[UPDATE] export parquet, include score tables into export if present

[singjc]

This pull request introduces a new option to control whether transition-level data is included in the exported parquet files, and significantly improves the flexibility and completeness of score column handling in the export process. It also enhances the robustness of SQL query generation and table creation for various export scenarios.

The most important changes are:

New feature: Control inclusion of transition data in exports

• Added an include_transition_data option to both the configuration (ExportIOConfig) and the CLI (export_parquet), allowing users to choose whether to include transition-level data in the exported parquet files. The export logic now respects this option and skips transition data export if disabled. [1] [2] [3] [4] [5] [6] [7] [8]

Score column and SQL query improvements

• Added dynamic detection and inclusion of transition-level score columns in export queries, using the new _build_transition_score_columns_and_join helper. Transition score columns are now included only if the relevant table exists, and appropriate NULL columns are added to maintain table consistency between precursor and transition exports. [1] [2] [3] [4] [5] [6] [7] [8]

• Enhanced temporary table creation to dynamically include all relevant score columns (MS1, MS2, IPF, peptide, protein, and transition scores) with appropriate types, ensuring the exported schema matches the data. [1] [2]

Robustness and correctness fixes

• Improved filtering of columns when preparing export column info to ensure only columns with valid types are included, preventing potential SQL errors.

• Fixed SQL join logic in peptide/protein score table generation to include RUN_ID in the join condition, ensuring correct mapping of global and non-global scores. [1] [2]