pip install requests lxml
cd 13_mystics && python download_script.py

Scripts skip already-downloaded files and resume cleanly after interruption.

To re-download the entire corpus:

pip install requests lxml
bash run_all.sh

Task 02 (Corpus Corporum) is the longest — expect 12–24 hours for the full Patrologia Latina.

Rebuild the Manifest

python build_manifest.py

Regenerates master_manifest.json (machine-readable) and SUMMARY.md (human-readable) from disk state.

Licensing

The vast majority of this corpus is public domain (pre-1928 U.S. publications or authors deceased 70+ years). Two items carry Creative Commons licenses:

SBLGNT: CC BY 4.0 (Society of Biblical Literature)
Rahlfs Septuagint 1935: CC BY-NC-SA 4.0 (Eliran Wong)

See LICENSE_NOTES.md for the full breakdown, and check each file's _source.json sidecar for per-item license details.

Attribution

If you use this corpus in research, please cite:

CatholicCorpus — An open-access corpus of Catholic texts for NLP and digital humanities research. https://catholiccorpus.org

See ACKNOWLEDGMENTS.md for the full list of institutions, projects, and individuals whose work made this possible.

Project Structure

CatholicCorpus/
  01_git_repos/          — Git-cloned text repositories
  02_corpus_corporum/    — Corpus Corporum Latin texts (TEI XML)
  03_gutenberg/          — Project Gutenberg texts
  ...
  16_patrologia_graeca/  — Migne PG volumes
  17_modern_magisterium/ — Download script only (content is copyrighted)
  text_extraction/       — Plain text extraction pipeline (code only, not yet run)
  stats/                 — Tokenization and composition scripts
  docs/                  — Schema documentation
  archive_helpers.py     — Shared archive.org download utilities
  build_manifest.py      — Manifest + summary generator
  master_manifest.json   — Full file inventory (machine-readable)
  dataset.json           — Machine-readable dataset metadata
  SUMMARY.md             — Current corpus statistics
  LICENSE_NOTES.md       — Licensing details
  LIMITATIONS.md         — Known limitations, biases, and scope
  ROADMAP.md             — Versioned release milestones
  CHANGELOG.md           — Version history
  CONTRIBUTING.md        — How to contribute
  ACKNOWLEDGMENTS.md     — Credits and attribution
  BACKLOG.md             — Known gaps and future work
  LENDING_RESTRICTED.md  — Archive.org borrow-only items

Limitations

See LIMITATIONS.md for the full discussion of scope, biases, and known gaps. Key items:

Patrologia Graeca vols 16 & 86: Not found on archive.org (98.8% coverage without them).
23 archive.org items are lending-restricted (borrow-only). See LENDING_RESTRICTED.md.
Text extraction not yet run (code included for review).
Token counts cover text-extractable files only (2.6B tokens counted; PDF-only collections pending text extraction).
OCR quality varies by collection; no systematic scoring performed yet.

Requirements

Python 3.8+
requests and lxml (pip install requests lxml)
~40 GB free disk space for the full corpus
Patience for archive.org rate limits

License

The download scripts in this repository are released under the MIT License. The downloaded content is subject to the individual licenses documented in each file's _source.json sidecar and in LICENSE_NOTES.md.

Downloads last month: 73

Total file size:

38.5 GB