Fix image input

app.py

@@ -160,6 +160,29 @@ def _clean_history_for_display(history: list[dict[str, Any]]) -> list[dict[str,
 
     return cleaned
 
+def format_message_with_image(
+    text: str, role: str, image: Optional[Image.Image] = None
+) -> Dict[str, Any]:
+    """Format message for VLLM API with optional image."""
+    if image is not None:
+        # Convert PIL image to base64
+        buffered = io.BytesIO()
+        image.save(buffered, format="JPEG")
+        img_base64 = base64.b64encode(buffered.getvalue()).decode()
+
+        return {
+            "role": role,
+            "content": [
+                {"type": "text", "text": text},
+                {
+                    "type": "image_url",
+                    "image_url": {"url": f"data:image/jpeg;base64,{img_base64}"},
+                },
+            ],
+        }
+    else:
+        return {"role": role, "content": text}
+
 
 @spaces.GPU
 def bot(
@@ -207,6 +230,8 @@ def bot(
                     formatted_content = []
                     for item in content:
                         if isinstance(item, dict):
+
+                            # Add text
                             if item.get("type") == "text":
                                 formatted_content.append({"type": "text", "text": item.get("text", "")})
                             elif item.get("type") == "image":
@@ -222,7 +247,14 @@ def bot(
                                     pil_image = Image.open(item["path"])
 
                                 if pil_image is not None:
-                                    formatted_content.append({"type": "image", "image": pil_image})
+                                    # formatted_content.append({"type": "image", "image": pil_image})
+                                    buffered = io.BytesIO()
+                                    pil_image.save(buffered, format="JPEG")
+                                    img_base64 = base64.b64encode(buffered.getvalue()).decode()
+                                    {
+                                        "type": "image_url",
+                                        "image_url": {"url": f"data:image/jpeg;base64,{img_base64}"},
+                                    }
                     if formatted_content:
                         processor_history.append({"role": role, "content": formatted_content})