Searching for works by an author fails

[Edward Betts]

I should be able to use the API to search for all works by an author.

I would expect this URLs to return a list of works by William Shakespeare:

http://openlibrary.org/query.json?type=/type/work&authors.author=/a/OL9388A

http://upstream.openlibrary.org/query.json?type=/type/work&authors.author=/a/OL9388A

They don't give an error, I just get an empty list.

It doesn't work because they're linked by an embedded type.

This would be very useful when working with the data in the database.

[Anand Chitipothu (anandology)]

On 17-Feb-10, at 11:00 PM, Edward Betts wrote:

> Public bug reported:
>
> I should be able to use the API to search for all works by an author.
>
> I would expect this URLs to return a list of works by William
> Shakespeare:
>
> http://openlibrary.org/query.json?type=/type/work&authors.author=/a/OL9388A
>
> http://upstream.openlibrary.org/query.json?type=/type/work&authors.author=/a/OL9388A
>
> They don't give an error, I just get an empty list.
>
> It doesn't work because they're linked by an embedded type.
>
> This would be very useful when working with the data in the database.

Use this instead:

http://upstream.openlibrary.org/query.json?query={"type": "/type/
work", "authors": {"author": {"key": "/authors/OL9388A"}}}

[Edward Betts (edwardbetts)]

That includes duplicates:

http://upstream.openlibrary.org/query.json?query={%22limit%22:500,%22type%22:%20%22/type/work%22,%20%22authors%22:%20{%22author%22:%20{%22key%22:%20%22/authors/OL32983A%22}}}

/works/OL495477W appears more than one

[Edward Betts (edwardbetts)]

Also some searches are missing results:

from openlibrary.api import OpenLibrary, Reference
ol = OpenLibrary("http://openlibrary.org")
q = {'type': '/type/work', 'authors': {'author': Reference('/a/OL2837555A')}}
print ol.query(q)
print ol.get('/works/OL8493302W')['authors']

Gives:

[]
[{'type': '/type/author_role', 'author': <ref: u'/a/OL2837555A'>}]

[George (george-archive)]

Bumping this in priority (High -> Critical) since it's a blocker for Edward.

[Anand Chitipothu (anandology)]

Started reindexing the work_ref table.

  3.4% ( 479397/14091822) [=> ] ETA 25:15:14

[Anand Chitipothu (anandology)]

Finished reindexing the work_ref table. Queries should work now.

Edward, please verify and close this bug.

[Edward Betts (edwardbetts)]

Anand, I ran my test from a previous bug comment and I get the same results, so the bug isn't fixed.

[Edward Betts (edwardbetts)]

I rewrote my example, now it works:

from openlibrary.api import OpenLibrary, Reference
ol = OpenLibrary("http://openlibrary.org")
q = {'type': '/type/work', 'authors': {'author': {'key': '/a/OL2837555A'}}}
print ol.query(q)
print ol.get('/works/OL8493302W')['authors']

Output is:

[<ref: u'/works/OL8493303W'>, <ref: u'/works/OL8493302W'>]

So I guess it is fixed.

[Edward Betts (edwardbetts)]

It is still happening. Here is an example

from openlibrary.api import OpenLibrary
ol = OpenLibrary("http://openlibrary.org")
q = {'type': '/type/work', 'authors': {'author': {'key': '/authors/OL6876811A'}}}
print ol.query(q)
print ol.get('/works/OL15598361W')['authors']

Output is:

[]
[{'type': <ref: u'/type/author_role'>, 'author': <ref: u'/authors/OL6876811A'>}]

[George (george-archive)]

Anand - this bug is critical. Please address this week.

[Anand Chitipothu (anandology)]

Noticed that none of the work-author queries are working. There was a NULL == NULL comparison in the SQL query. Fixed it by avoid that case.

$ curl -G http://openlibrary.org/api/things --data-urlencode 'query={"type": "/type/work", "authors": {"author": {"key": "/authors/OL6876811A"}}}'
{"status": "ok", "result": ["/works/OL15598357W", "/works/OL15598361W"]}

Edward: can you please verify?