perf(query): optimize field processing with caching

[fyvri]

This PR introduces a caching mechanism to optimize the processing of struct fields in the query package. The main goal is to reduce redundant reflection operations when encoding structs into URL query parameters, especially for nested or deeply embedded structs.

Key Changes:

1. Added Caching:

   • Introduced a structCache map to store field information for each struct type, avoiding repeated reflection operations.
   • Field information is computed once and reused for subsequent encoding of the same struct type.

2. Improved Performance:

   • Reduced the overhead of reflection by caching field metadata, leading to faster encoding of structs.
   • This is particularly beneficial for applications that frequently encode large or nested structs.

3. Backward Compatibility:

   • The external API remains unchanged, ensuring no breaking changes for existing users.

Impact:

• Faster Encoding: The caching mechanism significantly improves performance by reducing redundant reflection calls.
• Low Memory Overhead: The cache introduces minimal memory usage, with a clear performance benefit.