Handle STR0NNN in a generic way

[tcely]

The function should return 'VP9' for:

• vp9.1.2.3
• vp09.4.5.6

As well as other variations.
It also works for avc01 or av01, or any others that fit the pattern.

[tcely]

I've been running with this code for a while now. I'm seeing more exact matches than before.

[meeb]

This looks fine, but preferably could do with some tests for the new function.

Edit: and thanks for the further contributions!

[tcely]

The matching tests are already exercising that function.
I checked the test data. I believe the coverage of the function is even complete.

Was there anything specific you were wanting to test?

[meeb]

Ah, I just meant as you're (slightly) modifying how the codec string was being parsed if there was a test that explicitly confirmed this behavior. Current tests are probably sufficient.

[tcely]

I'm up for adding more tests, especially as the boring metadata doesn't have any vcodec matching this vp09.01.2.40 pattern, that I saw in my database.

The 20230629 metadata does have both vp09 and av01 codecs though.

[meeb]

Yeah YouTube especially seem to tweak the codecs quite frequently. The version number postfixes are a new-ish change.