Closes #113 - Add Chebi (Chapti) #525
Conversation
napsternxg
requested review from
hakunanatasha,
jason-fries,
sunnnymskang,
ruisi-su,
galtay,
leonweber,
sg-wbi and
debajyotidatta
as code owners
sg-wbi
changed the title
|
@phlobo Please have a look at the implementation. |
There was a problem hiding this comment.
@mariosaenger I have some questions regarding this dataset, could you please have a look?
| issn = {0305-1048}, | ||
| pages = {D344—50}, | ||
| url = {https://europepmc.org/articles/PMC2238832}, | ||
| biburl = {https://aclanthology.org/W19-5008.bib}, |
There was a problem hiding this comment.
Looks like biburl and bibsource belong to a different dataset
| issn = {0305-1048}, | ||
| pages = {D344—50}, | ||
| url = {https://europepmc.org/articles/PMC2238832}, | ||
| biburl = {https://aclanthology.org/W19-5008.bib}, |
There was a problem hiding this comment.
Looks like biburl and bibsource belong to a different dataset
|
|
||
| DATA_URL = "https://github.com/bigscience-workshop/biomedical/files/8568960/PatentAnnotations_GoldStandard.tar.gz" | ||
| _URLS = { | ||
| # The original dataset is hosted on CVS on sourceforge. Hence I have downloaded and reuploded it as tar.gz format. |
There was a problem hiding this comment.
The provenance of this dataset seems to be tricky. It is not mentioned in the original ChEBI publication in NAR, that we have as a citation. Is there any other source of information about the annotation project? Otherwise, it is a bit hard to tell if we got the number of annotations, IDs, etc. right.
| # Converted via the following command: | ||
| # cvs -z3 -d:pserver:[email protected]:/cvsroot/chebi co \ | ||
| # chapati/patentsGoldStandard/PatentAnnotations_GoldStandard.tgz | ||
| # mkdir -p ./MoNERo |
There was a problem hiding this comment.
The MoNERo part belongs to a different dataset?
| "offsets": [[e["start"], e["end"]]], | ||
| "type": e["attrs"]["type"], | ||
| "normalized": [ | ||
| {"db_name": "chebi", "db_id": chebi_id.strip()} for chebi_id in e["attrs"]["chebi-id"].split(",") |
There was a problem hiding this comment.
The IDs extracted this way look inconsistent, e.g.:
{'id': 'WO2007000651-E8',
'type': 'CHEMICAL',
'text': ['Zinc oxide'],
'offsets': [[613, 623]],
'normalized': [{'db_name': 'chebi', 'db_id': 'CHEBI:36560'}]},
{'id': 'WO2007000651-E9',
'type': 'ONT',
'text': ['astringent'],
'offsets': [[690, 700]],
'normalized': [{'db_name': 'chebi', 'db_id': 'WO2007000651:157583'}]},
Maybe the db_id should just contain the last numerical bit?
| "offsets": [[e["start"], e["end"]]], | ||
| "type": e["attrs"]["type"], | ||
| "normalized": [ | ||
| {"db_name": "chebi", "db_id": chebi_id.strip()} for chebi_id in e["attrs"]["chebi-id"].split(",") |
There was a problem hiding this comment.
Moreover, there seem to be more identifiers attached to each entity, e.g., in the source version there are entries like 'epochem-id': 'EPOCHEM:NEW:CLASS:4',. Shall we include them as additional normalized entries with db_id : epochem? Might be a relevant NED task for some users of the dataset.
Please name your PR after the issue it closes. You can use the following line: "Closes #ISSUE-NUMBER" where you replace the ISSUE-NUMBER with the one corresponding to your dataset.
Fixes #113
If the following information is NOT present in the issue, please populate:
Checkbox
biodatasets/my_dataset/my_dataset.py(please use only lowercase and underscore for dataset naming)._CITATION,_DATASETNAME,_DESCRIPTION,_HOMEPAGE,_LICENSE,_URLs,_SUPPORTED_TASKS,_SOURCE_VERSION, and_BIGBIO_VERSIONvariables._info(),_split_generators()and_generate_examples()in dataloader script.BUILDER_CONFIGSclass attribute is a list with at least oneBigBioConfigfor the source schema and one for a bigbio schema.datasets.load_datasetfunction.python -m tests.test_bigbio biodatasets/my_dataset/my_dataset.py.