Spaces:
Sleeping
Sleeping
Update app.py
Browse files
app.py
CHANGED
|
@@ -146,64 +146,134 @@ def encode_image(image_file):
|
|
| 146 |
return base64.b64encode(image_file.getvalue()).decode("utf-8")
|
| 147 |
|
| 148 |
# Streamlit page setup
|
| 149 |
-
st.set_page_config(page_title="MTSS Image Accessibility Alt Text Generator", layout="centered", initial_sidebar_state="
|
| 150 |
|
| 151 |
-
#
|
| 152 |
-
|
|
|
|
| 153 |
|
|
|
|
| 154 |
st.header('VisionText™ | Accessibility')
|
| 155 |
st.subheader(':green[_Image Alt Text Generator_]')
|
| 156 |
|
| 157 |
# Retrieve the OpenAI API Key from secrets
|
| 158 |
openai.api_key = st.secrets["openai_api_key"]
|
| 159 |
|
|
|
|
|
|
|
|
|
|
| 160 |
# Move the file uploader to the sidebar
|
| 161 |
uploaded_file = st.sidebar.file_uploader("Upload an image", type=["jpg", "png", "jpeg"])
|
| 162 |
|
| 163 |
-
|
| 164 |
-
|
| 165 |
-
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 166 |
|
| 167 |
-
# Define toggles and input areas for additional details in the main area, not the sidebar
|
| 168 |
-
show_details = st.toggle("Add details about the image.", value=False)
|
| 169 |
if show_details:
|
|
|
|
| 170 |
additional_details = st.text_area(
|
| 171 |
-
"The details could include specific information that is important to include in the alt text or reflect why the image is being used:"
|
|
|
|
| 172 |
)
|
| 173 |
|
| 174 |
-
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 175 |
|
| 176 |
# Button to trigger the analysis
|
| 177 |
-
analyze_button = st.button("Analyze the Image")
|
| 178 |
|
| 179 |
-
#
|
| 180 |
-
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 181 |
|
| 182 |
-
#
|
| 183 |
-
if uploaded_file:
|
| 184 |
-
with st.expander("Uploaded Image Preview"):
|
| 185 |
-
st.image(uploaded_file, caption="Uploaded Image", width=300) # Display uploaded image with adjusted width
|
| 186 |
|
| 187 |
-
|
| 188 |
-
|
| 189 |
-
with st.spinner("Analyzing the image..."):
|
| 190 |
-
# Encode the image for analysis
|
| 191 |
base64_image = encode_image(uploaded_file)
|
| 192 |
|
| 193 |
-
#
|
| 194 |
-
|
| 195 |
-
|
| 196 |
-
|
| 197 |
-
|
| 198 |
-
|
| 199 |
-
|
| 200 |
-
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 201 |
|
| 202 |
-
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 203 |
else:
|
| 204 |
-
#
|
| 205 |
-
|
| 206 |
-
|
| 207 |
-
|
|
|
|
|
|
|
| 208 |
|
| 209 |
|
|
|
|
| 146 |
return base64.b64encode(image_file.getvalue()).decode("utf-8")
|
| 147 |
|
| 148 |
# Streamlit page setup
|
| 149 |
+
st.set_page_config(page_title="MTSS Image Accessibility Alt Text Generator", layout="centered", initial_sidebar_state="expanded")
|
| 150 |
|
| 151 |
+
#Add the image with a specified width
|
| 152 |
+
image_width = 300 # Set the desired width in pixels
|
| 153 |
+
st.image('MTSS.ai_Logo.png', width=image_width)
|
| 154 |
|
| 155 |
+
# st.title('MTSS:grey[.ai]')
|
| 156 |
st.header('VisionText™ | Accessibility')
|
| 157 |
st.subheader(':green[_Image Alt Text Generator_]')
|
| 158 |
|
| 159 |
# Retrieve the OpenAI API Key from secrets
|
| 160 |
openai.api_key = st.secrets["openai_api_key"]
|
| 161 |
|
| 162 |
+
# File uploader allows user to add their own image
|
| 163 |
+
# uploaded_file = st.file_uploader("Upload an image", type=["jpg", "png", "jpeg"])
|
| 164 |
+
|
| 165 |
# Move the file uploader to the sidebar
|
| 166 |
uploaded_file = st.sidebar.file_uploader("Upload an image", type=["jpg", "png", "jpeg"])
|
| 167 |
|
| 168 |
+
if uploaded_file:
|
| 169 |
+
# Display the uploaded image with specified width
|
| 170 |
+
image_width = 100 # Set the desired width in pixels
|
| 171 |
+
with st.expander("Image", expanded=True):
|
| 172 |
+
st.image(uploaded_file, caption=uploaded_file.name, width=image_width, use_column_width=False)
|
| 173 |
+
|
| 174 |
+
# Toggle for showing additional details input
|
| 175 |
+
show_details = st.toggle("Add details about the image. ", value=False)
|
| 176 |
|
|
|
|
|
|
|
| 177 |
if show_details:
|
| 178 |
+
# Text input for additional details about the image, shown only if toggle is True
|
| 179 |
additional_details = st.text_area(
|
| 180 |
+
"The details could include specific information that is important to include in the alt text or reflect why the image is being used:",
|
| 181 |
+
disabled=not show_details
|
| 182 |
)
|
| 183 |
|
| 184 |
+
# Toggle for modifying the prompt for complex images
|
| 185 |
+
complex_image = st.toggle("Is this a complex image? ", value=False)
|
| 186 |
+
|
| 187 |
+
if complex_image:
|
| 188 |
+
# Text input for additional details about the image, shown only if toggle is True
|
| 189 |
+
complex_image_details = st.caption(
|
| 190 |
+
"By clicking this toggle, it will inform MTSS.ai to create a description that exceeds the 125 character limit. "
|
| 191 |
+
"Add the description in a placeholder behind the image and 'Description in the content placeholder' in the alt text box. "
|
| 192 |
+
)
|
| 193 |
|
| 194 |
# Button to trigger the analysis
|
| 195 |
+
analyze_button = st.button("Analyze the Image", type="secondary")
|
| 196 |
|
| 197 |
+
# Optimized prompt for complex images
|
| 198 |
+
complex_image_prompt_text = (
|
| 199 |
+
"As an expert in image accessibility and alternative text, thoroughly describe the image provided. "
|
| 200 |
+
"Provide a brief description using not more than 500 characters that convey the essential information conveyed by the image in eight or fewer clear and concise sentences. "
|
| 201 |
+
"Skip phrases like 'image of' or 'picture of.' "
|
| 202 |
+
"Your description should form a clear, well-structured, and factual paragraph that avoids bullet points, focusing on creating a seamless narrative."
|
| 203 |
+
)
|
| 204 |
|
| 205 |
+
# Check if an image has been uploaded, if the API key is available, and if the button has been pressed
|
| 206 |
+
if uploaded_file is not None and analyze_button:
|
|
|
|
|
|
|
| 207 |
|
| 208 |
+
with st.spinner("Analyzing the image ..."):
|
| 209 |
+
# Encode the image
|
|
|
|
|
|
|
| 210 |
base64_image = encode_image(uploaded_file)
|
| 211 |
|
| 212 |
+
# Determine which prompt to use based on the complexity of the image
|
| 213 |
+
if complex_image:
|
| 214 |
+
prompt_text = complex_image_prompt_text
|
| 215 |
+
else:
|
| 216 |
+
prompt_text = (
|
| 217 |
+
"As an expert in image accessibility and alternative text, succinctly describe the image provided in less than 125 characters. "
|
| 218 |
+
"Provide a brief description using not more than 125 characters that convey the essential information conveyed by the image in three or fewer clear and concise sentences for use as alt text. "
|
| 219 |
+
"Skip phrases like 'image of' or 'picture of.' "
|
| 220 |
+
"Your description should form a clear, well-structured, and factual paragraph that avoids bullet points and newlines, focusing on creating a seamless narrative that serves as effective alternative text for accessibility purposes."
|
| 221 |
+
)
|
| 222 |
+
|
| 223 |
+
if show_details and additional_details:
|
| 224 |
+
prompt_text += (
|
| 225 |
+
f"\n\nAdditional Context Provided by the User:\n{additional_details}"
|
| 226 |
+
)
|
| 227 |
|
| 228 |
+
|
| 229 |
+
# Create the payload for the completion request
|
| 230 |
+
messages = [
|
| 231 |
+
{
|
| 232 |
+
"role": "user",
|
| 233 |
+
"content": [
|
| 234 |
+
{"type": "text", "text": prompt_text},
|
| 235 |
+
{
|
| 236 |
+
"type": "image_url",
|
| 237 |
+
"image_url": f"data:image/jpeg;base64,{base64_image}",
|
| 238 |
+
},
|
| 239 |
+
],
|
| 240 |
+
}
|
| 241 |
+
]
|
| 242 |
+
|
| 243 |
+
# Make the request to the OpenAI API
|
| 244 |
+
try:
|
| 245 |
+
# Without Stream
|
| 246 |
+
|
| 247 |
+
# response = openai.chat.completions.create(
|
| 248 |
+
# model="gpt-4-vision-preview", messages=messages, max_tokens=250, stream=False
|
| 249 |
+
# )
|
| 250 |
+
|
| 251 |
+
# Stream the response
|
| 252 |
+
full_response = ""
|
| 253 |
+
message_placeholder = st.empty()
|
| 254 |
+
for completion in client.chat.completions.create(
|
| 255 |
+
model="gpt-4-vision-preview", messages=messages,
|
| 256 |
+
max_tokens=1200, stream=True
|
| 257 |
+
):
|
| 258 |
+
# Check if there is content to display
|
| 259 |
+
if completion.choices[0].delta.content is not None:
|
| 260 |
+
full_response += completion.choices[0].delta.content
|
| 261 |
+
message_placeholder.markdown(full_response + "▌")
|
| 262 |
+
# Final update to placeholder after the stream ends
|
| 263 |
+
message_placeholder.markdown(full_response)
|
| 264 |
+
|
| 265 |
+
# # Display the response in a text area
|
| 266 |
+
# st.text_area('Response:', value=full_response, height=250, key="response_text_area")
|
| 267 |
+
|
| 268 |
+
st.success('Powered by MTSS GPT. AI can make mistakes. Consider checking important information.')
|
| 269 |
+
except Exception as e:
|
| 270 |
+
st.error(f"An error occurred: {e}")
|
| 271 |
else:
|
| 272 |
+
# Warnings for user action required
|
| 273 |
+
if not uploaded_file and analyze_button:
|
| 274 |
+
st.warning("Please upload an image.")
|
| 275 |
+
|
| 276 |
+
|
| 277 |
+
|
| 278 |
|
| 279 |
|