Fix duplicate check for CREATE MODEL and CREATE PROMPT

[anasdorbani]

CREATE MODEL and CREATE PROMPT commands were not properly detecting duplicates when a model/prompt already existed in a different table (global vs local). This allowed duplicate creations to succeed, showing "created successfully" for both attempts.

[Copilot]

Pull request overview

The PR adds audio transcription support to the Flock extension, along with comprehensive test coverage and a new metrics tracking system. It enables CREATE MODEL and CREATE PROMPT commands to properly detect duplicates across global and local scopes.

Key Changes:

• Added audio transcription functionality for OpenAI models with gpt-4o-mini-transcribe
• Implemented a metrics tracking system with flock_get_metrics(), flock_get_debug_metrics(), and flock_reset_metrics() functions
• Enhanced test coverage across scalar and aggregate LLM functions
• Modified model providers to support transcription requests and improved image handling

Reviewed changes

Copilot reviewed 99 out of 102 changed files in this pull request and generated no comments.

Show a summary per file

File	Description
test/unit/prompt_manager/prompt_manager_test.cpp	Added tests for audio transcription column handling
test/unit/model_manager/model_providers_test.cpp	Updated model name and added transcription tests
test/unit/model_manager/model_manager_test.cpp	Updated model name reference
test/unit/functions/scalar/metrics_test.cpp	New comprehensive metrics tracking tests
test/unit/functions/scalar/llm_function_test_base_instantiations.cpp	Added Ollama secret creation
test/unit/functions/scalar/llm_filter.cpp	Added audio transcription tests and updated type casts
test/unit/functions/scalar/llm_embedding.cpp	Updated type casts from TEXT to VARCHAR
test/unit/functions/scalar/llm_complete.cpp	Added audio transcription tests and updated type casts
test/unit/functions/mock_provider.hpp	Added mock methods for transcription
test/unit/functions/aggregate/llm_*.cpp	Added audio transcription tests and improved test structure
test/unit/functions/aggregate/llm_aggregate_function_test_base.hpp	Updated mock provider setup with factory pattern
test/integration/src/integration/tests/**/*.py	Added with_secrets parameter to all CLI calls
test/integration/src/integration/setup_test_db.py	Simplified setup by removing secret creation logic
test/integration/src/integration/conftest.py	Added secret injection and audio file path helper
src/registry/scalar.cpp	Registered new metrics functions
src/prompt_manager/prompt_manager.cpp	Added audio transcription support and null handling
src/model_manager/providers/adapters/*.cpp	Updated to support transcription and improved media handling
src/model_manager/model.cpp	Enhanced model details resolution and added transcription methods
src/metrics/*.cpp	New metrics tracking implementation
src/include/flock/**/*.hpp	Updated headers for new functionality

Comments suppressed due to low confidence (3)

test/unit/model_manager/model_providers_test.cpp:1

• The model version 'gemma3:4b' does not exist. The Gemma model family uses version numbers like 'gemma:2b', 'gemma:7b', or 'gemma2:2b'. This should be changed to an existing Gemma model version.
  test/unit/model_manager/model_manager_test.cpp:1
• The model version 'gemma3:4b' does not exist. The Gemma model family uses version numbers like 'gemma:2b', 'gemma:7b', or 'gemma2:2b'. This should be changed to an existing Gemma model version.
  test/integration/src/integration/tests/secret_manager/test_secret_manager.py:1
• The secret name changed from 'test_openai_secret' to 'test_persistent_lifecycle_secret', but this creates potential naming conflicts with other tests. Consider using a more unique name like 'test_persistent_secret_lifecycle' to better indicate its purpose and avoid conflicts.

---

💡 Add Copilot custom instructions for smarter, more guided reviews. Learn how to get started.