Spaces:
Sleeping
Sleeping
Update app.py
Browse files
app.py
CHANGED
|
@@ -39,88 +39,7 @@ MODELS = {
|
|
| 39 |
}
|
| 40 |
}
|
| 41 |
|
| 42 |
-
#
|
| 43 |
-
tab1, tab2 = st.tabs(["PDF Summarizer", "Invoice Extractor"])
|
| 44 |
-
|
| 45 |
-
# PDF Summarizer Tab
|
| 46 |
-
with tab1:
|
| 47 |
-
st.title("PDF to Bullet Point Summarizer π π")
|
| 48 |
-
|
| 49 |
-
# File uploader for the PDF
|
| 50 |
-
uploaded_file = st.file_uploader("Upload your PDF document", type="pdf", key="pdf_uploader")
|
| 51 |
-
|
| 52 |
-
# Slider for users to select the summarization extent
|
| 53 |
-
summary_scale = st.slider("Select the extent of summarization (%)", min_value=1, max_value=100, value=20, key="summary_scale")
|
| 54 |
-
|
| 55 |
-
# Submit button
|
| 56 |
-
submit_button = st.button("Generate Summary", key="summary_button")
|
| 57 |
-
|
| 58 |
-
# Check if the submit button is pressed
|
| 59 |
-
if submit_button and uploaded_file is not None:
|
| 60 |
-
with st.spinner('Processing...'):
|
| 61 |
-
# Read the PDF content
|
| 62 |
-
text = read_pdf(io.BytesIO(uploaded_file.getvalue()))
|
| 63 |
-
|
| 64 |
-
# Extract key phrases from the text
|
| 65 |
-
key_phrases = extract_key_phrases(text)
|
| 66 |
-
|
| 67 |
-
# Score sentences based on the key phrases
|
| 68 |
-
sentence_scores = score_sentences(text, key_phrases)
|
| 69 |
-
|
| 70 |
-
# Determine the number of bullet points based on the selected summarization scale
|
| 71 |
-
total_sentences = len(list(sentence_scores.keys()))
|
| 72 |
-
num_points = max(1, total_sentences * summary_scale // 100)
|
| 73 |
-
|
| 74 |
-
# Generate the bullet-point summary
|
| 75 |
-
summary = summarize_text(sentence_scores, num_points=num_points)
|
| 76 |
-
|
| 77 |
-
# Display the summary as bullet points
|
| 78 |
-
st.subheader("Here's the summary: ")
|
| 79 |
-
st.markdown(summary)
|
| 80 |
-
|
| 81 |
-
# Invoice Extractor Tab
|
| 82 |
-
with tab2:
|
| 83 |
-
st.title("π Invoice Extractor from PDF")
|
| 84 |
-
st.write("Upload an invoice PDF to extract key details")
|
| 85 |
-
|
| 86 |
-
# Model selection
|
| 87 |
-
model_choice = st.selectbox(
|
| 88 |
-
"Select AI Model",
|
| 89 |
-
list(MODELS.keys()),
|
| 90 |
-
index=0,
|
| 91 |
-
help="Choose which AI model to use for extraction",
|
| 92 |
-
key="model_choice"
|
| 93 |
-
)
|
| 94 |
-
|
| 95 |
-
# File uploader for the invoice PDF
|
| 96 |
-
invoice_pdf = st.file_uploader("Upload Invoice PDF", type="pdf", key="invoice_pdf_uploader")
|
| 97 |
-
|
| 98 |
-
if st.button("Extract Invoice Information", key="invoice_button") and invoice_pdf is not None:
|
| 99 |
-
with st.spinner('Reading PDF...'):
|
| 100 |
-
# Read the PDF content
|
| 101 |
-
invoice_text = read_pdf(io.BytesIO(invoice_pdf.getvalue()))
|
| 102 |
-
|
| 103 |
-
# Process in status container
|
| 104 |
-
with st.status("Processing...", expanded=True) as status:
|
| 105 |
-
st.write(f"π€ Querying {model_choice} API...")
|
| 106 |
-
invoice_data = extract_invoice_info(model_choice, invoice_text)
|
| 107 |
-
|
| 108 |
-
if invoice_data:
|
| 109 |
-
status.update(label="β
Extraction Complete!", state="complete")
|
| 110 |
-
display_invoice_data(model_choice, invoice_data)
|
| 111 |
-
else:
|
| 112 |
-
status.update(label="β Extraction Failed", state="error")
|
| 113 |
-
st.error("Failed to extract information. Try simplifying the text.")
|
| 114 |
-
|
| 115 |
-
# Debug information outside the status container
|
| 116 |
-
if invoice_data and "last_api_response" in st.session_state:
|
| 117 |
-
with st.expander("Debug Information"):
|
| 118 |
-
st.write("API Response:")
|
| 119 |
-
st.json(st.session_state.last_api_response)
|
| 120 |
-
st.write("Raw API Response:")
|
| 121 |
-
st.code(st.session_state.get("last_api_response_raw", "No response"))
|
| 122 |
-
|
| 123 |
-
# Invoice Extractor Functions
|
| 124 |
def get_api_key(model_choice):
|
| 125 |
"""Get the appropriate API key based on model choice"""
|
| 126 |
api_key_env = MODELS[model_choice]["api_key_env"]
|
|
@@ -438,4 +357,85 @@ def extract_invoice_info(model_choice, text):
|
|
| 438 |
if field not in item:
|
| 439 |
item[field] = None if field != "quantity" else 0
|
| 440 |
|
| 441 |
-
return parsed_data
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 39 |
}
|
| 40 |
}
|
| 41 |
|
| 42 |
+
# Define all invoice extraction functions first
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 43 |
def get_api_key(model_choice):
|
| 44 |
"""Get the appropriate API key based on model choice"""
|
| 45 |
api_key_env = MODELS[model_choice]["api_key_env"]
|
|
|
|
| 357 |
if field not in item:
|
| 358 |
item[field] = None if field != "quantity" else 0
|
| 359 |
|
| 360 |
+
return parsed_data
|
| 361 |
+
|
| 362 |
+
# Create tabs for different functionalities
|
| 363 |
+
tab1, tab2 = st.tabs(["PDF Summarizer", "Invoice Extractor"])
|
| 364 |
+
|
| 365 |
+
# PDF Summarizer Tab
|
| 366 |
+
with tab1:
|
| 367 |
+
st.title("PDF to Bullet Point Summarizer π π")
|
| 368 |
+
|
| 369 |
+
# File uploader for the PDF
|
| 370 |
+
uploaded_file = st.file_uploader("Upload your PDF document", type="pdf", key="pdf_uploader")
|
| 371 |
+
|
| 372 |
+
# Slider for users to select the summarization extent
|
| 373 |
+
summary_scale = st.slider("Select the extent of summarization (%)", min_value=1, max_value=100, value=20, key="summary_scale")
|
| 374 |
+
|
| 375 |
+
# Submit button
|
| 376 |
+
submit_button = st.button("Generate Summary", key="summary_button")
|
| 377 |
+
|
| 378 |
+
# Check if the submit button is pressed
|
| 379 |
+
if submit_button and uploaded_file is not None:
|
| 380 |
+
with st.spinner('Processing...'):
|
| 381 |
+
# Read the PDF content
|
| 382 |
+
text = read_pdf(io.BytesIO(uploaded_file.getvalue()))
|
| 383 |
+
|
| 384 |
+
# Extract key phrases from the text
|
| 385 |
+
key_phrases = extract_key_phrases(text)
|
| 386 |
+
|
| 387 |
+
# Score sentences based on the key phrases
|
| 388 |
+
sentence_scores = score_sentences(text, key_phrases)
|
| 389 |
+
|
| 390 |
+
# Determine the number of bullet points based on the selected summarization scale
|
| 391 |
+
total_sentences = len(list(sentence_scores.keys()))
|
| 392 |
+
num_points = max(1, total_sentences * summary_scale // 100)
|
| 393 |
+
|
| 394 |
+
# Generate the bullet-point summary
|
| 395 |
+
summary = summarize_text(sentence_scores, num_points=num_points)
|
| 396 |
+
|
| 397 |
+
# Display the summary as bullet points
|
| 398 |
+
st.subheader("Here's the summary: ")
|
| 399 |
+
st.markdown(summary)
|
| 400 |
+
|
| 401 |
+
# Invoice Extractor Tab
|
| 402 |
+
with tab2:
|
| 403 |
+
st.title("π Invoice Extractor from PDF")
|
| 404 |
+
st.write("Upload an invoice PDF to extract key details")
|
| 405 |
+
|
| 406 |
+
# Model selection
|
| 407 |
+
model_choice = st.selectbox(
|
| 408 |
+
"Select AI Model",
|
| 409 |
+
list(MODELS.keys()),
|
| 410 |
+
index=0,
|
| 411 |
+
help="Choose which AI model to use for extraction",
|
| 412 |
+
key="model_choice"
|
| 413 |
+
)
|
| 414 |
+
|
| 415 |
+
# File uploader for the invoice PDF
|
| 416 |
+
invoice_pdf = st.file_uploader("Upload Invoice PDF", type="pdf", key="invoice_pdf_uploader")
|
| 417 |
+
|
| 418 |
+
if st.button("Extract Invoice Information", key="invoice_button") and invoice_pdf is not None:
|
| 419 |
+
with st.spinner('Reading PDF...'):
|
| 420 |
+
# Read the PDF content
|
| 421 |
+
invoice_text = read_pdf(io.BytesIO(invoice_pdf.getvalue()))
|
| 422 |
+
|
| 423 |
+
# Process in status container
|
| 424 |
+
with st.status("Processing...", expanded=True) as status:
|
| 425 |
+
st.write(f"π€ Querying {model_choice} API...")
|
| 426 |
+
invoice_data = extract_invoice_info(model_choice, invoice_text)
|
| 427 |
+
|
| 428 |
+
if invoice_data:
|
| 429 |
+
status.update(label="β
Extraction Complete!", state="complete")
|
| 430 |
+
display_invoice_data(model_choice, invoice_data)
|
| 431 |
+
else:
|
| 432 |
+
status.update(label="β Extraction Failed", state="error")
|
| 433 |
+
st.error("Failed to extract information. Try simplifying the text.")
|
| 434 |
+
|
| 435 |
+
# Debug information outside the status container
|
| 436 |
+
if invoice_data and "last_api_response" in st.session_state:
|
| 437 |
+
with st.expander("Debug Information"):
|
| 438 |
+
st.write("API Response:")
|
| 439 |
+
st.json(st.session_state.last_api_response)
|
| 440 |
+
st.write("Raw API Response:")
|
| 441 |
+
st.code(st.session_state.get("last_api_response_raw", "No response"))
|