Add scoping support to gathering events

[fsimonis]

Main changes of this PR

This PR changes gatherstats to use scopes for extracting timings from the profiling data, making it robust against the introduction of additional scopes in between.

Related to precice/precice#2066

Author's checklist

• I used the pre-commit hook and used pre-commit run --all to apply all available hooks.
• I added a test to cover the proposed changes in our test suite.
• I updated the documentation in docs/README.md.
• I added a changelog entry in ./changelog-entries/ (create if necessary).
• I updated potential breaking changes in the tutorial precice/tutorials/aste-turbine.

[davidscn]

Ah that's nice. Did you check the result? Since you are already taking a look here: we have the script as part of our CI. Could we check as part of our CI that the gathered stats are 'correct' or within bounds? This would make testing way easier, as the script might not complain/proceed, when a timing is 'not present'. In the current CI, this would be a pass, but insufficient from a functionality POV.

[fsimonis]

Ah that's nice. Did you check the result?

Results are now identical to the before.
I checked with preCICE precice/precice#2066

Since you are already taking a look here: we have the script as part of our CI. Could we check as part of our CI that the gathered stats are 'correct' or within bounds? This would make testing way easier, as the script might not complain/proceed, when a timing is 'not present'. In the current CI, this would be a pass, but insufficient from a functionality POV.

A CI python script that checks if the timings contain possible values (as in >0) would be a good first step. But checking runtime values is always a problem without controlled hardware.

[davidscn]

It would be sufficient to test that we have a "reasonable" value. Examples: a lately introduced 'sync' event gave me some zeros. The renaming of events gives nothing. The event scoping runs into the risk of a similar 'nothing' value (putting aside that you tested it for now).

[fsimonis]

Right. Opening an issue regarding this

[davidscn]

The issue is insufficient to ensure functionality for upcoming changes though