Dataset Viewer
Duplicate
The dataset viewer is not available for this split.
Cannot load the dataset split (in streaming mode) to extract the first rows.
Error code:   StreamingRowsError
Exception:    TypeError
Message:      Couldn't cast array of type string to null
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 2118, in cast_array_to_feature
                  casted_array_values = _c(array.values, feature.feature)
                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 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 2152, in cast_array_to_feature
                  return array_cast(
                      array,
                  ...<2 lines>...
                      allow_decimal_to_str=allow_decimal_to_str,
                  )
                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 2014, in array_cast
                  raise TypeError(f"Couldn't cast array of type {_short_str(array.type)} to {_short_str(pa_type)}")
              TypeError: Couldn't cast array of type string to null

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.

INNER ATLAS — display shards

828,131 rights-cleared quotations in 31 languages, projected for display across 3,330 shards and served to the browser one shard at a time, with an inverted index for searching them.

This is a derived artifact, not a corpus. It carries only the fields a reader sees, and it exists so that a static site with no runtime database can show the whole corpus rather than a sample. For the corpora themselves, with full provenance, verification status and per-record rights blocks, use the source datasets.

Layout

Path What it holds
manifests/<lang>.json.gz every shard and search bucket for that language, with digests
shards/<lang>/NNNN.json.gz display records, 250 per shard, ordered by author
search/<lang>/NN.json.gz inverted index, 64 buckets keyed by a hash of the token

Languages

Language Quotations Shards Searchable With another rendering
English (en) 427,226 1709 yes 101,383
Pali (pli) 127,428 510 yes 61,182
Ancient Greek (grc) 105,730 423 yes 23,288
German (de) 49,068 197 yes 41,897
Russian (ru) 30,102 121 yes 26,471
Italian (it) 20,690 83 yes 14,504
Serbian (sr) 20,158 81 yes 19,062
Latin (la) 18,629 75 yes 6,842
Spanish (es) 8,809 36 yes 602
Lithuanian (lt) 5,731 23 yes 5,435
French (fr) 4,937 20 yes 4,410
Portuguese (pt) 2,331 10 yes 2,295
Polish (pl) 1,738 7 browse only 1,668
Hebrew (he) 1,648 7 browse only 349
Gujarati (gu) 948 4 browse only 926
Hindi (hi) 594 3 browse only 588
Vietnamese (vi) 587 3 browse only 264
Turkish (tr) 384 2 browse only 201
Thai (th) 382 2 browse only 381
Lao (lo) 318 2 browse only 318
Georgian (ka) 261 2 browse only 229
Estonian (et) 182 1 browse only 136
Indonesian (id) 101 1 browse only 84
Undetermined (und) 82 1 browse only 0
Czech (cs) 36 1 browse only 31
Mixed (misc) 11 1 browse only 0
Finnish (fi) 10 1 browse only 10
Swiss German (gsw) 4 1 browse only 3
Prakrit (pra) 3 1 browse only 3
Sanskrit (san) 2 1 browse only 0
Burmese (my) 1 1 browse only 0

12 of 31 languages carry a search index; the rest have too few records for searching to mean anything.

Parallel renderings

Renderings of one utterance share a work_key. Which of them is a source text is declared per provider from that provider's own model, never inferred from the records: the corpus does not say. An unmarked rendering is unstated, not a translation.

  • suttacentral: pli, pra, san
  • perseus: grc, la
  • sefaria: he

Licence

CC BY-SA 4.0. A combined artifact drawing on four corpora, some of them ShareAlike, so the combination is distributed under CC BY-SA 4.0. Each record carries its own licence and attribution, and both must be displayed with it.

Integrity

Every manifest's sha256 is recorded in src/data/atlas/corpus-index.json in the INNER ATLAS repository, which ships inside the published site rather than beside these files; each manifest carries the digest of every shard and bucket it names. So the origin serving the page vouches for the manifest and the manifest for the bytes, and a file that does not match is refused by the client rather than displayed.

Shards are deterministic: the same records produce the same bytes.

Downloads last month
178