Change hash index concurrent build

[ray6080]

Our hash index builder is not scalable. We need to rework on that.
Currently, I'm putting locks each time a thread flushes into the index. This can improve copying. On my local M1 Air laptop, it improves copying of ldbc10 comment csv files from ~28s to ~12s, and parquet files from ~22.4s to 5.6s.

[codecov]

Codecov Report

Patch coverage: 100.00% and project coverage change: -0.01 ⚠️

Comparison is base (91204de) 91.73% compared to head (43e0393) 91.72%.

❗ Current head 43e0393 differs from pull request most recent head faa5756. Consider uploading reports for the commit faa5756 to get more accurate results

Additional details and impacted files

@@            Coverage Diff             @@
##           master    #1564      +/-   ##
==========================================
- Coverage   91.73%   91.72%   -0.01%     
==========================================
  Files         698      698              
  Lines       25128    25113      -15     
==========================================
- Hits        23050    23034      -16     
- Misses       2078     2079       +1     

Impacted Files	Coverage Δ
src/storage/copier/node_copy_executor.cpp	91.30% <ø> (ø)
src/storage/index/hash_index_builder.cpp	89.58% <ø> (-0.99%)	⬇️
src/binder/bound_statement_result.cpp	92.30% <100.00%> (ø)
src/include/storage/index/hash_index_builder.h	79.16% <100.00%> (-2.66%)	⬇️
src/include/storage/storage_info.h	100.00% <100.00%> (ø)
src/storage/copier/node_copier.cpp	91.55% <100.00%> (+0.11%)	⬆️

... and 1 file with indirect coverage changes

☔ View full report in Codecov by Sentry.
📢 Do you have feedback about the report comment? Let us know in this issue.