feat: extend image plugins

api/v1/image.py

@@ -73,8 +73,8 @@ async def process_image(request: ImageProcessingRequest):
         plugin_info.class_name, ImageProcessingFunction, str(filepath)
     )
 
-    extracted_text = await asyncio.get_running_loop().run_in_executor(
+    model_response = await asyncio.get_running_loop().run_in_executor(
         None, lambda: job.get(blocking=True, preserve=True)
     )
 
-    return ImageProcessingResponse(text=extracted_text)
+    return ImageProcessingResponse.parse_obj(model_response.dict())

api/v1/models.py

@@ -26,8 +26,26 @@ class ImageProcessingRequest(BaseModel):
     image_model: str
 
 
+class IPRPoint(BaseModel):
+    x: int
+    y: int
+
+
+class IPRRectangle(BaseModel):
+    left_top: IPRPoint
+    right_top: IPRPoint
+    left_bottom: IPRPoint
+    right_bottom: IPRPoint
+
+
+class IPRTextBox(BaseModel):
+    text: str
+    coordinates: IPRRectangle
+
+
 class ImageProcessingResponse(BaseModel):
     text: str
+    boxes: List[IPRTextBox]
 
 
 class AudioProcessingRequest(BaseModel):

core/plugins/__init__.py

@@ -4,4 +4,11 @@ from core.plugins.loader import (
     load_plugins,
     register_plugin,
 )
-from core.plugins.base import AudioChunk, AudioProcessingResult
+from core.plugins.base import (
+    AudioChunk,
+    AudioProcessingResult,
+    Point,
+    Rectangle,
+    ImageTextBox,
+    ImageProcessingResult,
+)

core/plugins/base.py

@@ -13,6 +13,28 @@ class AudioProcessingResult(BaseModel):
     segments: List[AudioChunk]
 
 
+class Point(BaseModel):
+    x: int
+    y: int
+
+
+class Rectangle(BaseModel):
+    left_top: Point
+    right_top: Point
+    left_bottom: Point
+    right_bottom: Point
+
+
+class ImageTextBox(BaseModel):
+    text: str
+    coordinates: Rectangle
+
+
+class ImageProcessingResult(BaseModel):
+    text: str
+    boxes: List[ImageTextBox]
+
+
 @runtime_checkable
 class BasePlugin(Protocol):
     """