I've been reading mostly the archives and the GitHub tickets so far, but given the interest in the Primary Sources Tool maybe it's time I joined the mailinglist ;-)

Following up on the discussions the last weeks I added two new features to the backend:

1. for any request listing statements it is now possible to filter by state, with the default being "unapproved"

For example, you can select 10 random statements that have been marked "wrong" (i.e. rejected in the UI) with

curl -i "https://tools.wmflabs.org/wikidata-primary-sources/statements/any?state=wrong" 

or retrieve all approved statements with

curl -i "https://tools.wmflabs.org/wikidata-primary-sources/statements/all?state=approved" 

(NB: the /all endpoint is using paging, use the offset= and limit= parameters to control how much is returned)

the different acceptable states are defined in https://github.com/google/primarysources/blob/master/backend/Statement.h#L14

2. all statements that already had some form of interaction (e.g. have been approved or rejected) now contain a new JSON field "activities" listing the activities acting on the statement; even though usually there will be at most one activity (i.e. approved or rejected), the system stores (and already stored since we launched it) a complete history, e.g. for transitions like unapproved -> wrong -> unapproved -> approved.

You can try it out by retrieving a random selection of statements in other states than "unapproved", e.g. as before:

will give you results like:

Hope it is useful ;-)

Otherwise let me know if you are interested in other analysis data. I'll try adding features as time permits.

Cheers!