Spaces:
Paused
Paused
Update app.py
Browse files
app.py
CHANGED
|
@@ -59,18 +59,197 @@ def chat_with_pdf(api_key, pdf_text, user_question, history):
|
|
| 59 |
else:
|
| 60 |
pdf_context = pdf_text
|
| 61 |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 62 |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 63 |
|
|
|
|
|
|
|
| 64 |
|
|
|
|
|
|
|
| 65 |
|
|
|
|
| 66 |
|
|
|
|
|
|
|
| 67 |
|
|
|
|
|
|
|
| 68 |
|
|
|
|
|
|
|
|
|
|
| 69 |
|
|
|
|
|
|
|
| 70 |
|
|
|
|
|
|
|
| 71 |
|
|
|
|
|
|
|
| 72 |
|
|
|
|
| 73 |
|
|
|
|
|
|
|
| 74 |
|
| 75 |
|
| 76 |
-
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 59 |
else:
|
| 60 |
pdf_context = pdf_text
|
| 61 |
|
| 62 |
+
|
| 63 |
+
system_message = f"""You are an intelligent assistant designed to read, understand, and extract information from PDF documents.
|
| 64 |
+
Based on any question or query the user asks—whether it's about content, summaries, data extraction, definitions, insights, or interpretation—you will
|
| 65 |
+
analyze the following PDF content and provide an accurate, helpful response grounded in the document. Always respond with clear, concise, and context-aware information.
|
| 66 |
+
PDF CONTENT:
|
| 67 |
+
{pdf_context}
|
| 68 |
+
Answer the user's questions only based on the PDF content above. If the answer cannot be found in the PDF, politely state that the information is not available in the provided document."""
|
| 69 |
+
|
| 70 |
+
messages = [
|
| 71 |
+
{"role": "system", "content": system_message},
|
| 72 |
+
]
|
| 73 |
+
|
| 74 |
+
# Add chat history
|
| 75 |
+
for h_user, h_bot in history:
|
| 76 |
+
messages.append({"role": "user", "content": h_user}) #conv1, .....
|
| 77 |
+
messages.append({"role": "assistant", "content": h_bot}) #conv1res, ....
|
| 78 |
|
| 79 |
+
messages.append({"role": "user", "content": user_question}) #convx ----- History (context)
|
| 80 |
+
|
| 81 |
+
response=client.chat.completions.create(
|
| 82 |
+
model="meta-llama/Llama-3.3-70B-Instruct-Turbo-Free",
|
| 83 |
+
messages=messages,
|
| 84 |
+
max_tokens=5000, #5000
|
| 85 |
+
temperature=0.7,
|
| 86 |
+
)
|
| 87 |
+
|
| 88 |
+
# Extract the assistant's response
|
| 89 |
+
assistant_response = response.choices[0].message.content ## convx? Ans: convxres (error: ans (something is wrong: model out of limit))
|
| 90 |
+
|
| 91 |
+
new_history = history + [(user_question, assistant_response)]
|
| 92 |
+
|
| 93 |
+
return new_history, new_history
|
| 94 |
+
|
| 95 |
+
except Exception as e:
|
| 96 |
+
error_message = f"Error: {str(e)}"
|
| 97 |
+
return history + [(user_question, error_message)], history ### history + current question + error message
|
| 98 |
|
| 99 |
+
|
| 100 |
+
def process_pdf(pdf_file, api_key_input):
|
| 101 |
|
| 102 |
+
if pdf_file is None:
|
| 103 |
+
return "Please upload a PDF file.", "", []
|
| 104 |
|
| 105 |
+
try:
|
| 106 |
|
| 107 |
+
# Get the file name
|
| 108 |
+
file_name = os.path.basename(pdf_file.name) if hasattr(pdf_file, 'name') else "Upload PDF"
|
| 109 |
|
| 110 |
+
# Extract text from the PDF
|
| 111 |
+
pdf_text = extract_text_from_pdf(pdf_file)
|
| 112 |
|
| 113 |
+
# Check if there was an error in extraction
|
| 114 |
+
if pdf_text.startswith("Error extracting text from PDF"):
|
| 115 |
+
return f"❌ {pdf_text}", "", []
|
| 116 |
|
| 117 |
+
if not pdf_text.strip() or pdf_text.startswith("No text could be extracted"):
|
| 118 |
+
return f"⚠️ {pdf_text}", "", []
|
| 119 |
|
| 120 |
+
# Count words for information
|
| 121 |
+
word_count = len(pdf_text.split())
|
| 122 |
|
| 123 |
+
# Return a message with the file name and text content
|
| 124 |
+
status_message = f"✅ Successfully processed PDF: {file_name} ({word_count} words extracted)"
|
| 125 |
|
| 126 |
+
return status_message, pdf_text, []
|
| 127 |
|
| 128 |
+
except Exception as e:
|
| 129 |
+
return f"❌ Error processing PDF: {str(e)}", "", []
|
| 130 |
|
| 131 |
|
| 132 |
+
def validate_api_key(api_key):
|
| 133 |
+
"""Simple validation for API key format"""
|
| 134 |
+
if not api_key or not api_key.strip():
|
| 135 |
+
return "❌ API Key is required"
|
| 136 |
+
|
| 137 |
+
if len(api_key.strip()) < 10:
|
| 138 |
+
return "❌ API Key appears to be too short"
|
| 139 |
+
|
| 140 |
+
return "✓ API Key format looks valid (not verified with server)"
|
| 141 |
+
|
| 142 |
+
###UI
|
| 143 |
+
|
| 144 |
+
with gr.Blocks(title="ChatPDF with Together AI", theme=gr.themes.Ocean()) as app:
|
| 145 |
+
gr.Markdown("# 📄 ChatPDF with Together AI")
|
| 146 |
+
gr.Markdown("Upload a PDF and chat with it using the Llama-3.3-70B model.")
|
| 147 |
+
|
| 148 |
+
with gr.Row():
|
| 149 |
+
with gr.Column(scale=1):
|
| 150 |
+
# API Key input
|
| 151 |
+
api_key_input = gr.Textbox(
|
| 152 |
+
label="Together API Key",
|
| 153 |
+
placeholder="Enter your Together API key here...",
|
| 154 |
+
type="password"
|
| 155 |
+
)
|
| 156 |
+
|
| 157 |
+
# API key validation
|
| 158 |
+
api_key_status = gr.Textbox(
|
| 159 |
+
label="API Key Status",
|
| 160 |
+
interactive=False
|
| 161 |
+
)
|
| 162 |
+
|
| 163 |
+
# PDF upload
|
| 164 |
+
pdf_file = gr.File(
|
| 165 |
+
label="Upload PDF",
|
| 166 |
+
file_types=[".pdf"],
|
| 167 |
+
type="binary" # Ensure we get binary data
|
| 168 |
+
)
|
| 169 |
+
|
| 170 |
+
# Process PDF button
|
| 171 |
+
process_button = gr.Button("Process PDF")
|
| 172 |
+
|
| 173 |
+
# Status message
|
| 174 |
+
status_message = gr.Textbox(
|
| 175 |
+
label="Status",
|
| 176 |
+
interactive=False
|
| 177 |
+
)
|
| 178 |
+
|
| 179 |
+
with gr.Accordion("PDF Content Preview", open=True):
|
| 180 |
+
pdf_preview = gr.Textbox(
|
| 181 |
+
label="Extracted Text Preview",
|
| 182 |
+
interactive=False,
|
| 183 |
+
max_lines=10,
|
| 184 |
+
show_copy_button=True
|
| 185 |
+
)
|
| 186 |
+
|
| 187 |
+
with gr.Column(scale=2):
|
| 188 |
+
# Chat interface
|
| 189 |
+
chatbot = gr.Chatbot(
|
| 190 |
+
label="Chat with PDF",
|
| 191 |
+
height=500,
|
| 192 |
+
show_copy_button=True
|
| 193 |
+
)
|
| 194 |
+
|
| 195 |
+
# Question input
|
| 196 |
+
question = gr.Textbox(
|
| 197 |
+
label="Ask a question about the PDF",
|
| 198 |
+
placeholder="What is the main topic of this document?",
|
| 199 |
+
lines=2
|
| 200 |
+
)
|
| 201 |
+
|
| 202 |
+
# Submit button
|
| 203 |
+
submit_button = gr.Button("Submit Question")
|
| 204 |
+
|
| 205 |
+
|
| 206 |
+
# Event handlers
|
| 207 |
+
def update_preview(text):
|
| 208 |
+
"""Update the preview with the first few lines of the PDF text"""
|
| 209 |
+
if not text or text.startswith("Error") or text.startswith("No text"):
|
| 210 |
+
return text
|
| 211 |
+
|
| 212 |
+
# Get the first ~500 characters for preview
|
| 213 |
+
preview = text[:500]
|
| 214 |
+
if len(text) > 500:
|
| 215 |
+
preview += "...\n[Text truncated for preview. Full text will be used for chat.]"
|
| 216 |
+
return preview
|
| 217 |
+
|
| 218 |
+
# API key validation event
|
| 219 |
+
api_key_input.change(
|
| 220 |
+
fn=validate_api_key,
|
| 221 |
+
inputs=[api_key_input],
|
| 222 |
+
outputs=[api_key_status]
|
| 223 |
+
)
|
| 224 |
+
|
| 225 |
+
process_button.click(
|
| 226 |
+
fn=process_pdf,
|
| 227 |
+
inputs=[pdf_file, api_key_input],
|
| 228 |
+
outputs=[status_message, pdf_text, chatbot]
|
| 229 |
+
).then(
|
| 230 |
+
fn=update_preview,
|
| 231 |
+
inputs=[pdf_text],
|
| 232 |
+
outputs=[pdf_preview]
|
| 233 |
+
)
|
| 234 |
+
|
| 235 |
+
submit_button.click(
|
| 236 |
+
fn=chat_with_pdf,
|
| 237 |
+
inputs=[api_key_input, pdf_text, question, chatbot],
|
| 238 |
+
outputs=[chatbot, chatbot]
|
| 239 |
+
).then(
|
| 240 |
+
fn=lambda: "",
|
| 241 |
+
outputs=question
|
| 242 |
+
)
|
| 243 |
+
|
| 244 |
+
question.submit(
|
| 245 |
+
fn=chat_with_pdf,
|
| 246 |
+
inputs=[api_key_input, pdf_text, question, chatbot],
|
| 247 |
+
outputs=[chatbot, chatbot]
|
| 248 |
+
).then(
|
| 249 |
+
fn=lambda: "",
|
| 250 |
+
outputs=question
|
| 251 |
+
)
|
| 252 |
+
|
| 253 |
+
# Launch the app
|
| 254 |
+
if __name__ == "__main__":
|
| 255 |
+
app.launch(share=True)
|