Querying structural alerts with CHeMBL currently does not work

[jranke]

Currently, the test for the ChEMBL resource "compound_structural_alert" is disabled (PR #442), because it fails.

The URL put together in the function foo() is

https://www.ebi.ac.uk/chembl/api/data/compound_structural_alert/79048021.json

which gives a 404. A working URL for returning structural alerts [1] is e.g.

https://www.ebi.ac.uk/chembl/api/data/compound_structural_alert.json?molecule_chembl_id=CHEMBL266429

[1] chembl/chembl_webservices_2#109 (comment)

[stitam]

Thanks @jranke for flagging this. While most urls are constructed using the same pattern, it seems there are exceptions and "compound_structural_alert" is one of them ("similarity" is another exception)

[jranke]

Yes, this seems to be an inconsistency on the server side... but could be worked around in our package as well, I just do not need this right now so I won't bother to do it.

[stitam]

I already fixed this for "similarity" on my fork, so the framework is there. I'll try to wrap up and then I'll add an exception for "compound_structural_alert" as well.

[stitam]

@jranke, chembl_query() should now work with compound structural alerts. I think we're not 100% there yet:

• The response is paginated and the url seems to return the first page. I think we should hide pagination and return all results.
• Default formatting does not format the output correctly for this resource. Because of this I added a "tidy" argument which is TRUE by default but turned it off for the compound structural alerts example. Not sure whether we should keep this new argument for long term (adds a mechanism to return the raw response for those who do not like the way we tidy the responses) or not (simpler UI, we should tidy in a way that everyone likes).

I think we should keep this issue open until access to the resource is fully fixed.

[jranke]

@stitam OK, we can leave this open, but I currently do not plan to work on it. Thanks for your patience and efforts with this!

[stitam]

That's fine. I'm working on ChEMBL stuff anyway so I may work on this in the coming weeks.