added image detection code to display predicted bboxes

QA_bot.py

@@ -44,7 +44,7 @@ def tyre_synap_bot(filter_agent,image_file_path):
         with st.chat_message("assistant"):
             message_placeholder = st.empty()
             full_response = ""
-            if 'mask' in ai_response['output']:
+            if 'mask' in ai_response['output'] or 'predicted_image' in ai_response['output']:
                 display_mask_image('final_mask.png')
             
             for chunk in re.split(r'(\s+)', response):

extract_tools.py

@@ -44,7 +44,7 @@ def get_groq_model(model_name = "gemma2-9b-it"):
 def panoptic_image_segemntation(image_path:str)->str:
     """
     The tool is used to create a Panoptic segmentation mask . It uses Maskformer network to create a panoptic segmentation of all \
-    the objects present in the image . Use the tool in case user ask to create a panoptic segmentation.
+    the objects present in the image . Use the tool in case user ask to create a panoptic image segmentation or panoptic mask .
     """
     if image_path.startswith('https'):
         image = Image.open(requests.get(image_path, stream=True).raw).convert('RGB')
@@ -240,7 +240,8 @@ def get_all_tools():
             func = panoptic_image_segemntation,
             description = "The tool is used to create a Panoptic segmentation mask . It uses Maskformer network to create a panoptic segmentation of all \
                         the objects present in the image . Use the tool in case user ask to create a panoptic segmentation or count objects in the image.\
-                        The tool also provides a list of objects along with the mask image of the all segmented objects found in the image ."
+                        The tool also provides a list of objects along with the mask image of the all segmented objects found in the image .\
+                        Us the tool if user ask to create a  panoptic image segmentation or panoptic mask"
         )
 
         tools = [

tool_utils/yolo_world.py

@@ -78,6 +78,6 @@ class YoloWorld:
                          boxes=processed_predictions[0]['boxes'],
                          labels=processed_predictions[0]['labels']
                          )
-        cv2.imwrite('final_mask.jpg',detected_image)
-        return "Predicted image : final_mask.jpg . Details :{}".format(processed_predictions[0])
+        cv2.imwrite('final_mask.png',detected_image)
+        return "Predicted image mask : final_mask.jpg . Details :{}".format(processed_predictions[0])