Implement conversion tables for text and binary protocol parsing

Implement conversion tables for text and binary protocol parsing

Building on the new ColumnsDefinitionPacket (object-of-arrays) structure, implement precomputed conversion tables for both text and binary protocol row parsing.

Instead of performing type checks (if/elif) while decoding each row, conversion functions should be resolved once during metadata processing and stored in per-column lookup tables.

Scope

• Add conversion table initialization to ColumnsDefinitionPacket
• Support text protocol and binary protocol separately
• Conversion tables should be indexed by column number
• Row parsing should only:
  • extract raw field data
  • call the preassigned conversion function

Goals

• Eliminate per-row type branching
• Reduce attribute lookups while parsing
• Share conversion logic between text and binary protocols
• Centralize type handling in column metadata

Benefits

• Improved parsing performance
• Better cache locality
• Cleaner separation between:
  • metadata preparation
  • row decoding
• Easier future C-level optimization

Dependencies

Requires implementation of ColumnsDefinitionPacket (CONPY-335)