Improve name parsing

[larsgw]

Examples:

Input	Current out	Expected out
Rossana De Leo	'Rossana De', 'Leo'	'Rossana', 'De Leo'
Rossana de Leo	'Rossana', 'de'	'Rossana', 'de Leo'

---

Possible solutions:

• Use of P735, P734 (not always possible)
• Use of builtin tussenvoegsels (and similar) [edit: fixed in db60dd4 (v0.3.3)]
• Fix whatever bug causes the second example [edit: fixed in 2ad82af (v0.3.1)]

[digitalheir]

Hey. I stumbled on this project and figured you might want to re-use some of my stuff. I've made a pretty decent bibtex parser: https://github.com/digitalheir/bibtex-js/

See if you can use it

[larsgw]

Note: known issue of the current parser is that it can't handle lowercase particles in the middle of the family name (i.e. Given Family_1 y Family_2) if there are no lowercase particles at the start of the family name (e.g. with Pablo Diego Ruiz y Picasso). It's basically impossible to determine whether Diego and Ruiz are family names or given names (without understanding of the language, that is). The only possibility (AFAICT) would be to make a special case for y (assuming there is only ever one family name before the first y), and that's annoying and difficult, and not really clean code.

---

@digitalheir I'll take a look at it, thanks!

[digitalheir]

Don't try to be smarter than the spec, I guess. :^)

Standard BibTeX behaviour is to treat all capitalized names before "y" as first names, ie Firstname von Lastname. If user wants Ruiz it to be last names user should re-format the field as Ruiz y Picasso, Pablo Diego.

https://github.com/digitalheir/bibtex-js/blob/master/src/bibfile/bib-entry/bibliographic-entity/Author.ts

See function parseAuthorName.

[larsgw]

Well, this name parsing function is used in other parsers (like Wikidata) as well, so I was talking more generally.

[digitalheir]

Ah yeah. Parsing names can be a real headache generally. I had the same problem when I tried to look for Dutch names in a big collection of text files. In the end I just prepared a database of known last names:

https://github.com/digitalheir/family-names-in-the-netherlands

Meertens also keep a list of first names but I think it's a little harder to scrape.

[larsgw]

So there's another bug...

'First M. Last, Jr.' => {given: 'Jr.', family: 'First M. Last'}