Make realtime indexes pluggable

[richardstartin]

This is PR simply moves static realtime index provisioning logic out of MutableSegmentImpl's constructor and into a default implementation of an SPI which can overridden. This PR follows an identical approach to making offline indexes pluggable, the rationale and approach were described in #7895.

The changes depend on a of package move (#8278) which I would like to get merged in before rebasing this PR to reduce the surface area of this one.

[codecov-commenter]

Codecov Report

❌ Patch coverage is 0% with 137 lines in your changes missing coverage. Please review.
✅ Project coverage is 14.14%. Comparing base (e498a71) to head (b8c860b).
⚠️ Report is 6316 commits behind head on master.

Files with missing lines	Patch %	Lines
...pi/index/mutable/provider/MutableIndexContext.java	0.00%	62 Missing ⚠️
...exsegment/mutable/DefaultMutableIndexProvider.java	0.00%	35 Missing ⚠️
...local/indexsegment/mutable/MutableSegmentImpl.java	0.00%	17 Missing ⚠️
...che/pinot/segment/spi/index/IndexingOverrides.java	0.00%	17 Missing ⚠️
...dindex/RealtimeLuceneIndexReaderRefreshThread.java	0.00%	2 Missing ⚠️
...me/impl/invertedindex/RealtimeLuceneTextIndex.java	0.00%	2 Missing ⚠️
...time/impl/invertedindex/RealtimeInvertedIndex.java	0.00%	1 Missing ⚠️
...local/realtime/impl/json/MutableJsonIndexImpl.java	0.00%	1 Missing ⚠️

❗ There is a different number of reports uploaded between BASE (e498a71) and HEAD (b8c860b). Click for more details.

HEAD has 3 uploads less than BASE

Flag	BASE (e498a71)	HEAD (b8c860b)
unittests1	1	0
integration1	1	0
integration2	1	0

Additional details and impacted files

@@              Coverage Diff              @@
##             master    #8279       +/-   ##
=============================================
- Coverage     70.82%   14.14%   -56.69%     
+ Complexity     4239       81     -4158     
=============================================
  Files          1631     1588       -43     
  Lines         85461    83678     -1783     
  Branches      12877    12677      -200     
=============================================
- Hits          60527    11833    -48694     
- Misses        20749    70961    +50212     
+ Partials       4185      884     -3301     

Flag	Coverage Δ
integration1	?
integration2	?
unittests1	?
unittests2	14.14% <0.00%> (-0.03%)	⬇️

Flags with carried forward coverage won't be shown. Click here to find out more.

☔ View full report in Codecov by Sentry.
📢 Have feedback on the report? Share it here.

🚀 New features to boost your workflow:

• 📦 JS Bundle Analysis: Save yourself from yourself by tracking and limiting bundle sizes in JS merges.

[Jackie-Jiang]

LGTM otherwise

[sajjad-moradi]

Looks good. I left a few minor comments.

[kishoreg]

@sajjad-moradi We rarely block a PR unless the code design is completely broken. I looked at the comments and I see that you have mentioned that most of them are subjective.

Can you please unblock us if what you are requesting is subjective. Thanks for understanding.

[sajjad-moradi]

@sajjad-moradi We rarely block a PR unless the code design is completely broken. I looked at the comments and I see that you have mentioned that most of them are subjective.

Can you please unblock us if what you are requesting is subjective. Thanks for understanding.

Of course. Didn't know about this convention. The PR really looks good in general.

[richardstartin]

@sajjad-moradi We rarely block a PR unless the code design is completely broken. I looked at the comments and I see that you have mentioned that most of them are subjective.
Can you please unblock us if what you are requesting is subjective. Thanks for understanding.

Of course. Didn't know about this convention. The PR really looks good in general.

Thanks, I felt it was important to run the changes past you or @mcvsubbu as a courtesy, because you might need to debug something introduced here, but when it comes to naming (or composition vs inheritance) there tend to be as many opinions as there are people.