[QDP] Add unit tests for Parquet readers

[0lai0]

Related Issues

Closes #1184

Changes

• Bug fix
• New feature
• Refactoring
• Documentation
• Test
• CI/CD pipeline
• Other

Why

qdp/qdp-core/src/readers/parquet.rs had less code coverage. This means there was no regression protection for sample size validation, state initialization, or the streaming encoding path.

How

Added 10 unit tests in qdp/qdp-core/src/readers/parquet.rs :

1. test_parquet_reader_missing_file — validates missing file returns MahoutError::Io
2. test_parquet_reader_bad_schema — rejects non-list schema (Int32 column)
3. test_parquet_reader_too_many_columns — rejects files with more than one column
4. test_parquet_reader_bad_list_type — rejects List (must be List)
5. test_parquet_reader_list_f64 — roundtrip List with value and metadata assertions
6. test_parquet_reader_fixed_size_list_f64 — roundtrip FixedSizeList with value and metadata assertions
7. test_parquet_reader_inconsistent_sample_sizes — verifies error when list lengths differ across rows
8. test_parquet_streaming_reader_scalar_f64 — validates scalar Float64 basis encoding path with chunked read_chunk() calls
9. test_parquet_streaming_reader_list_f64_chunked — streaming read_batch() over List with small batch size
10. test_parquet_streaming_reader_empty_data — explicit assertion on empty Parquet file error

Checklist

• Added or updated unit tests for all changes
• Added or updated documentation for all changes