PubMed citation error with PubmedBookArticle PMID:26962613

[agitter]

Does Manubot support parsing PubmedBookArticle? I encountered the following error with manubot cite pubmed:26962613 and Manubot v0.5.1:

## ERROR
Error parsing the following PubMed metadata for PMID 26962613:
...
## WARNING
Generating csl_item for 'pubmed:26962613' failed due to a ValueError:
Expected article to be an XML element with tag PubmedArticle, received tag 'PubmedBookArticle'
[]
## CRITICAL
Failure: exiting with code 1 due to logged errors

The XML is https://eutils.ncbi.nlm.nih.gov/entrez/eutils/efetch.fcgi?db=pubmed&id=26962613&rettype=full

This line suggests PubmedBookArticle is supported

manubot/manubot/cite/pubmed.py

Line 143 in c27e5b6

assert xml_article.tag in ["PubmedArticle", "PubmedBookArticle"]

but this one does not

manubot/manubot/cite/pubmed.py

Line 164 in c27e5b6

if not article.tag == "PubmedArticle":

[dhimmel]

This came up in the past with pubmed:29227604 at #24 (comment).

See also this test:

manubot/manubot/cite/tests/test_citekey_api.py

Lines 179 to 190 in 3ff3000

	def test_citekey_to_csl_item_pubmed_book(caplog):
	"""
	Extracting CSL metadata from books in PubMed is not supported.
	Logic not implemented to parse XML returned by
	https://eutils.ncbi.nlm.nih.gov/entrez/eutils/efetch.fcgi?db=pubmed&id=29227604&rettype=full
	"""
	csl_item = citekey_to_csl_item("pmid:29227604", log_level="ERROR")
	assert csl_item is None
	assert (
	"Expected article to be an XML element with tag PubmedArticle, received tag 'PubmedBookArticle'"
	in caplog.text
	)

So we're aware that it's not supported and adding a PubmedBookArticle parser would be a bit of work. Maybe we could just fall back to the zotero translator here?

[agitter]

So we're aware that it's not supported and adding a PubmedBookArticle parser would be a bit of work.

I agree that adding a new parser isn't going to be high priority, and this is a rare use case. I was somewhat confused because line 143 allows PubmedBookArticle but then that tag isn't supported by a parser. The existing warning message is sufficiently clear though.

Maybe we could just fall back to the zotero translator here?

That seems like the right solution. In my case I switched to the PubMed URL, and Zotero worked well. The question is whether we want the user to change the identifier or try to automate it for them when this error occurs.