Long term idea: add Wrattler support

[sebastian]

Wrattler is a notebook environment (much like Jupyter notebooks) that is somewhat unique in that it supports multiple languages that can share data. This allows for a workflow where you can load data in one language, process it in another, and render graphs in a third. The idea sounds very neat.

In traditional notebook environments we have the issue that if we were to make an Aircloak specific SQL-querying kernel, it would be very hard to make the data available for further processing. Wrattler solves this problem.

The idea would be to have a kernel/aircloak language support which would allow someone to query the system directly using SQL and nothing but SQL. The results would then be packed into whatever the internal data format is, and then directly available for further analysis. It would not be a major improvement over say issuing queries using psycopg2 in Python, but still a meaningful one. Since the environment in which the query is added/edited is Aircloak specific we could:

• have Aircloak SQL auto-completion (which we can't in a Python kernel – where we can't even do general SQL auto-completion)
• data source specific auto-completion
• whatever other fancy things we might support in our web interface (like for example showing which part of a query is anonymizing, which is not etc).

From a brief look at the docs it seems we are entirely free to impact the UX the plugin presents, so we could for example have something which provides a text box that allows the user to specify two things: the name of the data frame the results should be stored in, the SQL for the query with a proper SQL editor.

---

Not actionable at the moment. Just food for thought.

[sebastian]

One concern is that this might be very much an "academic system", meaningful it might not at all be production ready and might not be maintained beyond the time frame they want it to publish papers...