pyagentic-core 2.2.1a2__tar.gz → 2.3.0a1__tar.gz

{pyagentic_core-2.2.1a2 → pyagentic_core-2.3.0a1}/PKG-INFO

@@ -1,6 +1,6 @@
 Metadata-Version: 2.4
 Name: pyagentic-core
-Version: 2.2.1a2
+Version: 2.3.0a1
 Summary: Build LLM Agents in a Pythonic way
 Author-email: Ryan Mikulec <rmikulec.dev@gmail.com>
 License: MIT
@@ -21,6 +21,7 @@ Requires-Dist: c3linearize>=0.1.0
 Requires-Dist: anthropic>=0.62.0
 Requires-Dist: google-generativeai>=0.8.0
 Requires-Dist: transitions>=0.9.3
+Requires-Dist: pillow>=12.1.0
 Dynamic: license-file
 
 # PyAgentic

{pyagentic_core-2.2.1a2 → pyagentic_core-2.3.0a1}/pyagentic/_base/_agent/_agent_state.py

@@ -60,10 +60,10 @@ class _AgentState(BaseModel):
         self._machine = machine
 
     def _update_state_machine(self, phases):
-        for to_, from_, condition in phases:
+        for source, dest, condition in phases:
             with self._state_lock:
-                if condition(self) and self.phase == from_:
-                    trigger = f"{to_}_to_{from_}"
+                if condition(self) and self.phase == source:
+                    trigger = f"{source}_to_{dest}"
                     getattr(self._machine, trigger)()
 
     def model_post_init(self, state):

{pyagentic_core-2.2.1a2 → pyagentic_core-2.3.0a1}/pyagentic/llm/_openai.py

@@ -9,8 +9,11 @@ import openai
 from openai.types.responses import Response as OpenAIResponse
 from openai.types.responses import ParsedResponse as OpenAIParsedResponse
 
+import base64
+import io
 from typing import List, Optional, Type
 from pydantic import BaseModel
+from PIL.Image import Image
 
 from pyagentic._base._agent._agent_state import _AgentState
 from pyagentic._base._tool import _ToolDefinition
@@ -85,10 +88,27 @@ class OpenAIProvider(LLMProvider):
         """
         return OpenAIMessage(type="function_call_output", call_id=id_, output=result)
 
+    def _encode_image(self, image: Image) -> str:
+        """
+        Convert a PIL Image to a base64-encoded data URL.
+
+        Args:
+            image: PIL Image object to encode
+
+        Returns:
+            Base64-encoded data URL string
+        """
+        buffer = io.BytesIO()
+        image.save(buffer, format="PNG")
+        image_bytes = buffer.getvalue()
+        base64_image = base64.b64encode(image_bytes).decode("utf-8")
+        return f"data:image/png;base64,{base64_image}"
+
     async def generate(
         self,
         state: _AgentState,
         *,
+        images: Optional[Image] = None,
         tool_defs: Optional[List[_ToolDefinition]] = None,
         response_format: Optional[Type[BaseModel]] = None,
         **kwargs,
@@ -102,6 +122,7 @@ class OpenAIProvider(LLMProvider):
 
         Args:
             state: Agent state containing conversation history and system messages
+            images: Optional PIL Image to include in the request
             tool_defs: List of available tools the model can call
             response_format: Optional Pydantic model for structured output
             **kwargs: Additional parameters for the OpenAI API call
@@ -113,11 +134,24 @@ class OpenAIProvider(LLMProvider):
         if tool_defs is None:
             tool_defs = []
 
+        # Prepare input messages
+        input_messages = [message.to_dict() for message in state._messages]
+
+        # Add image if provided
+        if images:
+            image_url = self._encode_image(images)
+            # Add image as a user message with image content
+            image_message = {
+                "role": "user",
+                "content": [{"type": "image_url", "image_url": {"url": image_url}}],
+            }
+            input_messages.append(image_message)
+
         if response_format:
             response: OpenAIParsedResponse[Type[BaseModel]] = await self.client.responses.parse(
                 model=self._model,
                 instructions=state.system_message,
-                input=[message.to_dict() for message in state._messages],
+                input=input_messages,
                 tools=[tool.to_openai_spec() for tool in tool_defs],
                 text_format=response_format,
                 **kwargs,
@@ -142,7 +176,7 @@ class OpenAIProvider(LLMProvider):
             response: OpenAIResponse = await self.client.responses.create(
                 model=self._model,
                 instructions=state.system_message,
-                input=[message.to_dict() for message in state._messages],
+                input=input_messages,
                 tools=[tool.to_openai_spec() for tool in tool_defs],
                 **kwargs,
             )

{pyagentic_core-2.2.1a2 → pyagentic_core-2.3.0a1}/pyagentic/llm/_provider.py

@@ -8,6 +8,7 @@ compatible with the pyagentic framework.
 from typing import Optional, Type
 from abc import ABC, abstractmethod
 from pydantic import BaseModel
+from PIL.Image import Image
 
 from pyagentic._base._tool import _ToolDefinition
 from pyagentic._base._agent._agent_state import _AgentState
@@ -80,6 +81,7 @@ class LLMProvider(ABC):
         self,
         state: _AgentState,
         *,
+        images: Optional[Image] = None,
         tool_defs: Optional[list[_ToolDefinition]] = None,
         response_format: Optional[Type[BaseModel]] = None,
         **kwargs,

{pyagentic_core-2.2.1a2 → pyagentic_core-2.3.0a1}/pyagentic_core.egg-info/PKG-INFO

@@ -1,6 +1,6 @@
 Metadata-Version: 2.4
 Name: pyagentic-core
-Version: 2.2.1a2
+Version: 2.3.0a1
 Summary: Build LLM Agents in a Pythonic way
 Author-email: Ryan Mikulec <rmikulec.dev@gmail.com>
 License: MIT
@@ -21,6 +21,7 @@ Requires-Dist: c3linearize>=0.1.0
 Requires-Dist: anthropic>=0.62.0
 Requires-Dist: google-generativeai>=0.8.0
 Requires-Dist: transitions>=0.9.3
+Requires-Dist: pillow>=12.1.0
 Dynamic: license-file
 
 # PyAgentic

{pyagentic_core-2.2.1a2 → pyagentic_core-2.3.0a1}/pyagentic_core.egg-info/requires.txt

@@ -9,3 +9,4 @@ c3linearize>=0.1.0
 anthropic>=0.62.0
 google-generativeai>=0.8.0
 transitions>=0.9.3
+pillow>=12.1.0

{pyagentic_core-2.2.1a2 → pyagentic_core-2.3.0a1}/pyproject.toml

@@ -1,6 +1,6 @@
 [project]
 name = "pyagentic-core"
-version = "2.2.1-a.2"
+version = "2.3.0-a.1"
 description = "Build LLM Agents in a Pythonic way"
 readme = "README.md"
 requires-python = ">=3.13"
@@ -24,6 +24,7 @@ dependencies = [
     "anthropic>=0.62.0",
     "google-generativeai>=0.8.0",
     "transitions>=0.9.3",
+    "pillow>=12.1.0",
 ]
 
 [dependency-groups]