✨ Update the /v1/responses endpoint (#75)

* Remove the unused auto-refresh functionality and related imports.

They are no longer needed since the underlying library issue has been resolved.

* Enhance error handling in client initialization and message sending

* Refactor link handling to extract file paths and simplify Google search links

* Fix regex pattern for Google search link matching

* Fix regex patterns for Markdown escaping, code fence and Google search link matching

* Increase timeout value in configuration files from 60 to 120 seconds to better handle heavy tasks

* Fix Image generation

* Refactor tool handling to support standard and image generation tools separately

* Fix: use "ascii" decoding for base64-encoded image data consistency

* Fix: replace `running` with `_running` for internal client status checks

* Refactor: replace direct `_running` access with `running()` method in client status checks

* Extend models with new fields for annotations, reasoning, audio, log probabilities, and token details; adjust response handling accordingly.

* Extend models with new fields (annotations, error), add `normalize_output_text` validator, rename `created` to `created_at`, and update response handling accordingly.

* Extend response models to support tool choices, image output, and improved streaming of response items. Refactor image generation handling for consistency and add compatibility with output content.

* Set default `text` value to an empty string for `ResponseOutputContent` and ensure consistent initialization in image output handling.

* feat: Add /images endpoint with dedicated router and improved image management

Add dedicated router for /images endpoint and refactor image handling logic for better modularity. Enhance temporary image management with secure naming, token verification, and cleanup functionality.

* feat: Add token-based verification for image access

* Refactor: rename image store directory to `ai_generated_images` for clarity

* fix: Update create_response to use FastAPI Request object for base_url and refactor variable handling

* fix: Correct attribute access in request_data handling within `chat.py` for tools, tool_choice, and streaming settings

* fix: Save generated images to persistent storage

* fix: Remove unused `output_image` type from `ResponseOutputContent` and update response handling for consistency

* fix: Update image URL generation in chat response to use Markdown format for compatibility

Author: luuquangvu

app/main.py

@@ -6,7 +6,12 @@
 
 from .server.chat import router as chat_router
 from .server.health import router as health_router
-from .server.middleware import add_cors_middleware, add_exception_handler
+from .server.images import router as images_router
+from .server.middleware import (
+    add_cors_middleware,
+    add_exception_handler,
+    cleanup_expired_images,
+)
 from .services import GeminiClientPool, LMDBConversationStore
 
 RETENTION_CLEANUP_INTERVAL_SECONDS = 6 * 60 * 60  # 6 hours
@@ -28,6 +33,7 @@ async def _run_retention_cleanup(stop_event: asyncio.Event) -> None:
     while not stop_event.is_set():
         try:
             store.cleanup_expired()
+            cleanup_expired_images(store.retention_days)
         except Exception:
             logger.exception("LMDB retention cleanup task failed.")
 
@@ -93,5 +99,6 @@ def create_app() -> FastAPI:
 
     app.include_router(health_router, tags=["Health"])
     app.include_router(chat_router, tags=["Chat"])
+    app.include_router(images_router, tags=["Images"])
 
     return app

app/models/models.py

@@ -3,7 +3,7 @@
 from datetime import datetime
 from typing import Any, Dict, List, Literal, Optional, Union
 
-from pydantic import BaseModel, Field
+from pydantic import BaseModel, Field, model_validator
 
 
 class ContentItem(BaseModel):
@@ -12,7 +12,9 @@ class ContentItem(BaseModel):
     type: Literal["text", "image_url", "file", "input_audio"]
     text: Optional[str] = None
     image_url: Optional[Dict[str, str]] = None
+    input_audio: Optional[Dict[str, Any]] = None
     file: Optional[Dict[str, str]] = None
+    annotations: List[Dict[str, Any]] = Field(default_factory=list)
 
 
 class Message(BaseModel):
@@ -22,6 +24,10 @@ class Message(BaseModel):
     content: Union[str, List[ContentItem], None] = None
     name: Optional[str] = None
     tool_calls: Optional[List["ToolCall"]] = None
+    refusal: Optional[str] = None
+    reasoning_content: Optional[str] = None
+    audio: Optional[Dict[str, Any]] = None
+    annotations: List[Dict[str, Any]] = Field(default_factory=list)
 
 
 class Choice(BaseModel):
@@ -30,6 +36,7 @@ class Choice(BaseModel):
     index: int
     message: Message
     finish_reason: str
+    logprobs: Optional[Dict[str, Any]] = None
 
 
 class FunctionCall(BaseModel):
@@ -81,6 +88,8 @@ class Usage(BaseModel):
     prompt_tokens: int
     completion_tokens: int
     total_tokens: int
+    prompt_tokens_details: Optional[Dict[str, int]] = None
+    completion_tokens_details: Optional[Dict[str, int]] = None
 
 
 class ModelData(BaseModel):
@@ -161,6 +170,15 @@ class ResponseInputContent(BaseModel):
     file_url: Optional[str] = None
     file_data: Optional[str] = None
     filename: Optional[str] = None
+    annotations: List[Dict[str, Any]] = Field(default_factory=list)
+
+    @model_validator(mode="before")
+    @classmethod
+    def normalize_output_text(cls, data: Any) -> Any:
+        """Allow output_text (from previous turns) to be treated as input_text."""
+        if isinstance(data, dict) and data.get("type") == "output_text":
+            data["type"] = "input_text"
+        return data
 
 
 class ResponseInputItem(BaseModel):
@@ -216,7 +234,8 @@ class ResponseOutputContent(BaseModel):
     """Content item for Responses API output."""
 
     type: Literal["output_text"]
-    text: Optional[str] = None
+    text: Optional[str] = ""
+    annotations: List[Dict[str, Any]] = Field(default_factory=list)
 
 
 class ResponseOutputMessage(BaseModel):
@@ -254,20 +273,22 @@ class ResponseCreateResponse(BaseModel):
 
     id: str
     object: Literal["response"] = "response"
-    created: int
+    created_at: int
     model: str
     output: List[Union[ResponseOutputMessage, ResponseImageGenerationCall, ResponseToolCall]]
-    output_text: Optional[str] = None
     status: Literal[
         "in_progress",
         "completed",
         "failed",
         "incomplete",
+        "cancelled",
         "requires_action",
     ] = "completed"
+    tool_choice: Optional[Union[str, ResponseToolChoice]] = None
+    tools: Optional[List[Union[Tool, ResponseImageTool]]] = None
     usage: ResponseUsage
+    error: Optional[Dict[str, Any]] = None
     metadata: Optional[Dict[str, Any]] = None
-    system_fingerprint: Optional[str] = None
     input: Optional[Union[str, List[ResponseInputItem]]] = None