Commit
•
9d7e9d0
1
Parent(s):
4fa5a46
Update app.py
Browse files
app.py
CHANGED
@@ -16,23 +16,30 @@ import numpy as np
|
|
16 |
import nltk
|
17 |
nltk.download('punkt')
|
18 |
|
19 |
-
#
|
20 |
USER_ID_IMG = 'openai'
|
21 |
APP_ID_IMG = 'dall-e'
|
22 |
MODEL_ID_IMG = 'dall-e-3'
|
23 |
MODEL_VERSION_ID_IMG = 'dc9dcb6ee67543cebc0b9a025861b868'
|
24 |
|
|
|
25 |
USER_ID_AUDIO = 'eleven-labs'
|
26 |
APP_ID_AUDIO = 'audio-generation'
|
27 |
MODEL_ID_AUDIO = 'speech-synthesis'
|
28 |
MODEL_VERSION_ID_AUDIO = 'f2cead3a965f4c419a61a4a9b501095c'
|
29 |
|
30 |
-
#
|
31 |
USER_ID_OBJECT = 'clarifai'
|
32 |
APP_ID_OBJECT = 'main'
|
33 |
MODEL_ID_OBJECT = 'general-image-detection'
|
34 |
MODEL_VERSION_ID_OBJECT = '1580bb1932594c93b7e2e04456af7c6f'
|
35 |
|
|
|
|
|
|
|
|
|
|
|
|
|
36 |
# Retrieve PAT from environment variable
|
37 |
PAT = os.getenv('CLARIFAI_PAT')
|
38 |
|
@@ -141,8 +148,43 @@ def get_image_concepts(image_bytes):
|
|
141 |
|
142 |
return post_model_outputs_response.outputs[0].data.regions
|
143 |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
144 |
|
145 |
-
|
146 |
|
147 |
# Function to split text into sentences and then chunk them
|
148 |
|
@@ -420,6 +462,10 @@ with tab3:
|
|
420 |
|
421 |
|
422 |
st.image(st.session_state['image_paths'][image_index])
|
|
|
|
|
|
|
|
|
423 |
|
424 |
# Button for actions related to the selected image
|
425 |
if st.button("Details", key=f"details_{image_index}"):
|
|
|
16 |
import nltk
|
17 |
nltk.download('punkt')
|
18 |
|
19 |
+
# Image Variables
|
20 |
USER_ID_IMG = 'openai'
|
21 |
APP_ID_IMG = 'dall-e'
|
22 |
MODEL_ID_IMG = 'dall-e-3'
|
23 |
MODEL_VERSION_ID_IMG = 'dc9dcb6ee67543cebc0b9a025861b868'
|
24 |
|
25 |
+
# Audio variables
|
26 |
USER_ID_AUDIO = 'eleven-labs'
|
27 |
APP_ID_AUDIO = 'audio-generation'
|
28 |
MODEL_ID_AUDIO = 'speech-synthesis'
|
29 |
MODEL_VERSION_ID_AUDIO = 'f2cead3a965f4c419a61a4a9b501095c'
|
30 |
|
31 |
+
# Object Detection variables
|
32 |
USER_ID_OBJECT = 'clarifai'
|
33 |
APP_ID_OBJECT = 'main'
|
34 |
MODEL_ID_OBJECT = 'general-image-detection'
|
35 |
MODEL_VERSION_ID_OBJECT = '1580bb1932594c93b7e2e04456af7c6f'
|
36 |
|
37 |
+
# Vision variables
|
38 |
+
USER_ID_GPT4 = 'openai'
|
39 |
+
APP_ID_GPT4 = 'chat-completion'
|
40 |
+
MODEL_ID_GPT4 = 'openai-gpt-4-vision'
|
41 |
+
MODEL_VERSION_ID_GPT4 = '266df29bc09843e0aee9b7bf723c03c2'
|
42 |
+
|
43 |
# Retrieve PAT from environment variable
|
44 |
PAT = os.getenv('CLARIFAI_PAT')
|
45 |
|
|
|
148 |
|
149 |
return post_model_outputs_response.outputs[0].data.regions
|
150 |
|
151 |
+
# GPT4 Image Description Creation +++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++
|
152 |
+
|
153 |
+
# Method to process the text output
|
154 |
+
def process_text_output(text_output):
|
155 |
+
st.write("Processed Text Output:")
|
156 |
+
st.write(text_output)
|
157 |
+
|
158 |
+
def analyze_image(uploaded_file):
|
159 |
+
|
160 |
+
channel = ClarifaiChannel.get_grpc_channel()
|
161 |
+
stub = service_pb2_grpc.V2Stub(channel)
|
162 |
+
metadata = (('authorization', 'Key ' + PAT),)
|
163 |
+
userDataObject = resources_pb2.UserAppIDSet(user_id=USER_ID_GPT4, app_id=APP_ID_GPT4)
|
164 |
+
|
165 |
+
try:
|
166 |
+
bytes_data = uploaded_file.getvalue()
|
167 |
+
|
168 |
+
response = stub.PostModelOutputs(
|
169 |
+
service_pb2.PostModelOutputsRequest(
|
170 |
+
user_app_id=userDataObject,
|
171 |
+
model_id=MODEL_ID_GPT4,
|
172 |
+
version_id=MODEL_VERSION_ID_GPT4,
|
173 |
+
inputs=[resources_pb2.Input(data=resources_pb2.Data(image=resources_pb2.Image(base64=bytes_data)))]
|
174 |
+
),
|
175 |
+
metadata=metadata
|
176 |
+
)
|
177 |
+
|
178 |
+
if response.status.code != SUCCESS:
|
179 |
+
st.error("Error in API call: " + response.status.description)
|
180 |
+
return None
|
181 |
+
|
182 |
+
return response.outputs[0].data.text.raw
|
183 |
+
|
184 |
+
except Exception as e:
|
185 |
+
st.error(f"An error occurred: {e}")
|
186 |
+
return None
|
187 |
|
|
|
188 |
|
189 |
# Function to split text into sentences and then chunk them
|
190 |
|
|
|
462 |
|
463 |
|
464 |
st.image(st.session_state['image_paths'][image_index])
|
465 |
+
|
466 |
+
image_text_output = analyze_image(st.session_state['image_paths'][image_index])
|
467 |
+
|
468 |
+
st.write(image_text_output)
|
469 |
|
470 |
# Button for actions related to the selected image
|
471 |
if st.button("Details", key=f"details_{image_index}"):
|