Update completer part 3

[trungleduc]

References

This PR allows completer providers to customize the completer widget model and fixes an issue of activating completer before all providers are registered.

Code changes

• Add a new signal (providersActivated), that is emitted when all providers are registered.
• Add an optional modelFactory method to the provider interface.

User-facing changes

N/A

Backwards-incompatible changes

N/A

[jupyterlab-probot]

Thanks for making a pull request to jupyterlab!
To try out this branch on binder, follow this link:

[trungleduc]

Restart CI

[fcollonval]

Design question, are you interested by listening to the changes in the active providers or in the readiness of the manager at load time?

If this is the second, I would use a ready attribute:

get ready(): Promise<void>

If not, the signal is good, I would use anoterh name like activeProvidersChanged to better reflect when the signal is emitted (as it will be emitted on settings update too).

Look for example at jupyterlab/jupyterlab/packages/apputils/src/sessioncontext.tsx

[trungleduc]

Indeed I need to call the slot on the provider changed event since I need to update existing completer handlers with the new provider setting.
I'll rename the signal to reflect its intent.

[fcollonval]

I got a question about the intent.

[fcollonval]

Minor suggestions on the doc and it is good to go for me.

[trungleduc]

Thanks @fcollonval for the review and suggestion!

[fcollonval]

Thanks @trungleduc

[github-actions]

Benchmark report

The execution time (in milliseconds) are grouped by test file, test type and browser.
For each case, the following values are computed: min <- [1st quartile - median - 3rd quartile] -> max.

The mean relative comparison is computed with 95% confidence.

Results table

Test file	large_code_notebook	large_md_notebook
open
chromium
actual	3453 <- [3762 - 3875 - 3963] -> 4398	2443 <- [2743 - 2793 - 2865] -> 3120
expected	3412 <- [3899 - 3960 - 4071] -> 5041	2390 <- [2681 - 2728 - 2800] -> 3060
Mean relative change	-3.1% ± 1.4%	2.3% ± 1.2%
switch-from
chromium
actual	620 <- [674 - 695 - 753] -> 938	479 <- [523 - 544 - 935] -> 1131
expected	624 <- [704 - 731 - 760] -> 948	486 <- [531 - 890 - 941] -> 1206
Mean relative change	-2.5% ± 2.9%	-8.9% ± 7.5%
switch-to
chromium
actual	303 <- [372 - 379 - 390] -> 428	246 <- [280 - 288 - 296] -> 682
expected	309 <- [381 - 390 - 399] -> 426	229 <- [276 - 282 - 291] -> 633
Mean relative change	-2.3% ± 1.5%	1.9% ± 3.9%
close
chromium
actual	648 <- [901 - 921 - 937] -> 1018	450 <- [474 - 480 - 492] -> 527
expected	576 <- [912 - 936 - 956] -> 1004	437 <- [470 - 477 - 490] -> 833
Mean relative change	-1.3% ± 1.7%	0.0% ± 1.7%

Changes are computed with expected as reference.