remove DataTypeInfo and use LogicalType and column names

[russell-liu]

fixes #2423

[codecov]

Codecov Report

Attention: Patch coverage is 81.69935% with 28 lines in your changes are missing coverage. Please review.

Project coverage is 93.31%. Comparing base (d946982) to head (84700b3).
Report is 11 commits behind head on master.

❗ Current head 84700b3 differs from pull request most recent head 4d9b7bf. Consider uploading reports for the commit 4d9b7bf to get more accurate results

Files	Patch %	Lines
src/common/arrow/arrow_row_batch.cpp	78.46%	28 Missing ⚠️

Additional details and impacted files

@@            Coverage Diff             @@
##           master    #2539      +/-   ##
==========================================
+ Coverage   92.28%   93.31%   +1.03%     
==========================================
  Files        1161     1036     -125     
  Lines       44150    38819    -5331     
==========================================
- Hits        40744    36225    -4519     
+ Misses       3406     2594     -812     

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

[andyfengHKU]

Add test to cover all uncovered lines.

[andyfengHKU]

use common::logical_types_t

[russell-liu]

After attempting to add more Arrow tests, I've discovered a problem with this removal of DataTypeInfo that I do not know how to resolve. Details:

I tried to return a node, rel, or internal ID in the Arrow tests, and this led to segmentation faults. After further investigation, I believe this is because the following logic in the old implementation using DataTypeInfo is not preserved:
https://github.com/kuzudb/kuzu/blob/master/src/main/query_result.cpp#L24
https://github.com/kuzudb/kuzu/blob/master/src/main/query_result.cpp#L102
https://github.com/kuzudb/kuzu/blob/master/src/main/query_result.cpp#L115
That is, when DataTypeInfo used to be created for internal IDs, nodes, or rels, some additional childrenTypesInfo would be created; when we switch to using DataType only, these fields are missing.

I tried to modify arrow_row_batch.cpp (including templateInitializeVector and templateCopyNonNullValue for internal IDs, nodes, and rels) to replicate this logic, but I didn't really know what I was doing and I ended up with a segmentation fault occurring during std::memcpy that I didn't know what to do about. My approach was probably incorrect and messy I won't push the code here.

[mxwli]

Should close #3224 as well

[andyfengHKU]

Make sure we add corresponding test cases for coverage.