improvement: proper data types mapping

[github-actions]

Problem

Currently, Olake maps all integer and floating-point types to generalized types like Int64 or Float64, without considering more specific types such as Int32, UInt32, Float32, etc.

This results in:

• Loss of type precision
• Potential compatibility issues for downstream systems
• Increased memory usage for smaller types

Goals

• Implement accurate and granular type mappings based on the source column types
• Support all major numeric variants:
  • Signed and unsigned integers: tinyint, smallint, int, bigint, etc.
  • Floating-point numbers: float, real, double precision
• Ensure consistent type preservation across different sources (e.g., MySQL, PostgreSQL)

Tasks

• Audit current type mapping logic
• Add proper mappings for Int32, UInt32, Int16, UInt16, Float32, etc.
• Ensure compatibility across all supported source systems
• Document all supported type mappings in the developer guide

Outcome

• Enhanced data accuracy and fidelity
• Better downstream integration and usability

Example: below types not properly mapped

---

olake/drivers/postgres/internal/datatype_conversion.go

Lines 8 to 13 in ed705d2

	// TODO: add proper types (not only int64)
	"bigint": types.Int64,
	"tinyint": types.Int64,
	"integer": types.Int64,
	"smallint": types.Int64,
	"smallserial": types.Int64,

---

This issue was generated by todo-issue based on a TODO comment in ed705d2. It's been assigned to @hash-data because they committed the code.