Booting the project!

[Amanpreet Singh]

Dear everybody,
So its time to start Google Summer of Code 2014.

First steps first:
I have created two public repositories on Github regarding the project.
First is the Wikidata Annotation Tool itself: https://github.com/apsdehal/WikidataAnnotationFeeder

Second is the Python Bot, I named it Teressa: https://github.com/apsdehal/Teressa

You may watch or star these repositories if you would like to receive regular and quick updates on the project.
Proposal is still hosted at: https://www.mediawiki.org/wiki/Wikidata_annotation_tool

I am going to post updates about the project in a separate namespace under https://www.mediawiki.org/wiki/Wikidata_annotation_tool/milestones and also on my blog.
Please be sure to check them as you feel comfortable.

Now I want to know what are the next steps that I should take, like which documentation I should probably read etc. Mentors kindly guide me on this.

Regards
Amanpreet Singh 

[Simone Fonda]

On Wed, Apr 23, 2014 at 6:07 PM, Amanpreet Singh
<amanpreet...@gmail.com> wrote:

> First steps first:
> I have created two public repositories on Github regarding the project.
> First is the Wikidata Annotation Tool itself:
> https://github.com/apsdehal/WikidataAnnotationFeeder
> Second is the Python Bot, I named it Teressa:
> https://github.com/apsdehal/Teressa

Great stuff!


> [ .. ]

> Now I want to know what are the next steps that I should take, like which
> documentation I should probably read etc. Mentors kindly guide me on this.

Okie dokie!

Did you have a good read on those topics we already talked about
(predicates, vocabolaries, configuration, ..) ?

If you want to develop a new GUI which does annotations, you might
want to have a look at the triple composer to see how it serializes
back Items, ready to be saved to the server.

Moreover, i would look at the API documentation, to have an idea on
the ways you can interact with the server.


I dont remember if we had a final decision on all of the questions we
raised during the discussion phase (how/where to save, how to model
the knowledge, new GUI or not, etc).

Simone

[David Cuenca]

On Mon, Apr 28, 2014 at 12:51 PM, Simone Fonda <fo...@netseven.it> wrote:

I dont remember if we had a final decision on all of the questions we

raised during the discussion phase (how/where to save, how to model
the knowledge, new GUI or not, etc).

IIRC, it was mentioned that annotations would be saved in the Pundit server and then imported by a bot running on Wikimedia Labs [1] or transferred via API [2].

The bot approach is the easiest, but maybe it is not possible to preserve the information about who created the annotation, which would make harder to identify and revert vandalism. If done via API, then OAuth login should be considered, either using the same OpenID login Wikimedia->Pundit (which is unlikely to be possible any time soon [3]), or by requesting the wikimedia oauth access. [4]

For an example of Oauth login, see:

Application to transfer files from archive.org to Commons: https://tools.wmflabs.org/ia-upload

Source code: https://github.com/Tpt/ia-upload

Eventually a future expansion of the project could be to use a Mediawiki installation as a Pundit annotation server, but I think that is out of scope for now.

My recommendation is to focus first on getting Pundit to annotate with Wikidata vocabulary and generate a "permanent annotation id URL" (if not done yet), and later on work on the annotation transfer Pundit->Wikidata.

Cheers,

Micru

[1] https://www.wikidata.org/wiki/Wikidata:Creating_a_bot

[2] https://www.mediawiki.org/wiki/Extension:Wikibase_Repository/API

[3] https://bugzilla.wikimedia.org/show_bug.cgi?id=9604

[4] https://www.mediawiki.org/wiki/OAuth/For_Developers

[Amanpreet Singh]

On Mon, Apr 28, 2014 at 12:51 PM, Simone Fonda <fo...@netseven.it> wrote:

I dont remember if we had a final decision on all of the questions we

raised during the discussion phase (how/where to save, how to model
the knowledge, new GUI or not, etc).

We have discussed to save it at Pundit Annotation sever and later use a bot to feed Wikidata from it. Also we finalized on a RFC based model where Pundit's Subject-Predicate-Object would be fed to Wikidata as Item-property-value.
For e.g. In Pundit:

London(Subject) - population(Predicate) - "8173456"(Object)

- and a reference

In Wikidata:

London(Item) - population(property) - "8173456"(value)

 

The bot approach is the easiest, but maybe it is not possible to preserve the information about who created the annotation, which would make harder to identify and revert vandalism. If done via API, then OAuth login should be considered, either using the same OpenID login Wikimedia->Pundit (which is unlikely to be possible any time soon [3]), or by requesting the wikimedia oauth access. [4]

I was also considering this approach so for this to happen, I will customize Pundit's GUI specifically for Wikimedia by removing OpenID login system, and adding Wikimedia OAuth login.

Thanks

 

Amanpreet Singh,

IIT Roorkee

[Amanpreet Singh]

About the login I have some doubts,
This should be specifically clarified by Wikimedia people.

As the login functionality as mentioned at [1] needs some server side code, it would be difficult to integrate it with Pundit.
Other option as seems suitable to me is at [2], this requires JS but I don't think its viable an safe to send passwords in JS queries, what we need is something like Facebook JS SDK which provides us a unique APP ID and APP SECRET and through them we can connect to facebook to check if user is logged in or not, if not he is presented with a page to login.

Kindly correct me if the service like this exists in mediawiki or provide me a path to do this.

[1] https://www.mediawiki.org/wiki/OAuth/For_Developers

[2] http://www.mediawiki.org/wiki/Example_login_code_in_JS_(using_JQuery)

Thanks 

--

Amanpreet Singh,

IIT Roorkee

[Luca Martinelli]

Hi Amanpreet, I'm completely sorry I didn't answer you still, I'll try
to talk with some techie later and give you an answer! Promise! :)

L.

> --
> You received this message because you are subscribed to the Google Groups
> "Annotation tool GSoC" group.
> To unsubscribe from this group and stop receiving emails from it, send an
> email to annotation-tool-...@googlegroups.com.
> Visit this group at http://groups.google.com/group/annotation-tool-gsoc.
> For more options, visit https://groups.google.com/d/optout.



--
Luca "Sannita" Martinelli
http://it.wikipedia.org/wiki/Utente:Sannita