Only log Clears when needed in face tracking example

[teh-cmc]

This is what caused the deadlocks/crashes we were after (which I'm glad it did -- those have been fixed in #4949).

Even though they are not a problem anymore, I'm not sure why they're here to begin with? Clearing scalars does nothing unless you specifically intend to use them with latest-at queries (i.e. not plots, which isn't the case here), otherwise it's just inefficient with no upside (which I'd rather not teach users). Am I missing something? 🤔

Checklist

• I have read and agree to Contributor Guide and the Code of Conduct
• I've included a screenshot or gif (if applicable)
• I have tested the web demo (if applicable):
  • Using newly built examples: app.rerun.io
  • Using examples from latest main build: app.rerun.io
  • Using full set of examples from nightly build: app.rerun.io
• The PR title and labels are set such as to maximize their usefulness for the next release's CHANGELOG

• PR Build Summary
• Docs preview
• Examples preview
• Recent benchmark results
• Wasm size tracking

[abey79]

I'm not convinced about this change. It would yield weird result when face(s) are no longer detected.

[abey79]

Those clear happen because the number of face detected can vary from 0 to N. That's why we clear faces/** and the re-populate faces/$i/**. Maybe a cleaner way to go about this would be to detect when faces disappear and clear only those ones.

[teh-cmc]

I don't understand the intent here.
Clearing effectively doesn't affect range queries at all, so what is the expected behavior we should expect? I don't get it.

I.e. this:

for step in range(0, 64):
    rr.set_time_sequence("step", step)
    rr.log("scalar", rr.TimeSeriesScalar(math.sin(step / 10.0)))

https://app.rerun.io/pr/4949/index.html?url=https://storage.googleapis.com/rerun-builds/pull_request/4950/scalar_no_clears.rrd

results in the same exact graph as this:

for step in range(0, 64):
    rr.set_time_sequence("step", step)
    rr.log("scalar", rr.TimeSeriesScalar(math.sin(step / 10.0)))
    rr.log("scalar", rr.Clear(recursive=True))

https://app.rerun.io/pr/4949/index.html?url=https://storage.googleapis.com/rerun-builds/pull_request/4950/scalar_clears.rrd

[abey79]

You're right, not clearing here as no effect, so it's ok to remove it (but I'd leave a TODO to explain why this one is not cleared while the others are).

Incidentally, it would be really nice if those Clears would actually translate into breaks in the timeseries to avoid those ugly jumps:

[teh-cmc]

Haaa interesting! Let me see if there's something quick we can do about this.

[teh-cmc]

I've updated the code to only clear faces when they disappear.

I might come up with a follow up PR soon so that the timeseries view interprets those clears as plot breakpoints.

[abey79]

Nice! Clear == break would actually be a very nice improvement tying nicely into the plotting story.