[openspending-dev] SpenDB Update

[Friedrich Lindenberg]

On a related note, I've spent another couple of evenings viciously stabbing
the SpenDB code base, with the following result:

* The Python code base is about as simple as it will ever be. It's a third
of the line count of what is deployed at openspending.org right now.

* It also doesn't have a lot of stuff, like FTS, a static HTML browser,
APIv1, APIv2, Badges, Scoreboards, etc.

* Web UI is coming together, you can now create a dataset and upload data
in one step. The major missing bits are a consistent nav for the dataset
manager, and a new model editor. Angular bits lack i18n and a proper
testing harness.

* The data model format is roughly the same as with cubepress, which should
have some nice implications.

Thinks I'm still unsure about:

* Unhappy with the way this does logging, which is basically to redirect
Python logging to S3. This should be more structured.

* Whether the new model editor can be based around source data columns,
rather than the actual model. Still looking for good examples.

* Can this be simple enough to run on a Heroku free tier?

My goal is to play with easy-to-use uploads to OffenerHaushalt by the end
of the month or so :)

Cheers,

- Friedrich
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.okfn.org/pipermail/openspending-dev/attachments/20150505/084fccf0/attachment.html>