Spaces:
Running
Running
Reality123b
commited on
Update app.py
Browse files
app.py
CHANGED
@@ -5,8 +5,7 @@ import gradio as gr
|
|
5 |
from huggingface_hub import InferenceClient
|
6 |
from dataclasses import dataclass
|
7 |
import pytesseract
|
8 |
-
from PIL import Image
|
9 |
-
import io
|
10 |
|
11 |
@dataclass
|
12 |
class ChatMessage:
|
@@ -31,8 +30,8 @@ class XylariaChat:
|
|
31 |
api_key=self.hf_token
|
32 |
)
|
33 |
|
34 |
-
# Image captioning API setup
|
35 |
-
self.image_api_url = "https://api-inference.huggingface.co/models/
|
36 |
self.image_api_headers = {"Authorization": f"Bearer {self.hf_token}"}
|
37 |
|
38 |
# Initialize conversation history and persistent memory
|
@@ -75,47 +74,38 @@ class XylariaChat:
|
|
75 |
"""
|
76 |
Caption an uploaded image using Hugging Face API
|
77 |
Args:
|
78 |
-
image (str
|
79 |
Returns:
|
80 |
-
str:
|
81 |
"""
|
82 |
try:
|
83 |
-
#
|
84 |
-
if
|
85 |
-
image
|
86 |
-
|
87 |
-
|
88 |
-
|
89 |
-
#
|
90 |
-
if
|
91 |
-
|
92 |
-
|
93 |
-
|
94 |
-
|
95 |
-
|
96 |
-
|
97 |
-
|
98 |
-
|
99 |
-
|
100 |
-
|
101 |
-
|
102 |
-
|
103 |
-
# Send request to Hugging Face API
|
104 |
-
response = requests.post(
|
105 |
-
self.image_api_url,
|
106 |
-
headers=self.image_api_headers,
|
107 |
-
data=data
|
108 |
-
)
|
109 |
-
|
110 |
-
# Check response
|
111 |
-
if response.status_code == 200:
|
112 |
-
caption = response.json()[0].get('generated_text', 'No caption generated')
|
113 |
-
captions.append(caption)
|
114 |
-
else:
|
115 |
-
captions.append(f"Error captioning image: {response.status_code} - {response.text}")
|
116 |
|
117 |
-
#
|
118 |
-
|
|
|
|
|
|
|
|
|
119 |
|
120 |
except Exception as e:
|
121 |
return f"Error processing image: {str(e)}"
|
@@ -141,13 +131,12 @@ class XylariaChat:
|
|
141 |
except Exception as e:
|
142 |
return f"Error during Math OCR: {e}"
|
143 |
|
144 |
-
def get_response(self, user_input,
|
145 |
"""
|
146 |
Generate a response using chat completions with improved error handling
|
147 |
Args:
|
148 |
user_input (str): User's message
|
149 |
-
|
150 |
-
math_ocr_image (str, optional): Path to math OCR image
|
151 |
Returns:
|
152 |
Stream of chat completions or error message
|
153 |
"""
|
@@ -178,25 +167,15 @@ class XylariaChat:
|
|
178 |
content=msg['content']
|
179 |
).to_dict())
|
180 |
|
181 |
-
# Process
|
182 |
-
|
183 |
-
|
184 |
-
|
185 |
-
image_context += f"Uploaded images: {image_caption}\n\n"
|
186 |
-
|
187 |
-
# Process math OCR image if uploaded
|
188 |
-
if math_ocr_image:
|
189 |
-
ocr_text = self.perform_math_ocr(math_ocr_image)
|
190 |
-
if not ocr_text.startswith("Error"):
|
191 |
-
image_context += f"Math OCR Result: {ocr_text}\n\n"
|
192 |
-
|
193 |
-
# Combine image context with user input
|
194 |
-
full_input = image_context + user_input
|
195 |
|
196 |
# Add user input
|
197 |
messages.append(ChatMessage(
|
198 |
role="user",
|
199 |
-
content=
|
200 |
).to_dict())
|
201 |
|
202 |
# Calculate available tokens
|
@@ -242,32 +221,31 @@ class XylariaChat:
|
|
242 |
|
243 |
|
244 |
def create_interface(self):
|
245 |
-
def
|
246 |
-
|
247 |
-
|
248 |
-
|
249 |
-
|
250 |
-
|
251 |
-
|
252 |
-
|
253 |
-
|
254 |
-
|
255 |
-
|
256 |
-
|
257 |
-
return None
|
258 |
-
|
259 |
-
def streaming_response(message, chat_history, image1, image2, image3, image4, image5, math_ocr_image_path):
|
260 |
-
# Collect non-None images
|
261 |
-
images = [img for img in [image1, image2, image3, image4, image5] if img is not None]
|
262 |
|
263 |
-
#
|
264 |
-
|
|
|
|
|
|
|
|
|
265 |
|
266 |
# Handle errors in get_response
|
267 |
if isinstance(response_stream, str):
|
268 |
# Return immediately with the error message
|
269 |
updated_history = chat_history + [[message, response_stream]]
|
270 |
-
yield
|
271 |
return
|
272 |
|
273 |
# Prepare for streaming response
|
@@ -283,12 +261,12 @@ class XylariaChat:
|
|
283 |
|
284 |
# Update the last message in chat history with partial response
|
285 |
updated_history[-1][1] = full_response
|
286 |
-
yield
|
287 |
except Exception as e:
|
288 |
print(f"Streaming error: {e}")
|
289 |
# Display error in the chat interface
|
290 |
updated_history[-1][1] = f"Error during response: {e}"
|
291 |
-
yield
|
292 |
return
|
293 |
|
294 |
# Update conversation history
|
@@ -303,9 +281,6 @@ class XylariaChat:
|
|
303 |
if len(self.conversation_history) > 10:
|
304 |
self.conversation_history = self.conversation_history[-10:]
|
305 |
|
306 |
-
# Reset image inputs after processing
|
307 |
-
yield ("", updated_history, None, None, None, None, None, None)
|
308 |
-
|
309 |
# Custom CSS for Inter font and improved styling
|
310 |
custom_css = """
|
311 |
@import url('https://fonts.googleapis.com/css2?family=Inter:wght@300;400;500;600;700&display=swap');
|
@@ -320,6 +295,38 @@ class XylariaChat:
|
|
320 |
.gradio-container button {
|
321 |
font-family: 'Inter', sans-serif !important;
|
322 |
}
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
323 |
"""
|
324 |
|
325 |
with gr.Blocks(theme='soft', css=custom_css) as demo:
|
@@ -331,6 +338,29 @@ class XylariaChat:
|
|
331 |
show_copy_button=True,
|
332 |
)
|
333 |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
334 |
# Input row with improved layout
|
335 |
with gr.Row():
|
336 |
with gr.Column(scale=4):
|
@@ -339,102 +369,70 @@ class XylariaChat:
|
|
339 |
placeholder="Type your message...",
|
340 |
container=False
|
341 |
)
|
342 |
-
|
343 |
-
# Image and Math upload buttons
|
344 |
-
with gr.Column(scale=1):
|
345 |
-
# Buttons for image and math uploads with symbolic icons
|
346 |
-
with gr.Row():
|
347 |
-
img_upload_btn = gr.Button("🖼️") # Image upload button
|
348 |
-
math_upload_btn = gr.Button("➗") # Math upload button
|
349 |
-
clipboard_btn = gr.Button("📋") # Clipboard paste button
|
350 |
-
|
351 |
-
# Multiple image inputs
|
352 |
-
with gr.Accordion("Images", open=False):
|
353 |
-
with gr.Column():
|
354 |
-
with gr.Row():
|
355 |
-
img1 = gr.Image(
|
356 |
-
sources=["upload", "webcam"],
|
357 |
-
type="filepath",
|
358 |
-
label="Image 1",
|
359 |
-
height=200
|
360 |
-
)
|
361 |
-
img2 = gr.Image(
|
362 |
-
sources=["upload", "webcam"],
|
363 |
-
type="filepath",
|
364 |
-
label="Image 2",
|
365 |
-
height=200
|
366 |
-
)
|
367 |
-
with gr.Row():
|
368 |
-
img3 = gr.Image(
|
369 |
-
sources=["upload", "webcam"],
|
370 |
-
type="filepath",
|
371 |
-
label="Image 3",
|
372 |
-
height=200
|
373 |
-
)
|
374 |
-
img4 = gr.Image(
|
375 |
-
sources=["upload", "webcam"],
|
376 |
-
type="filepath",
|
377 |
-
label="Image 4",
|
378 |
-
height=200
|
379 |
-
)
|
380 |
-
img5 = gr.Image(
|
381 |
-
sources=["upload", "webcam"],
|
382 |
-
type="filepath",
|
383 |
-
label="Image 5",
|
384 |
-
height=200
|
385 |
-
)
|
386 |
-
|
387 |
-
# Math OCR Image Upload
|
388 |
-
with gr.Accordion("Math Input", open=False):
|
389 |
-
math_ocr_img = gr.Image(
|
390 |
-
sources=["upload", "webcam"],
|
391 |
-
type="filepath",
|
392 |
-
label="Upload Image for math",
|
393 |
-
height=200
|
394 |
-
)
|
395 |
|
396 |
# Clear history and memory buttons
|
397 |
with gr.Row():
|
398 |
clear = gr.Button("Clear Conversation")
|
399 |
clear_memory = gr.Button("Clear Memory")
|
400 |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
401 |
# Submit functionality with streaming and image support
|
402 |
-
btn = gr.Button("Send")
|
403 |
btn.click(
|
404 |
fn=streaming_response,
|
405 |
-
inputs=[txt, chatbot,
|
406 |
-
outputs=[txt, chatbot,
|
407 |
)
|
408 |
txt.submit(
|
409 |
fn=streaming_response,
|
410 |
-
inputs=[txt, chatbot,
|
411 |
-
outputs=[txt, chatbot,
|
412 |
)
|
413 |
|
414 |
-
#
|
415 |
-
clipboard_btn.click(
|
416 |
-
fn=get_clipboard_image,
|
417 |
-
outputs=[img1]
|
418 |
-
)
|
419 |
-
|
420 |
-
# Clear conversation button
|
421 |
clear.click(
|
422 |
-
fn=
|
423 |
inputs=None,
|
424 |
-
outputs=[chatbot,
|
|
|
425 |
)
|
426 |
|
427 |
-
# Clear memory
|
428 |
clear_memory.click(
|
429 |
-
fn=
|
430 |
inputs=None,
|
431 |
-
outputs=[]
|
|
|
432 |
)
|
433 |
|
|
|
|
|
|
|
434 |
return demo
|
435 |
|
436 |
-
#
|
437 |
-
|
438 |
chat = XylariaChat()
|
439 |
interface = chat.create_interface()
|
440 |
-
interface.launch(
|
|
|
|
|
|
|
|
|
|
|
|
|
|
5 |
from huggingface_hub import InferenceClient
|
6 |
from dataclasses import dataclass
|
7 |
import pytesseract
|
8 |
+
from PIL import Image
|
|
|
9 |
|
10 |
@dataclass
|
11 |
class ChatMessage:
|
|
|
30 |
api_key=self.hf_token
|
31 |
)
|
32 |
|
33 |
+
# Image captioning API setup
|
34 |
+
self.image_api_url = "https://api-inference.huggingface.co/models/Salesforce/blip-image-captioning-large"
|
35 |
self.image_api_headers = {"Authorization": f"Bearer {self.hf_token}"}
|
36 |
|
37 |
# Initialize conversation history and persistent memory
|
|
|
74 |
"""
|
75 |
Caption an uploaded image using Hugging Face API
|
76 |
Args:
|
77 |
+
image (str): Base64 encoded image or file path
|
78 |
Returns:
|
79 |
+
str: Image caption or error message
|
80 |
"""
|
81 |
try:
|
82 |
+
# If image is a file path, read and encode
|
83 |
+
if isinstance(image, str) and os.path.isfile(image):
|
84 |
+
with open(image, "rb") as f:
|
85 |
+
data = f.read()
|
86 |
+
# If image is already base64 encoded
|
87 |
+
elif isinstance(image, str):
|
88 |
+
# Remove data URI prefix if present
|
89 |
+
if image.startswith('data:image'):
|
90 |
+
image = image.split(',')[1]
|
91 |
+
data = base64.b64decode(image)
|
92 |
+
# If image is a file-like object (unlikely with Gradio, but good to have)
|
93 |
+
else:
|
94 |
+
data = image.read()
|
95 |
+
|
96 |
+
# Send request to Hugging Face API
|
97 |
+
response = requests.post(
|
98 |
+
self.image_api_url,
|
99 |
+
headers=self.image_api_headers,
|
100 |
+
data=data
|
101 |
+
)
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
102 |
|
103 |
+
# Check response
|
104 |
+
if response.status_code == 200:
|
105 |
+
caption = response.json()[0].get('generated_text', 'No caption generated')
|
106 |
+
return caption
|
107 |
+
else:
|
108 |
+
return f"Error captioning image: {response.status_code} - {response.text}"
|
109 |
|
110 |
except Exception as e:
|
111 |
return f"Error processing image: {str(e)}"
|
|
|
131 |
except Exception as e:
|
132 |
return f"Error during Math OCR: {e}"
|
133 |
|
134 |
+
def get_response(self, user_input, image=None):
|
135 |
"""
|
136 |
Generate a response using chat completions with improved error handling
|
137 |
Args:
|
138 |
user_input (str): User's message
|
139 |
+
image (optional): Uploaded image
|
|
|
140 |
Returns:
|
141 |
Stream of chat completions or error message
|
142 |
"""
|
|
|
167 |
content=msg['content']
|
168 |
).to_dict())
|
169 |
|
170 |
+
# Process image if uploaded
|
171 |
+
if image:
|
172 |
+
image_caption = self.caption_image(image)
|
173 |
+
user_input = f"Uploaded image : {image_caption}\n\nUser's message: {user_input}"
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
174 |
|
175 |
# Add user input
|
176 |
messages.append(ChatMessage(
|
177 |
role="user",
|
178 |
+
content=user_input
|
179 |
).to_dict())
|
180 |
|
181 |
# Calculate available tokens
|
|
|
221 |
|
222 |
|
223 |
def create_interface(self):
|
224 |
+
def streaming_response(message, chat_history, image_filepath, math_ocr_image_path):
|
225 |
+
|
226 |
+
ocr_text = ""
|
227 |
+
if math_ocr_image_path:
|
228 |
+
ocr_text = self.perform_math_ocr(math_ocr_image_path)
|
229 |
+
if ocr_text.startswith("Error"):
|
230 |
+
# Handle OCR error
|
231 |
+
updated_history = chat_history + [[message, ocr_text]]
|
232 |
+
yield "", updated_history, None, None
|
233 |
+
return
|
234 |
+
else:
|
235 |
+
message = f"Math OCR Result: {ocr_text}\n\nUser's message: {message}"
|
|
|
|
|
|
|
|
|
|
|
236 |
|
237 |
+
# Check if an image was actually uploaded
|
238 |
+
if image_filepath:
|
239 |
+
response_stream = self.get_response(message, image_filepath)
|
240 |
+
else:
|
241 |
+
response_stream = self.get_response(message)
|
242 |
+
|
243 |
|
244 |
# Handle errors in get_response
|
245 |
if isinstance(response_stream, str):
|
246 |
# Return immediately with the error message
|
247 |
updated_history = chat_history + [[message, response_stream]]
|
248 |
+
yield "", updated_history, None, None
|
249 |
return
|
250 |
|
251 |
# Prepare for streaming response
|
|
|
261 |
|
262 |
# Update the last message in chat history with partial response
|
263 |
updated_history[-1][1] = full_response
|
264 |
+
yield "", updated_history, None, None
|
265 |
except Exception as e:
|
266 |
print(f"Streaming error: {e}")
|
267 |
# Display error in the chat interface
|
268 |
updated_history[-1][1] = f"Error during response: {e}"
|
269 |
+
yield "", updated_history, None, None
|
270 |
return
|
271 |
|
272 |
# Update conversation history
|
|
|
281 |
if len(self.conversation_history) > 10:
|
282 |
self.conversation_history = self.conversation_history[-10:]
|
283 |
|
|
|
|
|
|
|
284 |
# Custom CSS for Inter font and improved styling
|
285 |
custom_css = """
|
286 |
@import url('https://fonts.googleapis.com/css2?family=Inter:wght@300;400;500;600;700&display=swap');
|
|
|
295 |
.gradio-container button {
|
296 |
font-family: 'Inter', sans-serif !important;
|
297 |
}
|
298 |
+
/* Image Upload Styling */
|
299 |
+
.image-container {
|
300 |
+
border: 1px solid #ccc;
|
301 |
+
border-radius: 8px;
|
302 |
+
padding: 10px;
|
303 |
+
margin-bottom: 10px;
|
304 |
+
display: flex;
|
305 |
+
flex-direction: column;
|
306 |
+
align-items: center;
|
307 |
+
gap: 10px;
|
308 |
+
background-color: #f8f8f8;
|
309 |
+
}
|
310 |
+
.image-preview {
|
311 |
+
max-width: 200px;
|
312 |
+
max-height: 200px;
|
313 |
+
border-radius: 8px;
|
314 |
+
}
|
315 |
+
.image-buttons {
|
316 |
+
display: flex;
|
317 |
+
gap: 10px;
|
318 |
+
}
|
319 |
+
.image-buttons button {
|
320 |
+
padding: 8px 15px;
|
321 |
+
border-radius: 5px;
|
322 |
+
background-color: #4CAF50;
|
323 |
+
color: white;
|
324 |
+
border: none;
|
325 |
+
cursor: pointer;
|
326 |
+
}
|
327 |
+
.image-buttons button:hover {
|
328 |
+
background-color: #367c39;
|
329 |
+
}
|
330 |
"""
|
331 |
|
332 |
with gr.Blocks(theme='soft', css=custom_css) as demo:
|
|
|
338 |
show_copy_button=True,
|
339 |
)
|
340 |
|
341 |
+
# Enhanced Image Upload Section
|
342 |
+
with gr.Accordion("Image Input", open=False):
|
343 |
+
with gr.Column() as image_container: # Use a Column for the image container
|
344 |
+
img = gr.Image(
|
345 |
+
sources=["upload", "webcam"],
|
346 |
+
type="filepath",
|
347 |
+
label="", # Remove label as it's redundant
|
348 |
+
elem_classes="image-preview", # Add a class for styling
|
349 |
+
)
|
350 |
+
with gr.Row():
|
351 |
+
clear_image_btn = gr.Button("Clear Image")
|
352 |
+
|
353 |
+
with gr.Accordion("Math Input", open=False):
|
354 |
+
with gr.Column():
|
355 |
+
math_ocr_img = gr.Image(
|
356 |
+
sources=["upload", "webcam"],
|
357 |
+
type="filepath",
|
358 |
+
label="Upload Image for math",
|
359 |
+
elem_classes="image-preview"
|
360 |
+
)
|
361 |
+
with gr.Row():
|
362 |
+
clear_math_ocr_btn = gr.Button("Clear Math Image")
|
363 |
+
|
364 |
# Input row with improved layout
|
365 |
with gr.Row():
|
366 |
with gr.Column(scale=4):
|
|
|
369 |
placeholder="Type your message...",
|
370 |
container=False
|
371 |
)
|
372 |
+
btn = gr.Button("Send", scale=1)
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
373 |
|
374 |
# Clear history and memory buttons
|
375 |
with gr.Row():
|
376 |
clear = gr.Button("Clear Conversation")
|
377 |
clear_memory = gr.Button("Clear Memory")
|
378 |
|
379 |
+
# Clear image functionality
|
380 |
+
clear_image_btn.click(
|
381 |
+
fn=lambda: None,
|
382 |
+
inputs=None,
|
383 |
+
outputs=[img],
|
384 |
+
queue=False
|
385 |
+
)
|
386 |
+
|
387 |
+
# Clear Math OCR image functionality
|
388 |
+
clear_math_ocr_btn.click(
|
389 |
+
fn=lambda: None,
|
390 |
+
inputs=None,
|
391 |
+
outputs=[math_ocr_img],
|
392 |
+
queue=False
|
393 |
+
)
|
394 |
+
|
395 |
# Submit functionality with streaming and image support
|
|
|
396 |
btn.click(
|
397 |
fn=streaming_response,
|
398 |
+
inputs=[txt, chatbot, img, math_ocr_img],
|
399 |
+
outputs=[txt, chatbot, img, math_ocr_img]
|
400 |
)
|
401 |
txt.submit(
|
402 |
fn=streaming_response,
|
403 |
+
inputs=[txt, chatbot, img, math_ocr_img],
|
404 |
+
outputs=[txt, chatbot, img, math_ocr_img]
|
405 |
)
|
406 |
|
407 |
+
# Clear conversation history
|
|
|
|
|
|
|
|
|
|
|
|
|
408 |
clear.click(
|
409 |
+
fn=lambda: None,
|
410 |
inputs=None,
|
411 |
+
outputs=[chatbot],
|
412 |
+
queue=False
|
413 |
)
|
414 |
|
415 |
+
# Clear persistent memory and reset conversation
|
416 |
clear_memory.click(
|
417 |
+
fn=self.reset_conversation,
|
418 |
inputs=None,
|
419 |
+
outputs=[chatbot],
|
420 |
+
queue=False
|
421 |
)
|
422 |
|
423 |
+
# Ensure memory is cleared when the interface is closed
|
424 |
+
demo.load(self.reset_conversation, None, None)
|
425 |
+
|
426 |
return demo
|
427 |
|
428 |
+
# Launch the interface
|
429 |
+
def main():
|
430 |
chat = XylariaChat()
|
431 |
interface = chat.create_interface()
|
432 |
+
interface.launch(
|
433 |
+
share=True, # Optional: create a public link
|
434 |
+
debug=True # Show detailed errors
|
435 |
+
)
|
436 |
+
|
437 |
+
if __name__ == "__main__":
|
438 |
+
main()
|