jonathanjfshaw

Here's an example of intended use: ``` nodes: - id: data_source ... - id: filter class: timeflux.nodes.expression class: ExpressionFilter params: expression: i_data < 6 eval_on: ports edges: - source: data_source...

Because this is extremely simple, and allows for the deprecation of the core Gate node, I suggest it belongs in core not timeflux_dsp.

> Things got mixed up in the commit. test_states.py (which should not be there in the first place) has the contents of test_gate.py. There is no test_missing.py. Done. > dropna...

I am also having doubts about this. I'm wondering about creating a node called 'Method' that has parameters 'method' and 'args' which simply does getattr(self.i.data, method, args). Then we wouldn't...

Perhaps the solution is to create the dataframe, measure its length, then generate and add the index. This should work robustly with any valid source of data for a dataframe.

Hi Pierre, thanks for your detailed answer. I had found the problem easy to work round as you suggested. I think the 'set' name led me into into thinking this...

The states tests are throwing quite a lot of warnings, I should fix them

Looks like .gitpod.Dockerfile is not longer needed, the readme needs changing.

Is this needed given #322 which seems much cleaner? I really don't like hardcoding entity-type specific logic into a generic function like parseEntityFields, it's building a maintenance nightmare.

http://cgit.drupalcode.org/lightning/tree/tests/features/bootstrap/entity.behat.inc?h=8.x-1.x Looks like this functionality has been created for the Drupal Lightning distribution. Hopefully it's ripe for picking out and including here.