[SYSTEMDS-3932] CSV reader for out-of-core streams

[janniklinde]

This patch adds an out-of-core CSV reblock instruction. It supports reading single or multiple row partitioned CSV files into dense matrix blocks. Reads are currently performed by a single thread and thus performance is comparable to (slightly slower than) non-parallel dense CSV reads if all blocks per row can be held in cache. The number of maximum blen x blen matrix blocks that are constructed in memory simultaneously can be specified by MAX_BLOCKS_IN_CACHE.

[codecov]

Codecov Report

❌ Patch coverage is 80.06135% with 65 lines in your changes missing coverage. Please review.
✅ Project coverage is 72.32%. Comparing base (c7300f3) to head (c0e87a4).

Files with missing lines	Patch %	Lines
...ime/instructions/ooc/CSVReblockOOCInstruction.java	80.18%	39 Missing and 25 partials ⚠️
...rc/main/java/org/apache/sysds/lops/CSVReBlock.java	50.00%	0 Missing and 1 partial ⚠️

Additional details and impacted files

@@             Coverage Diff              @@
##               main    #2352      +/-   ##
============================================
+ Coverage     72.29%   72.32%   +0.02%     
- Complexity    46829    46880      +51     
============================================
  Files          1508     1509       +1     
  Lines        177638   177962     +324     
  Branches      34880    34938      +58     
============================================
+ Hits         128430   128708     +278     
- Misses        39511    39535      +24     
- Partials       9697     9719      +22     

☔ View full report in Codecov by Sentry.
📢 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.

[mboehm7]

LGTM - thanks for the patch and the additional simplification @janniklinde. For now we focus on lean code and optimizing the common case in terms of data characteristics.