add cache to optimize type checking

[MintsInc]

What does this PR do?

Context

Jira ticket

The validate_and_convert_types function was causing severe performance degradation when check_return_type=True, resulting in 5.0x slowdown for API calls.

Root Cause: Exponential validation overhead due to deeply nested API responses. A single dashboard response triggered 108,516+ validation calls with recursive validation of the same nested structures repeatedly.

Failed Optimization Attempts

Early Exit Strategies: Primitive type checks, inheritance checks, and exact type matching were ineffective because inputs were raw JSON data requiring conversion to typed models, not already-typed instances.

The core issue was that we couldn't avoid the validation work through early exits, but we could avoid repeating identical validation work through caching.

Changes

ApiClient-Level Validation Cache

1. Cache Infrastructure: Added _validation_cache to ApiClient.__init__() with configurable size limit (default: 1000 entries)

2. Cache Integration: Modified ApiClient.deserialize() to pass cache to validate_and_convert_types()

3. Validation Function Updates: Added request_cache parameter throughout validation call chain with _make_hashable() helper for unhashable types

4. Gentle FIFO Memory Management: When cache is full, evict 25% of entries while keeping 75% most recent

Design Decisions:

• ApiClient-level scope: Cache persists across requests within same client instance
• Conservative defaults: 1000 entry limit balances memory vs. performance
• Zero breaking changes: Transparent caching, existing APIs unchanged

Tests

Performance Test Setup

Using the exact reproduction case from the original issue:

# main.py - Original test case from the ticket
def get_dashboard(check_return_type, count=10):
    configuration = Configuration(check_return_type=check_return_type)
    with ApiClient(configuration) as api_client:
        api_instance = DashboardsApi(api_client)
        for i in range(count):
            api_instance.get_dashboard("kct-7gx-vz7")  # Specific dashboard ID

# Executed with: dd-auth -- python main.py

Performance Results:

# Before optimization
check_return_type=True  - 15.0s
check_return_type=False - 3.0s   # 5.0x slowdown

# After optimization  
check_return_type=True  - 4.33s
check_return_type=False - 3.09s  # 1.4x slowdown

Additional Notes

Review checklist

Please check relevant items below:

• This PR includes all newly recorded cassettes for any modified tests.

• This PR does not rely on API client schema changes.

  • The CI should be fully passing.

• Or, this PR relies on API schema changes and this is a Draft PR to include tests for that new functionality.

  • Note: CI shouldn't be run on this Draft PR, as its expected to fail without the corresponding schema changes.

[MintsInc]

• add cache to optimize type checking #2737 👈 (View in Graphite)
• master

This stack of pull requests is managed by Graphite. Learn more about stacking.