Split threads by default in memray transform csv

[godlygeek]

Closes: #996

memray transform always retrieved allocation records with threads merged, so allocations made on different threads at the same call stack collapsed into a single CSV row and the tid / thread_name columns no longer identified a single thread.

The point of the csv format is to be machine-parseable, so rather than adding an opt-in flag, make memray transform csv retrieve records with merge_threads=False by default. Each thread's allocations now stay on their own row carrying that thread's real tid and thread_name, which is also what the csv column documentation has always described. The gprof2dot and speedscope formats have no per-thread dimension and are unchanged.

[godlygeek]

@BhargavKumarNath this is your work from #1006, lightly edited. I simplified the tests quite a lot, dropping some things that I think were already adequately exercised by other tests, and switching to building a capture file "by hand" rather than running a test program to generate one to simplify the remaining test. All of the implementation and docs match yours, except that I changed the NotImplementedError to be raised whenever merge_threads == (self.format == "csv") - so we raise a NotImplementedError if merge_threads is true and the format is CSV, or when merge_threads is false and the format isn't CSV.

Sorry for not getting around to reviewing #1006 before you deleted your fork 😅 but I'm leaving the commit attribution as you since you did basically all of the work here, and I just tweaked the tests and one exception check / message.

[codecov-commenter]

Codecov Report

❌ Patch coverage is 91.66667% with 2 lines in your changes missing coverage. Please review.
✅ Project coverage is 92.31%. Comparing base (c3be2a3) to head (1ead335).

Files with missing lines	Patch %	Lines
src/memray/commands/transform.py	0.00%	1 Missing ⚠️
src/memray/reporters/transform.py	50.00%	1 Missing ⚠️

Additional details and impacted files

@@            Coverage Diff             @@
##             main    #1056      +/-   ##
==========================================
- Coverage   92.35%   92.31%   -0.04%     
==========================================
  Files         103      103              
  Lines       13363    13385      +22     
  Branches      483      484       +1     
==========================================
+ Hits        12341    12357      +16     
- Misses       1022     1028       +6     

Flag	Coverage Δ
cpp	92.31% <91.66%> (-0.04%)	⬇️
python_and_cython	92.31% <91.66%> (-0.04%)	⬇️

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

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

🚀 New features to boost your workflow:

• ❄️ Test Analytics: Detect flaky tests, report on failures, and find test suite problems.
• 📦 JS Bundle Analysis: Save yourself from yourself by tracking and limiting bundle sizes in JS merges.

[BhargavKumarNath]

Thanks for picking this up, @godlygeek, and no worries about #1006. I'm glad it's getting in. The simplified tests and the merge_threads == (self.format == "csv") check both look good to me. That condition covers both invalid combinations in one place, which is cleaner than what I had. Let me know if you'd like anything else changed on my side.