Description
This is an extreme corner case, but it is an actual real life case, but just rare. I'm not calling commonmeta-py so I'm not affected by this. General CSL file use is too challenging/buggy and commonmeta-py has too many dependencies.
I don't think this corner case is an important issue. But WFIW, this is a test case that epijats handles properly.
The JATS XML extracted from PMC and re-indented is here:
It comes from reference #26 of PMC11759912. I attached a csljson file generated from that JATS XML: input.json
Repo steps
commonmeta convert --via csl input.json --to citation
Result
Daniel, S., Venkateswaran, C., Hutchinson, A., & Johnson, M. (2021). 'I don't talk about my distress to others; I feel that I have to suffer my problems.' voices of indian women with breast cancer: a qualitative interview study. In <i>Support Care Cancer</i> (Vol. 29, Number 5, pp. 2591–2600).
Expected
The correct title has ... following the word problems in the title, not just a single period .
Commentary
I saw the commonmeta-py code has a reasonable hack to work around a limitation of citeproc-py. I've given up attempting to work with CSL files in general and instead just plan to support only a few (probably only 2) canned styles which are only loosely based on the Vancouver style. Technically I'm still using citeproc-py and two canned CSL files inside epijats but this is purely an implementation detail that is not intended to inter-operate with other CSL files.