Document post-mortem: Missing metadata and misc problems

[jvwong]

PubMed: Coordinated Conformational Processing of the Tumor Suppressor Protein p53 by the Hsp70 and Hsp90 Chaperone Machineries
Document URL: https://biofactoid.org/document/e3ad1fe0-7cc9-4075-b9a7-32626ce0c2ff

Problems

• 1. Biofactoid could not identify their article based on author input
  • Author added variant of the PubMed indexed title Coordinated conformational processing of p53 by the Hsp70 and Hsp90 chaperones machineries
  • It's worth noting that EUTILS ESEARCH returns nothing; PubMed website returns the correct article
• 2. No related papers
  • Document: No PubMedID
  • Network: INDRA responded with 504: Gateway Timeout
• 3. Display in carousel is different than explorer
  • Carousel:

• explorer

[maxkfranz]

Re. (iii): Did the user submit early, or might this have been related to the previous status values?

Re. (ii): Is this a one-off Indra issue for this document (https://biofactoid.org/document/a6b4351d-b474-49d9-bc3d-f044513d5a0c)? Could Metin add more retries to the Indra queries to resolve this? Or is this a reproducible error for the document specifically?

Another issue (iv): Is there the wrong grounding for this document https://biofactoid.org/document/aa18ee18-a78f-4d46-a244-9a6dc97bb238 ? 'Cofilin' doesn't exist for mouse in NCBI. Only 'cofilin-1' exists for mouse as the closest synonym.

[jvwong]

Re. (iii): Did the user submit early, or might this have been related to the previous status values?

Likely submitted followed by editing. I told one user that didn't like their submission that they could still edit the document via the same link.

Re. (ii): Is this a one-off Indra issue for this document (https://biofactoid.org/document/a6b4351d-b474-49d9-bc3d-f044513d5a0c)? Could Metin add more retries to the Indra queries to resolve this? Or is this a reproducible error for the document specifically?

One-off: I could update successfully by refreshing from the admin.

[jvwong]

Another issue (iv): Is there the wrong grounding for this document https://biofactoid.org/document/aa18ee18-a78f-4d46-a244-9a6dc97bb238 ? 'Cofilin' doesn't exist for mouse in NCBI. Only 'cofilin-1' exists for mouse as the closest synonym.

yeah probably mouse cofilin-1

[maxkfranz]

Another issue (iv): Is there the wrong grounding for this document https://biofactoid.org/document/aa18ee18-a78f-4d46-a244-9a6dc97bb238 ? 'Cofilin' doesn't exist for mouse in NCBI. Only 'cofilin-1' exists for mouse as the closest synonym.

yeah probably mouse cofilin-1

We could add to the patch file, but only if we’re sure users would always want cofilin-1 and not something like cofilin-2.

[jvwong]

Another issue (iv): Is there the wrong grounding for this document https://biofactoid.org/document/aa18ee18-a78f-4d46-a244-9a6dc97bb238 ? 'Cofilin' doesn't exist for mouse in NCBI. Only 'cofilin-1' exists for mouse as the closest synonym.

yeah probably mouse cofilin-1

See #861

[jvwong]

Re. (iii): Did the user submit early, or might this have been related to the previous status values?

Happened again with https://biofactoid.org/document/b2b68594-0ab2-4fdb-8305-3631b7b40319

[jvwong]

@maxkfranz re: author editing after the doc after a png has been created/cached

• one idea is to cache the doc lastEditedDate with the doc id to see if the .png needs to be (re)generated ?

[maxkfranz]

You'd want the timestamp in the cache, not with the ID as the key. But that would work for these cases where the user fiddles with things post-submission.

In future, we may also want a general interval-based cache clearing (e.g. nightly) to catch invalidations that aren't user edits (e.g. via admin ui).

[maxkfranz]

You can get the interval behaviour by just setting maxAge: https://www.npmjs.com/package/lru-cache

[jvwong]

I think this png issue is the only actionable item here.

[maxkfranz]

(iv) can be addressed by the patch file in the grounding service.

PathwayCommons/grounding-search#84