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

refactor: removed method from provider, should use model capabilities instead

Browse Source 
refactor: cleanup temperature factory method
This commit is contained in:
Fahad
2025-10-02 11:08:56 +04:00
parent 9c11ecc4bf
commit a254ff2220
25 changed files with 105 additions and 224 deletions
Download Patch File Download Diff File Expand all files Collapse all files

5
providers/base.py
View File

@@ -109,11 +109,6 @@ class ModelProvider(ABC):
constraint_desc = capabilities.temperature_constraint.get_description()
raise ValueError(f"Temperature {temperature} is invalid for model {model_name}. {constraint_desc}")
@abstractmethod
def supports_thinking_mode(self, model_name: str) -> bool:
"""Check if the model supports extended thinking mode."""
pass
def get_model_configurations(self) -> dict[str, ModelCapabilities]:
"""Get model configurations for this provider.

18
providers/custom.py
View File

@@ -284,24 +284,6 @@ class CustomProvider(OpenAICompatibleProvider):
**kwargs,
)
def supports_thinking_mode(self, model_name: str) -> bool:
"""Check if the model supports extended thinking mode.
Args:
model_name: Model to check
Returns:
True if model supports thinking mode, False otherwise
"""
# Check if model is in registry
config = self._registry.resolve(model_name) if self._registry else None
if config and config.is_custom:
# Trust the config from custom_models.json
return config.supports_extended_thinking
# Default to False for unknown models
return False
def get_model_configurations(self) -> dict[str, ModelCapabilities]:
"""Get model configurations from the registry.

25
providers/dial.py
View File

@@ -7,12 +7,7 @@ import time
from typing import Optional
from .openai_compatible import OpenAICompatibleProvider
from .shared import (
ModelCapabilities,
ModelResponse,
ProviderType,
create_temperature_constraint,
)
from .shared import ModelCapabilities, ModelResponse, ProviderType, TemperatureConstraint
logger = logging.getLogger(__name__)
@@ -48,7 +43,7 @@ class DIALModelProvider(OpenAICompatibleProvider):
supports_images=True,
max_image_size_mb=20.0,
supports_temperature=False, # O3 models don't accept temperature
temperature_constraint=create_temperature_constraint("fixed"),
temperature_constraint=TemperatureConstraint.create("fixed"),
description="OpenAI O3 via DIAL - Strong reasoning model",
aliases=["o3"],
),
@@ -66,7 +61,7 @@ class DIALModelProvider(OpenAICompatibleProvider):
supports_images=True,
max_image_size_mb=20.0,
supports_temperature=False, # O4 models don't accept temperature
temperature_constraint=create_temperature_constraint("fixed"),
temperature_constraint=TemperatureConstraint.create("fixed"),
description="OpenAI O4-mini via DIAL - Fast reasoning model",
aliases=["o4-mini"],
),
@@ -84,7 +79,7 @@ class DIALModelProvider(OpenAICompatibleProvider):
supports_images=True,
max_image_size_mb=5.0,
supports_temperature=True,
temperature_constraint=create_temperature_constraint("range"),
temperature_constraint=TemperatureConstraint.create("range"),
description="Claude Sonnet 4.1 via DIAL - Balanced performance",
aliases=["sonnet-4.1", "sonnet-4"],
),
@@ -102,7 +97,7 @@ class DIALModelProvider(OpenAICompatibleProvider):
supports_images=True,
max_image_size_mb=5.0,
supports_temperature=True,
temperature_constraint=create_temperature_constraint("range"),
temperature_constraint=TemperatureConstraint.create("range"),
description="Claude Sonnet 4.1 with thinking mode via DIAL",
aliases=["sonnet-4.1-thinking", "sonnet-4-thinking"],
),
@@ -120,7 +115,7 @@ class DIALModelProvider(OpenAICompatibleProvider):
supports_images=True,
max_image_size_mb=5.0,
supports_temperature=True,
temperature_constraint=create_temperature_constraint("range"),
temperature_constraint=TemperatureConstraint.create("range"),
description="Claude Opus 4.1 via DIAL - Most capable Claude model",
aliases=["opus-4.1", "opus-4"],
),
@@ -138,7 +133,7 @@ class DIALModelProvider(OpenAICompatibleProvider):
supports_images=True,
max_image_size_mb=5.0,
supports_temperature=True,
temperature_constraint=create_temperature_constraint("range"),
temperature_constraint=TemperatureConstraint.create("range"),
description="Claude Opus 4.1 with thinking mode via DIAL",
aliases=["opus-4.1-thinking", "opus-4-thinking"],
),
@@ -156,7 +151,7 @@ class DIALModelProvider(OpenAICompatibleProvider):
supports_images=True,
max_image_size_mb=20.0,
supports_temperature=True,
temperature_constraint=create_temperature_constraint("range"),
temperature_constraint=TemperatureConstraint.create("range"),
description="Gemini 2.5 Pro with Google Search via DIAL",
aliases=["gemini-2.5-pro-search"],
),
@@ -174,7 +169,7 @@ class DIALModelProvider(OpenAICompatibleProvider):
supports_images=True,
max_image_size_mb=20.0,
supports_temperature=True,
temperature_constraint=create_temperature_constraint("range"),
temperature_constraint=TemperatureConstraint.create("range"),
description="Gemini 2.5 Pro via DIAL - Deep reasoning",
aliases=["gemini-2.5-pro"],
),
@@ -192,7 +187,7 @@ class DIALModelProvider(OpenAICompatibleProvider):
supports_images=True,
max_image_size_mb=20.0,
supports_temperature=True,
temperature_constraint=create_temperature_constraint("range"),
temperature_constraint=TemperatureConstraint.create("range"),
description="Gemini 2.5 Flash via DIAL - Ultra-fast",
aliases=["gemini-2.5-flash"],
),

20
providers/gemini.py
View File

@@ -12,12 +12,7 @@ from google import genai
from google.genai import types
from .base import ModelProvider
from .shared import (
ModelCapabilities,
ModelResponse,
ProviderType,
create_temperature_constraint,
)
from .shared import ModelCapabilities, ModelResponse, ProviderType, TemperatureConstraint
logger = logging.getLogger(__name__)
@@ -46,7 +41,7 @@ class GeminiModelProvider(ModelProvider):
supports_images=True, # Vision capability
max_image_size_mb=32.0, # Higher limit for Pro model
supports_temperature=True,
temperature_constraint=create_temperature_constraint("range"),
temperature_constraint=TemperatureConstraint.create("range"),
max_thinking_tokens=32768, # Max thinking tokens for Pro model
description="Deep reasoning + thinking mode (1M context) - Complex problems, architecture, deep analysis",
aliases=["pro", "gemini pro", "gemini-pro"],
@@ -65,7 +60,7 @@ class GeminiModelProvider(ModelProvider):
supports_images=True, # Vision capability
max_image_size_mb=20.0, # Conservative 20MB limit for reliability
supports_temperature=True,
temperature_constraint=create_temperature_constraint("range"),
temperature_constraint=TemperatureConstraint.create("range"),
max_thinking_tokens=24576, # Same as 2.5 flash for consistency
description="Gemini 2.0 Flash (1M context) - Latest fast model with experimental thinking, supports audio/video input",
aliases=["flash-2.0", "flash2"],
@@ -84,7 +79,7 @@ class GeminiModelProvider(ModelProvider):
supports_images=False, # Does not support images
max_image_size_mb=0.0, # No image support
supports_temperature=True,
temperature_constraint=create_temperature_constraint("range"),
temperature_constraint=TemperatureConstraint.create("range"),
description="Gemini 2.0 Flash Lite (1M context) - Lightweight fast model, text-only",
aliases=["flashlite", "flash-lite"],
),
@@ -102,7 +97,7 @@ class GeminiModelProvider(ModelProvider):
supports_images=True, # Vision capability
max_image_size_mb=20.0, # Conservative 20MB limit for reliability
supports_temperature=True,
temperature_constraint=create_temperature_constraint("range"),
temperature_constraint=TemperatureConstraint.create("range"),
max_thinking_tokens=24576, # Flash 2.5 thinking budget limit
description="Ultra-fast (1M context) - Quick analysis, simple queries, rapid iterations",
aliases=["flash", "flash2.5"],
@@ -397,11 +392,6 @@ class GeminiModelProvider(ModelProvider):
return True
def supports_thinking_mode(self, model_name: str) -> bool:
"""Check if the model supports extended thinking mode."""
capabilities = self.get_capabilities(model_name)
return capabilities.supports_extended_thinking
def get_thinking_budget(self, model_name: str, thinking_mode: str) -> int:
"""Get actual thinking token budget for a model and thinking mode."""
resolved_name = self._resolve_model_name(model_name)

7
providers/openai_compatible.py
View File

@@ -734,13 +734,6 @@ class OpenAICompatibleProvider(ModelProvider):
"""
pass
def supports_thinking_mode(self, model_name: str) -> bool:
"""Check if the model supports extended thinking mode.
Default is False for OpenAI-compatible providers.
"""
return False
def _is_error_retryable(self, error: Exception) -> bool:
"""Determine if an error should be retried based on structured error codes.

30
providers/openai_provider.py
View File

@@ -7,12 +7,7 @@ if TYPE_CHECKING:
from tools.models import ToolModelCategory
from .openai_compatible import OpenAICompatibleProvider
from .shared import (
ModelCapabilities,
ModelResponse,
ProviderType,
create_temperature_constraint,
)
from .shared import ModelCapabilities, ModelResponse, ProviderType, TemperatureConstraint
logger = logging.getLogger(__name__)
@@ -41,7 +36,7 @@ class OpenAIModelProvider(OpenAICompatibleProvider):
supports_images=True, # GPT-5 supports vision
max_image_size_mb=20.0, # 20MB per OpenAI docs
supports_temperature=True, # Regular models accept temperature parameter
temperature_constraint=create_temperature_constraint("fixed"),
temperature_constraint=TemperatureConstraint.create("fixed"),
description="GPT-5 (400K context, 128K output) - Advanced model with reasoning support",
aliases=["gpt5"],
),
@@ -59,7 +54,7 @@ class OpenAIModelProvider(OpenAICompatibleProvider):
supports_images=True, # GPT-5-mini supports vision
max_image_size_mb=20.0, # 20MB per OpenAI docs
supports_temperature=True,
temperature_constraint=create_temperature_constraint("fixed"),
temperature_constraint=TemperatureConstraint.create("fixed"),
description="GPT-5-mini (400K context, 128K output) - Efficient variant with reasoning support",
aliases=["gpt5-mini", "gpt5mini", "mini"],
),
@@ -77,7 +72,7 @@ class OpenAIModelProvider(OpenAICompatibleProvider):
supports_images=True,
max_image_size_mb=20.0,
supports_temperature=True,
temperature_constraint=create_temperature_constraint("fixed"),
temperature_constraint=TemperatureConstraint.create("fixed"),
description="GPT-5 nano (400K context) - Fastest, cheapest version of GPT-5 for summarization and classification tasks",
aliases=["gpt5nano", "gpt5-nano", "nano"],
),
@@ -95,7 +90,7 @@ class OpenAIModelProvider(OpenAICompatibleProvider):
supports_images=True, # O3 models support vision
max_image_size_mb=20.0, # 20MB per OpenAI docs
supports_temperature=False, # O3 models don't accept temperature parameter
temperature_constraint=create_temperature_constraint("fixed"),
temperature_constraint=TemperatureConstraint.create("fixed"),
description="Strong reasoning (200K context) - Logical problems, code generation, systematic analysis",
aliases=[],
),
@@ -113,7 +108,7 @@ class OpenAIModelProvider(OpenAICompatibleProvider):
supports_images=True, # O3 models support vision
max_image_size_mb=20.0, # 20MB per OpenAI docs
supports_temperature=False, # O3 models don't accept temperature parameter
temperature_constraint=create_temperature_constraint("fixed"),
temperature_constraint=TemperatureConstraint.create("fixed"),
description="Fast O3 variant (200K context) - Balanced performance/speed, moderate complexity",
aliases=["o3mini"],
),
@@ -131,7 +126,7 @@ class OpenAIModelProvider(OpenAICompatibleProvider):
supports_images=True, # O3 models support vision
max_image_size_mb=20.0, # 20MB per OpenAI docs
supports_temperature=False, # O3 models don't accept temperature parameter
temperature_constraint=create_temperature_constraint("fixed"),
temperature_constraint=TemperatureConstraint.create("fixed"),
description="Professional-grade reasoning (200K context) - EXTREMELY EXPENSIVE: Only for the most complex problems requiring universe-scale complexity analysis OR when the user explicitly asks for this model. Use sparingly for critical architectural decisions or exceptionally complex debugging that other models cannot handle.",
aliases=["o3pro"],
),
@@ -149,7 +144,7 @@ class OpenAIModelProvider(OpenAICompatibleProvider):
supports_images=True, # O4 models support vision
max_image_size_mb=20.0, # 20MB per OpenAI docs
supports_temperature=False, # O4 models don't accept temperature parameter
temperature_constraint=create_temperature_constraint("fixed"),
temperature_constraint=TemperatureConstraint.create("fixed"),
description="Latest reasoning model (200K context) - Optimized for shorter contexts, rapid reasoning",
aliases=["o4mini"],
),
@@ -167,7 +162,7 @@ class OpenAIModelProvider(OpenAICompatibleProvider):
supports_images=True, # GPT-4.1 supports vision
max_image_size_mb=20.0, # 20MB per OpenAI docs
supports_temperature=True, # Regular models accept temperature parameter
temperature_constraint=create_temperature_constraint("range"),
temperature_constraint=TemperatureConstraint.create("range"),
description="GPT-4.1 (1M context) - Advanced reasoning model with large context window",
aliases=["gpt4.1"],
),
@@ -303,13 +298,6 @@ class OpenAIModelProvider(OpenAICompatibleProvider):
**kwargs,
)
def supports_thinking_mode(self, model_name: str) -> bool:
"""Check if the model supports extended thinking mode."""
try:
return self.get_capabilities(model_name).supports_extended_thinking
except ValueError:
return False
def get_preferred_model(self, category: "ToolModelCategory", allowed_models: list[str]) -> Optional[str]:
"""Get OpenAI's preferred model for a given category from allowed models.

14
providers/openrouter.py
View File

@@ -204,20 +204,6 @@ class OpenRouterProvider(OpenAICompatibleProvider):
**kwargs,
)
def supports_thinking_mode(self, model_name: str) -> bool:
"""Check if the model supports extended thinking mode.
Currently, no models via OpenRouter support extended thinking.
This may change as new models become available.
Args:
model_name: Model to check
Returns:
False (no OpenRouter models currently support thinking mode)
"""
return False
def list_models(
self,
*,

4
providers/openrouter_registry.py
View File

@@ -12,7 +12,7 @@ from utils.file_utils import read_json_file
from .shared import (
ModelCapabilities,
ProviderType,
create_temperature_constraint,
TemperatureConstraint,
)
@@ -178,7 +178,7 @@ class OpenRouterModelRegistry:
# Create ModelCapabilities directly from JSON data
# Handle temperature_constraint conversion
temp_constraint_str = model_data.get("temperature_constraint")
temp_constraint = create_temperature_constraint(temp_constraint_str or "range")
temp_constraint = TemperatureConstraint.create(temp_constraint_str or "range")
# Set provider-specific defaults based on is_custom flag
is_custom = model_data.get("is_custom", False)

2
providers/shared/__init__.py
View File

@@ -8,7 +8,6 @@ from .temperature import (
FixedTemperatureConstraint,
RangeTemperatureConstraint,
TemperatureConstraint,
create_temperature_constraint,
)
__all__ = [
@@ -19,5 +18,4 @@ __all__ = [
"FixedTemperatureConstraint",
"RangeTemperatureConstraint",
"DiscreteTemperatureConstraint",
"create_temperature_constraint",
]

35
providers/shared/temperature.py
View File

@@ -8,7 +8,6 @@ __all__ = [
"FixedTemperatureConstraint",
"RangeTemperatureConstraint",
"DiscreteTemperatureConstraint",
"create_temperature_constraint",
]
# Common heuristics for determining temperature support when explicit
@@ -102,7 +101,7 @@ class TemperatureConstraint(ABC):
"""
if constraint_hint:
constraint = create_temperature_constraint(constraint_hint)
constraint = TemperatureConstraint.create(constraint_hint)
supports_temperature = constraint_hint != "fixed"
reason = f"constraint hint '{constraint_hint}'"
return supports_temperature, constraint, reason
@@ -115,6 +114,19 @@ class TemperatureConstraint(ABC):
return supports_temperature, constraint, reason
@staticmethod
def create(constraint_type: str) -> "TemperatureConstraint":
"""Factory that yields the appropriate constraint for a configuration hint."""
if constraint_type == "fixed":
# Fixed temperature models (O3/O4) only support temperature=1.0
return FixedTemperatureConstraint(1.0)
if constraint_type == "discrete":
# For models with specific allowed values - using common OpenAI values as default
return DiscreteTemperatureConstraint([0.0, 0.3, 0.7, 1.0, 1.5, 2.0], 0.3)
# Default range constraint (for "range" or None)
return RangeTemperatureConstraint(0.0, 2.0, 0.3)
class FixedTemperatureConstraint(TemperatureConstraint):
"""Constraint for models that enforce an exact temperature (for example O3)."""
@@ -174,22 +186,3 @@ class DiscreteTemperatureConstraint(TemperatureConstraint):
def get_default(self) -> float:
return self.default_temp
def create_temperature_constraint(constraint_type: str) -> TemperatureConstraint:
"""Factory that yields the appropriate constraint for a model configuration.
The JSON configuration stored in ``conf/custom_models.json`` references this
helper via human-readable strings. Providers feed those values into this
function so that runtime logic can rely on strongly typed constraint
objects.
"""
if constraint_type == "fixed":
# Fixed temperature models (O3/O4) only support temperature=1.0
return FixedTemperatureConstraint(1.0)
if constraint_type == "discrete":
# For models with specific allowed values - using common OpenAI values as default
return DiscreteTemperatureConstraint([0.0, 0.3, 0.7, 1.0, 1.5, 2.0], 0.3)
# Default range constraint (for "range" or None)
return RangeTemperatureConstraint(0.0, 2.0, 0.3)

21
providers/xai.py
View File

@@ -7,12 +7,7 @@ if TYPE_CHECKING:
from tools.models import ToolModelCategory
from .openai_compatible import OpenAICompatibleProvider
from .shared import (
ModelCapabilities,
ModelResponse,
ProviderType,
create_temperature_constraint,
)
from .shared import ModelCapabilities, ModelResponse, ProviderType, TemperatureConstraint
logger = logging.getLogger(__name__)
@@ -42,7 +37,7 @@ class XAIModelProvider(OpenAICompatibleProvider):
supports_images=True, # Multimodal capabilities
max_image_size_mb=20.0, # Standard image size limit
supports_temperature=True,
temperature_constraint=create_temperature_constraint("range"),
temperature_constraint=TemperatureConstraint.create("range"),
description="GROK-4 (256K context) - Frontier multimodal reasoning model with advanced capabilities",
aliases=["grok", "grok4", "grok-4"],
),
@@ -60,7 +55,7 @@ class XAIModelProvider(OpenAICompatibleProvider):
supports_images=False, # Assuming GROK is text-only for now
max_image_size_mb=0.0,
supports_temperature=True,
temperature_constraint=create_temperature_constraint("range"),
temperature_constraint=TemperatureConstraint.create("range"),
description="GROK-3 (131K context) - Advanced reasoning model from X.AI, excellent for complex analysis",
aliases=["grok3"],
),
@@ -78,7 +73,7 @@ class XAIModelProvider(OpenAICompatibleProvider):
supports_images=False, # Assuming GROK is text-only for now
max_image_size_mb=0.0,
supports_temperature=True,
temperature_constraint=create_temperature_constraint("range"),
temperature_constraint=TemperatureConstraint.create("range"),
description="GROK-3 Fast (131K context) - Higher performance variant, faster processing but more expensive",
aliases=["grok3fast", "grokfast", "grok3-fast"],
),
@@ -153,14 +148,6 @@ class XAIModelProvider(OpenAICompatibleProvider):
**kwargs,
)
def supports_thinking_mode(self, model_name: str) -> bool:
"""Check if the model supports extended thinking mode."""
resolved_name = self._resolve_model_name(model_name)
capabilities = self.MODEL_CAPABILITIES.get(resolved_name)
if capabilities:
return capabilities.supports_extended_thinking
return False
def get_preferred_model(self, category: "ToolModelCategory", allowed_models: list[str]) -> Optional[str]:
"""Get XAI's preferred model for a given category from allowed models.