inferencesh 0.2.25__py3-none-any.whl → 0.2.27__py3-none-any.whl

inferencesh/models/llm.py

@@ -1,6 +1,6 @@
 from typing import Optional, List, Any, Callable, Dict, Generator
 from enum import Enum
-from pydantic import Field
+from pydantic import Field, BaseModel
 from queue import Queue
 from threading import Thread
 import time
@@ -9,6 +9,7 @@ import base64
 
 from .base import BaseAppInput, BaseAppOutput
 from .file import File
+from .types import ContextMessage
 
 class ContextMessageRole(str, Enum):
     USER = "user"
@@ -33,7 +34,8 @@ class ContextMessage(BaseAppInput):
         default=None
     )
 
-class LLMInput(BaseAppInput):
+class BaseLLMInput(BaseAppInput):
+    """Base class with common LLM fields."""
     system_prompt: str = Field(
         description="The system prompt to use for the model",
         default="You are a helpful assistant that can answer questions and help with tasks.",
@@ -47,25 +49,13 @@ class LLMInput(BaseAppInput):
     )
     context: List[ContextMessage] = Field(
         description="The context to use for the model",
+        default=[],
         examples=[
             [
-                {"role": "user", "content": [{"type": "text", "text": "What is the capital of France?"}]}, 
+                {"role": "user", "content": [{"type": "text", "text": "What is the capital of France?"}]},
                 {"role": "assistant", "content": [{"type": "text", "text": "The capital of France is Paris."}]}
-            ],
-            [
-                {"role": "user", "content": [{"type": "text", "text": "What is the weather like today?"}]}, 
-                {"role": "assistant", "content": [{"type": "text", "text": "I apologize, but I don't have access to real-time weather information. You would need to check a weather service or app to get current weather conditions for your location."}]}
-            ],
-            [
-                {"role": "user", "content": [{"type": "text", "text": "Can you help me write a poem about spring?"}]}, 
-                {"role": "assistant", "content": [{"type": "text", "text": "Here's a short poem about spring:\n\nGreen buds awakening,\nSoft rain gently falling down,\nNew life springs anew.\n\nWarm sun breaks through clouds,\nBirds return with joyful song,\nNature's sweet rebirth."}]}
-            ],
-            [
-                {"role": "user", "content": [{"type": "text", "text": "Explain quantum computing in simple terms"}]}, 
-                {"role": "assistant", "content": [{"type": "text", "text": "Quantum computing is like having a super-powerful calculator that can solve many problems at once instead of one at a time. While regular computers use bits (0s and 1s), quantum computers use quantum bits or \"qubits\" that can be both 0 and 1 at the same time - kind of like being in two places at once! This allows them to process huge amounts of information much faster than regular computers for certain types of problems."}]}
             ]
-        ],
-        default=[]
+        ]
     )
     text: str = Field(
         description="The user prompt to use for the model",
@@ -74,22 +64,41 @@ class LLMInput(BaseAppInput):
             "What is the weather like today?",
             "Can you help me write a poem about spring?",
             "Explain quantum computing in simple terms"
-        ],
-    )
-    image: Optional[File] = Field(
-        description="The image to use for the model",
-        default=None
+        ]
     )
-    # Optional parameters
     temperature: float = Field(default=0.7)
     top_p: float = Field(default=0.95)
     max_tokens: int = Field(default=4096)
     context_size: int = Field(default=4096)
-    
-    # Model specific flags
-    reasoning: bool = Field(default=False)
-    
-    tools: List[Dict[str, Any]] = Field(default=[])
+
+class ImageCapabilityMixin(BaseModel):
+    """Mixin for models that support image inputs."""
+    image: Optional[File] = Field(
+        description="The image to use for the model",
+        default=None
+    )
+
+class ReasoningCapabilityMixin(BaseModel):
+    """Mixin for models that support reasoning."""
+    reasoning: bool = Field(
+        description="Enable step-by-step reasoning",
+        default=False
+    )
+
+class ToolsCapabilityMixin(BaseModel):
+    """Mixin for models that support tool/function calling."""
+    tools: Optional[List[Dict[str, Any]]] = Field(
+        description="Tool definitions for function calling",
+        default=None
+    )
+
+# Example of how to use:
+class LLMInput(BaseLLMInput):
+    """Default LLM input model with no special capabilities."""
+    pass
+
+# For backward compatibility
+LLMInput.model_config["title"] = "LLMInput"
 
 class LLMUsage(BaseAppOutput):
     stop_reason: str = ""
@@ -102,12 +111,24 @@ class LLMUsage(BaseAppOutput):
     reasoning_time: float = 0.0
 
 
-class LLMOutput(BaseAppOutput):
-    response: str
-    reasoning: Optional[str] = None
-    tool_calls: Optional[List[Dict[str, Any]]] = None
-    usage: Optional[LLMUsage] = None
+class BaseLLMOutput(BaseAppOutput):
+    """Base class for LLM outputs with common fields."""
+    text: str = Field(description="The generated text response")
+    done: bool = Field(default=False, description="Whether this is the final chunk")
+
+class LLMUsageMixin(BaseModel):
+    """Mixin for models that provide token usage statistics."""
+    usage: Optional[LLMUsage] = Field(
+        description="Token usage statistics"
+    )
+
+# Example of how to use:
+class LLMOutput(BaseLLMOutput, LLMUsageMixin):
+    """Default LLM output model with token usage tracking."""
+    pass
 
+# For backward compatibility
+LLMOutput.model_config["title"] = "LLMOutput"
 
 @contextmanager
 def timing_context():

{inferencesh-0.2.25.dist-info → inferencesh-0.2.27.dist-info}/METADATA

@@ -1,6 +1,6 @@
 Metadata-Version: 2.4
 Name: inferencesh
-Version: 0.2.25
+Version: 0.2.27
 Summary: inference.sh Python SDK
 Author: Inference Shell Inc.
 Author-email: "Inference Shell Inc." <hello@inference.sh>

{inferencesh-0.2.25.dist-info → inferencesh-0.2.27.dist-info}/RECORD

@@ -2,13 +2,13 @@ inferencesh/__init__.py,sha256=WdADtOhfa3HDOunoE9HLFCTFlXRykYstBIH1FpyWvj8,613
 inferencesh/models/__init__.py,sha256=FDwcdtT6c4hbRitymjmN-hZMlQa8RbKSftkZZyjtUXA,536
 inferencesh/models/base.py,sha256=4gZQRi8J7y9U6PrGD9pRIehd1MJVJAqGakPQDs2AKFM,3251
 inferencesh/models/file.py,sha256=5xnpypcRahM1YcEjj64rv9g2gTimxrZb41YT4r440hU,7393
-inferencesh/models/llm.py,sha256=9d4JOlieJ-2bvwZQEAA69Qkmqk045gwibyzMcgkreFE,22362
+inferencesh/models/llm.py,sha256=CryJ6cQRmFVHrhR6zsTYYpBDQFH7DmwPn-7JFE3jI8M,22030
 inferencesh/utils/__init__.py,sha256=-xiD6uo2XzcrPAWFb_fUbaimmnW4KFKc-8IvBzaxNd4,148
 inferencesh/utils/download.py,sha256=7n5twvoNYDcFnKJyefImaj2YfzRI7vddQw4usZbj38c,1521
 inferencesh/utils/storage.py,sha256=E4J8emd4eFKdmdDgAqzz3TpaaDd3n0l8gYlMHuY8yIU,519
-inferencesh-0.2.25.dist-info/licenses/LICENSE,sha256=OsgqEWIh2el_QMj0y8O1A5Q5Dl-dxqqYbFE6fszuR4s,1086
-inferencesh-0.2.25.dist-info/METADATA,sha256=6PsBcetzAtbzXTGHBHVNkzoNVkRB9W49oaXJ0WYhBzY,2757
-inferencesh-0.2.25.dist-info/WHEEL,sha256=_zCd3N1l69ArxyTb8rzEoP9TpbYXkqRFSNOD5OuxnTs,91
-inferencesh-0.2.25.dist-info/entry_points.txt,sha256=6IC-fyozAqW3ljsMLGCXxJ0_ui2Jb-2fLHtoH1RTnEE,45
-inferencesh-0.2.25.dist-info/top_level.txt,sha256=TSMHg3T1ThMl1HGAWmzBClwOYH1ump5neof9BfHIwaA,12
-inferencesh-0.2.25.dist-info/RECORD,,
+inferencesh-0.2.27.dist-info/licenses/LICENSE,sha256=OsgqEWIh2el_QMj0y8O1A5Q5Dl-dxqqYbFE6fszuR4s,1086
+inferencesh-0.2.27.dist-info/METADATA,sha256=Hl5m-k_2dUxK0SgDrKIwSc1JugdPoRQoqJoXyh_a63I,2757
+inferencesh-0.2.27.dist-info/WHEEL,sha256=_zCd3N1l69ArxyTb8rzEoP9TpbYXkqRFSNOD5OuxnTs,91
+inferencesh-0.2.27.dist-info/entry_points.txt,sha256=6IC-fyozAqW3ljsMLGCXxJ0_ui2Jb-2fLHtoH1RTnEE,45
+inferencesh-0.2.27.dist-info/top_level.txt,sha256=TSMHg3T1ThMl1HGAWmzBClwOYH1ump5neof9BfHIwaA,12
+inferencesh-0.2.27.dist-info/RECORD,,