indra literature id_lookup bug

[Crispae]

I was checking Indra's literature id_lookup functionality. While i look up literature using doi it doesn't return PMID or PMCID. But , when i used PMID of the same literature it return the same DOI. I guess reverse mapping is not available.

[Crispae]

This endpoint of NCBI: https://eutils.ncbi.nlm.nih.gov/entrez/eutils/esearch.fcgi?db=pubmed&term={doi}[doi] will return PMID from DOI. It can be implemented for reverse.

Replace {doi} with original DOI.

[Crispae]

I guess the Endpoint being used in Indra: https://www.ncbi.nlm.nih.gov/pmc/utils/idconv/v1.0/ is not, may be I am implementing this Endpoint wrong.

https://github.com/sorgerlab/indra/blob/master/indra/literature/pmc_client.py. In this an additonal url(above mentioed) for DOI can be added, that will handle the DOI to PMID conversion.

[bgyori]

Hi @Crispae, the issue is indeed that there are multiple endpoints (e.g., idconv vs esearch) through multiple web services (PubMed, PMC, CrossRef) that can be used to convert between paper identifiers and they all provide somewhat different results. The indra.literature.id_lookup function uses https://www.ncbi.nlm.nih.gov/pmc/utils/idconv/v1.0/ which fails to retrieve other IDs for the DOI in your example even when called outside the context of INDRA. This is not generally the case though, here is an example DOI for which the lookup works:

In [1]: from indra.literature import id_lookup

In [2]: id_lookup('10.1093/nar/gks1195', 'doi')
Out[2]: {'doi': '10.1093/nar/gks1195', 'pmid': '23193287', 'pmcid': 'PMC3531190'}

If you have a specific set of DOIs you want to convert, you might want to test outside webservices like different endpoints that PubMed, PMC and CrossRef provide to see which of those works well on those.

[bgyori]

I'll close this, @Crispae if you have any further questions, please reach out.