Update app.py

app.py

@@ -16,23 +16,30 @@ import numpy as np
 import nltk
 nltk.download('punkt')
 
-# Constants for image and audio generation
+# Image Variables
 USER_ID_IMG = 'openai'
 APP_ID_IMG = 'dall-e'
 MODEL_ID_IMG = 'dall-e-3'
 MODEL_VERSION_ID_IMG = 'dc9dcb6ee67543cebc0b9a025861b868'
 
+# Audio variables
 USER_ID_AUDIO = 'eleven-labs'
 APP_ID_AUDIO = 'audio-generation'
 MODEL_ID_AUDIO = 'speech-synthesis'
 MODEL_VERSION_ID_AUDIO = 'f2cead3a965f4c419a61a4a9b501095c'
 
-# Renamed variables
+# Object Detection variables
 USER_ID_OBJECT = 'clarifai'
 APP_ID_OBJECT = 'main'
 MODEL_ID_OBJECT = 'general-image-detection'
 MODEL_VERSION_ID_OBJECT = '1580bb1932594c93b7e2e04456af7c6f'
 
+# Vision variables
+USER_ID_GPT4 = 'openai'
+APP_ID_GPT4 = 'chat-completion'
+MODEL_ID_GPT4 = 'openai-gpt-4-vision'
+MODEL_VERSION_ID_GPT4 = '266df29bc09843e0aee9b7bf723c03c2'
+
 # Retrieve PAT from environment variable
 PAT = os.getenv('CLARIFAI_PAT')
 
@@ -141,8 +148,43 @@ def get_image_concepts(image_bytes):
 
     return post_model_outputs_response.outputs[0].data.regions
 
+# GPT4 Image Description Creation +++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++
+
+# Method to process the text output
+def process_text_output(text_output):
+    st.write("Processed Text Output:")
+    st.write(text_output)
+
+def analyze_image(uploaded_file):
+
+    channel = ClarifaiChannel.get_grpc_channel()
+    stub = service_pb2_grpc.V2Stub(channel)
+    metadata = (('authorization', 'Key ' + PAT),)
+    userDataObject = resources_pb2.UserAppIDSet(user_id=USER_ID_GPT4, app_id=APP_ID_GPT4)
+    
+    try:
+        bytes_data = uploaded_file.getvalue()
+
+        response = stub.PostModelOutputs(
+            service_pb2.PostModelOutputsRequest(
+                user_app_id=userDataObject,
+                model_id=MODEL_ID_GPT4,
+                version_id=MODEL_VERSION_ID_GPT4,
+                inputs=[resources_pb2.Input(data=resources_pb2.Data(image=resources_pb2.Image(base64=bytes_data)))]
+            ),
+            metadata=metadata
+        )
+
+        if response.status.code != SUCCESS:
+            st.error("Error in API call: " + response.status.description)
+            return None
+
+        return response.outputs[0].data.text.raw
+
+    except Exception as e:
+        st.error(f"An error occurred: {e}")
+        return None
 
-        
 
 # Function to split text into sentences and then chunk them
 
@@ -420,6 +462,10 @@ with tab3:
                 
                 
                 st.image(st.session_state['image_paths'][image_index])
+
+                image_text_output = analyze_image(st.session_state['image_paths'][image_index])
+
+                st.write(image_text_output)
         
                 # Button for actions related to the selected image
                 if st.button("Details", key=f"details_{image_index}"):