Validate mocks - first step

[archmoj]

In addition to heavy cleanup and updating (by applying some non-deprecated attributes e.g. title object), this PR adds jasmine tests for validating all current mocks using Plotly.validate.
The test called mock_test is drafted in commit 8c1a9d9 which currently bypasses undefined layout, invisible traces, coloraxis, dynamic e.g. textposition arrays.
With this relaxation,
before modifying any mocks running the test resulted in
TOTAL: 1014 | FAILED: 487 | SUCCESS: 527
After updating the mocks the situation improved as listed below:
TOTAL: 1014 | FAILED: 174 | SUCCESS: 840.

@alexcjohnson
cc: @plotly/plotly_js

cc: #4733

[nicolaskruchten]

@alexcjohnson let's review this once shape-drawing is done :)

[alexcjohnson]

@archmoj I just had a quick look at this - very nice! Let's not touch it until 1.54.0 is released, but as this doesn't impact features at all we can include it in a patch after that.

My primary question is about the hard-coded list in mocks_test - seems like we should be able to

• read in the list of all mocks from the filesystem - perhaps if this step is fast enough we can just have karma.conf.js do it before every test run and write the results to a file in tests/jasmine/assets?
• only list a set of mocks or patterns to ignore - and then over time we can reduce that to an empty list.

[archmoj]

My primary question is about the hard-coded list in mocks_test - seems like we should be able to

• read in the list of all mocks from the filesystem - perhaps if this step is fast enough we can just have karma.conf.js do it before every test run and write the results to a file in tests/jasmine/assets?
• only list a set of mocks or patterns to ignore - and then over time we can reduce that to an empty list.

@alexcjohnson
That is a very good idea we could try in another PR.
At first I did try to adapt our image test for this purpose.
But it was rather complicated to acess the system files from the jasmine test.
I think the way the required mocks are listed is OK atm.
It is actually handy for fixing those many mocks.

[alexcjohnson]

That is a very good idea we could try in another PR.

🍻 fair enough - added to #4733 so we address it before considering that issue closed.

This looks great. You currently have the require statements for the mocks all commented out - can we enable the ones you've fixed already, just comment out the still-failing ones, and call this PR done?

[archmoj]

This looks great. You currently have the require statements for the mocks all commented out - can we enable the ones you've fixed already, just comment out the still-failing ones, and call this PR done?

Done in efd8518.

[alexcjohnson]

💃