Create dataset loader for BioASQ Task A (2013-2021)

[jason-fries]

Adding a Dataset

• Name: BioASQ v.2013 Task A
• Description: Labeling abstracts with MESH Keywords
• Task: TEXT
• Paper: https://bmcbioinformatics.biomedcentral.com/articles/10.1186/s12859-015-0564-6
• Data: http://participants-area.bioasq.org/general_information/Task9b/
• License: DUA

[mdmanurung]

#self-assign

[jason-fries]

Hi @mdmanurung can you let us know if you are still working on this so we can update our project board? Please just notify us the status by Friday April 8. You can response to this comment or ping us on Slack or Discord.

No worries if you are not finished but still intend to work on this!

[balikasg]

@jason-fries
Before committing a question:

• BioASQ Task A articles, while they are released every year, each year's set is a superset of previous years. Do we need a loader for all years here or just 2022? reference here. Also, the link for data in this description is wrong (we want the Task A data, you seem to already have added task B)

[hakunanatasha]

@balikasg Apologies on the incorrect link; this must have cropped up during our large compilation phase. Do you have the link to the most recent version?

Additionally - I suspect yes; there are possibly earlier bodies of work that reference an older version of the dataset. For reproducibility sake, it's nice to have a dataloader that has explicitly what we want, without needing to subset the earlier ones. That being said, thank you for letting us know about it.

Not to presumptively volunteer you, but if you think it may be useful to write a general dataloader for BioAsq based on earlier ones (with a changeable URL) or one for 2022, we would be very grateful!

[MFreidank]

#self-assign

[MFreidank]

@jason-fries @hakunanatasha @galtay I noticed that a dataset bioasq.py already exists (implemented by @galtay), but only seems to implement task B.
Should I implement task A into this same file, or as a separate one (e.g. named biodatasets/bioasq_task_A)?

[jason-fries]

Hi @MFreidank
Sorry for the delay in responding! We should rename bioasq.py to bioasq_task_b and use the full names for all the bioasq tasks. I would go with bioasq_task_a for this task.

[MFreidank]

Hi @jason-fries

Thank you for the clarification, I'll proceed that way (have most of the changes ready and working locally).
If I open a PR for it today, can we still make it count towards the hackathon? ;)

Took a bit longer than I had hoped, but I've now opened PR #518 .