Test text response structure

Test Model initialization with valid model configuration

Test TranscribeAudioColumn with named column

Test 1-tuple case: no LLM call needed, returns the single tuple directly

Test llm_reduce_json with SQL queries without GROUP BY - new API

Test 1-tuple case: no LLM call needed, returns the single tuple directly

Test single tuple: LLM is still called for reduce (to summarize)

Test 1-tuple case: no LLM call needed, returns the single tuple directly

Test llm_embedding with SQL queries - new API

Test llm_filter with SQL queries - new API

Test llm_complete with SQL queries

Test llm_complete with JSON responses using new SQL API

Common test methods that can be used by derived classes if needed

Convert TokenType to string

Constructor

Add input and output tokens (accumulative)

Record token usage (accumulative)

Validate file exists and is not empty

Performance test with larger dataset (commented out for regular testing).

Performance test - commented out with underscore prefix for optional execution

Performance test with larger dataset - commented out with underscore prefix for optional execution.

Performance test with larger dataset - commented out with underscore prefix for optional execution.

Performance test with larger dataset - commented out with underscore prefix for optional execution.

Performance test with larger dataset - commented out with underscore prefix for optional execution.