a few common special cases

[daryanypl]

Sister Souljah -- "Sister" is more of a title than last name.
His Holiness the Dalai Lama -- "His Holiness" -- the two words together is the title.
Bob Jones, composer
Bob Jones, author
Bob Jones, compositeur -- (French for composer)

Here's the code I used:

        name = HumanName('Sister Souljah')
        library_sort_name = u' '.join([name.first, name.middle, name.suffix, name.nickname, name.title])
        if name.last:
            library_sort_name = u''.join([name.last, ", ", library_sort_name])
        print "library_sort_name=%s" % library_sort_name
        
        name = HumanName('His Holiness the Dalai Lama')
        library_sort_name = u' '.join([name.first, name.middle, name.suffix, name.nickname, name.title])
        if name.last:
            library_sort_name = u''.join([name.last, ", ", library_sort_name])
        print "library_sort_name=%s" % library_sort_name
        
        name = HumanName('Bob Jones, author')
        library_sort_name = u' '.join([name.first, name.middle, name.suffix, name.nickname, name.title])
        if name.last:
            library_sort_name = u''.join([name.last, ", ", library_sort_name])
        print "library_sort_name=%s" % library_sort_name
        
        name = HumanName('Bob Jones, compositeur')
        library_sort_name = u' '.join([name.first, name.middle, name.suffix, name.nickname, name.title])
        if name.last:
            library_sort_name = u''.join([name.last, ", ", library_sort_name])
        print "library_sort_name=%s" % library_sort_name

[derek73]

I added some things to the titles constant.

It's not clear what you would want to do with the Dalai Lama because I think technically the entire thing is more of a title, depending if you believe in reincarnation I guess.

"Sister Souljah" is unchanged below, is it not correct?

Here's what it does now:

$ python tests.py "Sister Souljah"
<HumanName : [
	title: 'Sister' 
	first: 'Souljah' 
	middle: '' 
	last: '' 
	suffix: ''
	nickname: ''
]>
$ python tests.py "His Holiness the Dalai Lama"
<HumanName : [
	title: 'His Holiness the Dalai' 
	first: '' 
	middle: '' 
	last: 'Lama' 
	suffix: ''
	nickname: ''
]>
$ python tests.py "Bob Jones, composer"
<HumanName : [
	title: 'composer' 
	first: 'Bob Jones' 
	middle: '' 
	last: '' 
	suffix: ''
	nickname: ''
]>

[derek73]

btw, what are you doing with library_sort_name? I wonder if you could do something overriding __eq__()?

[daryanypl]

Good question, and thanks for working on it. I started writing the answer, and realized I need to consult with a coworker or two who are experts on the library science name parsing topic. I'll email around and get back to you. I'm a software engineer for the New York Public Library. I use HumanName to convert author names from the form they're displayed in ("Bob Jones") to the form they get sorted in ("Jones, Bob"). I also use it in a bit of code that normalizes names before I try to determine the likelihood that two authors from two different books are actually the same person (what's the chance that Bob Jones and Robert L. Jones are the same person?). The code is in its early stages, and the data is complex. For an example, check http://viaf.org/viaf/search?query=local.personalNames+all+%22mark%20twain%22&sortKeys=holdingscount&recordSchema=BriefVIAF . VIAF gets author names from library consortia all over the world. So Bob Jones writes "My Memoirs", that book is owned by some libraries in the US and let's say Netherlands. The US consortium tells VIAF they have a book "My Memoirs" written by Robert Jones, the Netherlands consortium tells VIAF they have a book with the translated title "Happy Life", published originally as "My Memoirs" by Bob Jones, Jr, with Alice Kale translator, and then VIAF creates records for two Bobs and one Alice. So when a library gets books, their authors are coming in in metadata in various formats from the reseller/open sourcer/etc., and we try our best to match those names against each other and VIAF and let them be sortable for library patrons. Best, Darya

…

On Wed, Feb 15, 2017 at 12:01 AM, derek73 ***@***.***> wrote: btw, what are you doing with library_sort_name? I wonder if you could do something overriding __eq__()? — You are receiving this because you authored the thread. Reply to this email directly, view it on GitHub <#57 (comment)>, or mute the thread <https://github.com/notifications/unsubscribe-auth/ARciorHFn-6T0KsQdjX9CWBx6nIpyV9Bks5rcoavgaJpZM4MBFkm> .

[daryanypl]

I wrote a little script that went through the MARC21 records in VIAF, looking at section 100, field "c", which is the field where librarians write down authors' personal titles. I wrote out the more common titles into the attached file. The good: these are awesome, from "anarchist" to "stroke survivor". The bad: they're a bit too much -- it would be rare that such detailed professional description is used in the author's name field. Still, it might be the best we can do for now.

[daryanypl]

contributor_titles_summary_1.txt

[derek73]

I added most of these names to the titles set and published them in v0.5.2.

Probably most nouns could be part of a title. Probably with some NLP and ML stuff you could get pretty good at figuring out the name parts, but for now those titles all seemed pretty on par with what we already have in there so may as well add them. Nice that they came from actual data someplace instead of just me putting my thinking cap on.