Dataset Viewer
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: CastError
Message: Couldn't cast
at: string
stage: string
profile: string
model: string
requested_model: string
temperature: int64
input_tokens: int64
output_tokens: int64
finish: string
cost_usd: double
ms: int64
parsed: null
ocr: struct<model: string>
child 0, model: string
text: string
page: int64
source: string
to
{'source': Value('string'), 'page': Value('int64'), 'text': Value('string'), 'ocr': {'model': Value('string')}}
because column names don't match
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 2951, in __iter__
for key, example in ex_iterable:
^^^^^^^^^^^
File "/usr/local/lib/python3.14/site-packages/datasets/iterable_dataset.py", line 2461, in __iter__
for key, pa_table in self._iter_arrow():
~~~~~~~~~~~~~~~~^^
File "/usr/local/lib/python3.14/site-packages/datasets/iterable_dataset.py", line 2486, 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 547, in _iter_arrow
for key, pa_table in iterator:
^^^^^^^^
File "/usr/local/lib/python3.14/site-packages/datasets/iterable_dataset.py", line 430, 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 2306, in cast_table_to_schema
raise CastError(
...<3 lines>...
)
datasets.table.CastError: Couldn't cast
at: string
stage: string
profile: string
model: string
requested_model: string
temperature: int64
input_tokens: int64
output_tokens: int64
finish: string
cost_usd: double
ms: int64
parsed: null
ocr: struct<model: string>
child 0, model: string
text: string
page: int64
source: string
to
{'source': Value('string'), 'page': Value('int64'), 'text': Value('string'), 'ocr': {'model': Value('string')}}
because column names don't matchNeed help to make the dataset viewer work? Make sure to review how to configure the dataset viewer, and open a discussion for direct support.
Myrias corpus
Source material and run outputs for Myrias, an open,
bilingual, source-linked evidence base on welfare technology for farmed aquatic animals. Every source PDF is
in raw/, with one index; the pipeline's inputs and outputs are in one folder per species, each with a
MANIFEST.json describing every file.
Private. The files include third-party abstracts and full texts under many licences. Do not make this repository public, or share files from it, until each licence has been checked.
| Folder | What it holds |
|---|---|
raw/ |
Every source PDF, organised as raw/<species>/zh/ and raw/<species>/en/, named year_author_title.pdf, with raw/index.csv (one row per PDF: bibliographic details, topic, relevance, origin, whether the pipeline has read it, checksum, old location). See raw/README.md |
<species>/intake/ |
New search across Europe PMC, Crossref, DOAJ, OpenAlex and Semantic Scholar, with the open full text it could fetch (the PDFs are now in raw/) |
<species>/first-run/ |
The first run on 20 Chinese papers: page text (source PDFs now in raw/), extraction, verification, mapping, English statements, model calls |
<species>/gap-pilot/ |
Language-gap pilot: English evidence for each intervention found in Chinese |
<species>/ground-truth/ |
Checked answers to score the pipeline against |
<species>/config/ |
Settings |
<species>/run-2/ |
Run 2: phase A re-read the Chinese papers with the current extraction; phase B read the 176 open-access papers. Each step has a numbered folder with its record |
<species>/store/ |
The species store: every claim that has passed in any run, merged, with permanent ids |
<species>/links.csv |
Every link in one table |
Built with pipeline/intake.mjs, pipeline/normalize.py and pipeline/run.mjs. Search and download
used no model; the first run's extraction, verification and mapping did, and every call is logged.
- Downloads last month
- 1,724