torbjorn/my-pal-mcp-server
1
0
Fork 0
Code Issues Pull Requests Actions Packages Projects Releases Wiki Activity

feat: added intelligence_score to the model capabilities schema; a 1-20 number that can be specified to influence the sort order of models presented to the CLI in auto selection mode

Browse Source 
fix: model definition re-introduced into the schema but intelligently and only a summary is generated per tool. Required to ensure CLI calls and uses the correct model
fix: removed `model` param from some tools where this wasn't needed
fix: fixed adherence to `*_ALLOWED_MODELS` by advertising only the allowed models to the CLI
fix: removed duplicates across providers when passing canonical names back to the CLI; the first enabled provider wins
This commit is contained in:
Fahad
2025-10-02 21:43:44 +04:00
parent e78fe35a1b
commit 6cab9e56fc
22 changed files with 525 additions and 110 deletions
Download Patch File Download Diff File Expand all files Collapse all files

4
tests/test_consensus.py
View File

@@ -122,8 +122,8 @@ class TestConsensusTool:
# relevant_files should be present as it's used by consensus
assert "relevant_files" in schema["properties"]
# model field should be present for Gemini compatibility (consensus uses 'models' as well)
assert "model" in schema["properties"]
# model field should NOT be present as consensus uses 'models' field instead
assert "model" not in schema["properties"]
# Verify workflow fields that should NOT be present
assert "files_checked" not in schema["properties"]

20
tests/test_image_support_integration.py
View File

@@ -26,6 +26,7 @@ from utils.conversation_memory import (
get_conversation_image_list,
get_thread,
)
from utils.model_context import ModelContext
@pytest.mark.no_mock_provider
@@ -180,17 +181,18 @@ class TestImageSupportIntegration:
try:
# Test with an invalid model name that doesn't exist in any provider
result = tool._validate_image_limits(small_images, "non-existent-model-12345")
# Use model_context parameter name (not positional)
result = tool._validate_image_limits(small_images, model_context=ModelContext("non-existent-model-12345"))
# Should return error because model not available or doesn't support images
assert result is not None
assert result["status"] == "error"
assert "is not available" in result["content"] or "does not support image processing" in result["content"]
# Test that empty/None images always pass regardless of model
result = tool._validate_image_limits([], "any-model")
result = tool._validate_image_limits([], model_context=ModelContext("gemini-2.5-pro"))
assert result is None
result = tool._validate_image_limits(None, "any-model")
result = tool._validate_image_limits(None, model_context=ModelContext("gemini-2.5-pro"))
assert result is None
finally:
@@ -215,7 +217,7 @@ class TestImageSupportIntegration:
small_image_path = temp_file.name
# Test with the default model from test environment (gemini-2.5-flash)
result = tool._validate_image_limits([small_image_path], "gemini-2.5-flash")
result = tool._validate_image_limits([small_image_path], ModelContext("gemini-2.5-flash"))
assert result is None # Should pass for Gemini models
# Create 150MB image (over typical limits)
@@ -223,7 +225,7 @@ class TestImageSupportIntegration:
temp_file.write(b"\x00" * (150 * 1024 * 1024)) # 150MB
large_image_path = temp_file.name
result = tool._validate_image_limits([large_image_path], "gemini-2.5-flash")
result = tool._validate_image_limits([large_image_path], ModelContext("gemini-2.5-flash"))
# Large images should fail validation
assert result is not None
assert result["status"] == "error"
@@ -429,14 +431,14 @@ class TestImageSupportIntegration:
images = [data_url]
# Test with a dummy model that doesn't exist in any provider
result = tool._validate_image_limits(images, "test-dummy-model-name")
result = tool._validate_image_limits(images, ModelContext("test-dummy-model-name"))
# Should return error because model not available or doesn't support images
assert result is not None
assert result["status"] == "error"
assert "is not available" in result["content"] or "does not support image processing" in result["content"]
# Test with another non-existent model to check error handling
result = tool._validate_image_limits(images, "another-dummy-model")
result = tool._validate_image_limits(images, ModelContext("another-dummy-model"))
# Should return error because model not available
assert result is not None
assert result["status"] == "error"
@@ -446,11 +448,11 @@ class TestImageSupportIntegration:
tool = ChatTool()
# Empty list should not fail validation (no need for provider setup)
result = tool._validate_image_limits([], "test_model")
result = tool._validate_image_limits([], ModelContext("gemini-2.5-pro"))
assert result is None
# None should not fail validation (no need for provider setup)
result = tool._validate_image_limits(None, "test_model")
result = tool._validate_image_limits(None, ModelContext("gemini-2.5-pro"))
assert result is None
@patch("utils.conversation_memory.get_storage")

41
tests/test_listmodels_restrictions.py
View File

@@ -70,11 +70,25 @@ class TestListModelsRestrictions(unittest.TestCase):
config = MagicMock()
config.model_name = "anthropic/claude-opus-4-20240229"
config.context_window = 200000
config.get_effective_capability_rank.return_value = 90 # High rank for Opus
return config
elif "sonnet" in model_name.lower():
config = MagicMock()
config.model_name = "anthropic/claude-sonnet-4-20240229"
config.context_window = 200000
config.get_effective_capability_rank.return_value = 80 # Lower rank for Sonnet
return config
elif "deepseek" in model_name.lower():
config = MagicMock()
config.model_name = "deepseek/deepseek-r1-0528:free"
config.context_window = 100000
config.get_effective_capability_rank.return_value = 70
return config
elif "qwen" in model_name.lower():
config = MagicMock()
config.model_name = "qwen/qwen3-235b-a22b-04-28:free"
config.context_window = 100000
config.get_effective_capability_rank.return_value = 60
return config
return None # No config for models without aliases
@@ -90,6 +104,9 @@ class TestListModelsRestrictions(unittest.TestCase):
mock_get_provider.side_effect = get_provider_side_effect
# Ensure registry is cleared before test
ModelProviderRegistry._registry = {}
# Mock available models
mock_get_models.return_value = {
"gemini-2.5-flash": ProviderType.GOOGLE,
@@ -131,6 +148,9 @@ class TestListModelsRestrictions(unittest.TestCase):
# Parse the output
lines = result.split("\n")
# Debug: print the actual result for troubleshooting
# print(f"DEBUG: Full result:\n{result}")
# Check that OpenRouter section exists
openrouter_section_found = False
openrouter_models = []
@@ -141,15 +161,18 @@ class TestListModelsRestrictions(unittest.TestCase):
openrouter_section_found = True
elif "Available Models" in line and openrouter_section_found:
in_openrouter_section = True
elif in_openrouter_section and line.strip().startswith("- "):
# Extract model name from various line formats:
# - `model-name` → `full-name` (context)
# - `model-name`
line_content = line.strip()[2:] # Remove "- "
if "`" in line_content:
# Extract content between first pair of backticks
model_name = line_content.split("`")[1]
openrouter_models.append(model_name)
elif in_openrouter_section:
# Check for lines with model names in backticks
# Format: - `model-name` (score X)
if line.strip().startswith("- ") and "`" in line:
# Extract model name between backticks
parts = line.split("`")
if len(parts) >= 2:
model_name = parts[1]
openrouter_models.append(model_name)
# Stop parsing when we hit the next section
elif "##" in line and in_openrouter_section:
break
self.assertTrue(openrouter_section_found, "OpenRouter section not found")
self.assertEqual(

2
tests/test_openrouter_provider.py
View File

@@ -174,6 +174,7 @@ class TestOpenRouterAutoMode:
mock_config = Mock()
mock_config.is_custom = False
mock_config.aliases = [] # Empty list of aliases
mock_config.get_effective_capability_rank = Mock(return_value=50) # Add ranking method
return mock_config
return None
@@ -220,6 +221,7 @@ class TestOpenRouterAutoMode:
# Mock the resolve method to return model configs with aliases
mock_model_config = Mock()
mock_model_config.aliases = [] # Empty aliases for simplicity
mock_model_config.get_effective_capability_rank = Mock(return_value=50) # Add ranking method
mock_registry.resolve.return_value = mock_model_config
ModelProviderRegistry.register_provider(ProviderType.OPENROUTER, OpenRouterProvider)