[omaha] Guest Speaker-Django

[Wes Turner]

On Thu, Jun 18, 2015 at 2:59 PM, Burch Kealey via Omaha <omaha at python.org>
wrote:

> Hi Everyone;
>
>
> We have been building a pretty cool backend using Django.  Here is one
> example:
>
>
> [cid:f2154b94-e8ae-4f14-874f-840de88676f7]
>
>
>
> Every night we pull SEC filings, parse out tables and try to add some
> value to specific tables.  In this case we pull executive and director
> compensation and try to identify the gender of the people mentioned in the
> table so we can add their gender with the other table data.  We use various
> cues in the document to try to discern their gender.  However, if the
> results fail some tests then we don't assign a value and it moves to a
> queue for a human to make the determination.  So the people whose gender
> the system could not automatically discern are moved to our Django system
> and it is set up so the analyst can look at the data and make the
> determination.
>

Wikipedia: https://en.wikipedia.org/wiki/XBRL

XBRL as RDF (e.g. for SPARQL queries)

* http://rhizomik.net/semanticxbrl/html/
* http://datahub.io/dataset/semantic-xbrl
*
https://github.com/openlink/Virtuoso-RDFIzer-Mapper-Scripts/blob/master/ontologies/xbrl/xbrl.owl

Django and SPARQL (RDF queries)

* https://code.google.com/p/djubby/
* https://github.com/wikier/djubby

Django and RDFLib

* https://wrdrd.com/docs/consulting/knowledge-engineering#rdflib
* https://github.com/odeoncg/django-rdflib


>
>
> Now the cool thing is that this is all web based so the analyst can be
> anywhere.  Further, once they make the determination they press another
> button (not visible in this screen shot) and the results get moved to a
> repository where they are always available for future processing runs.
>

That sounds pretty cool. Integration/linking with
https://github.com/quantopian/zipline would be cool.


>
>
> We have a lot of other stuff that is based on Django.
>
>
> The person who did all of the heavy lifting for us is with a company in
> India (PyDan) and is pretty active in user groups there.
>
>
> So , Siva would probably be delighted to speak remotely about Django -
> scheduling might be a little tricky but I think it is workable with a
> little bit of lead time.
>
>
> Before I ask him, you of course should opine as to whether or not you
> would like to have a guest speaker from India (remotely).
>

+1; though I may not be able to make it.


>
>
> We are embedding Python into a C# application.  We are not using IPython
> because we need LXML which is not available in IPython (as far as I could
> tell).  If I make enough progress before next month I might be able to
> share some of what we learned after the main presentation.
>

conda install ipython-notebook lxml html5lib beautifulsoup && pip install
django rdflib

https://github.com/quantopian/qgrid
http://pandas.pydata.org/pandas-docs/stable/ecosystem.html#quantopian-qgrid