File size: 4,030 Bytes
db806d0 |
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65 66 67 68 69 70 71 72 73 74 75 76 77 78 79 80 81 82 83 84 85 86 87 88 89 90 91 92 93 94 95 96 97 98 99 100 101 102 103 104 105 106 107 108 109 110 111 112 113 114 115 |
import streamlit as st
from transformers import pipeline
import fitz # PyMuPDF for PDF handling
import re
import tempfile
# Function to clean extracted text
def clean_text(text):
# Replace multiple spaces or newlines with a single space
text = re.sub(r'\s+', ' ', text)
return text.strip()
# Function to extract text from PDF and clean it
def extract_text_from_pdf(uploaded_file):
with tempfile.NamedTemporaryFile(delete=False) as tmp_file:
tmp_file.write(uploaded_file.read())
tmp_file_path = tmp_file.name
doc = fitz.open(tmp_file_path)
text = ""
for page in doc:
text += page.get_text()
doc.close()
return clean_text(text)
# Function to chunk large text for context
def chunk_text(text, max_length=1000):
# Split the text into chunks with a maximum character length
chunks = [text[i:i+max_length] for i in range(0, len(text), max_length)]
return chunks
# Initialize Hugging Face model pipeline
@st.cache_resource
def load_model():
return pipeline("text-generation", model="gpt2", clean_up_tokenization_spaces=True) # Use GPT-2 for simplicity
model_pipeline = load_model()
# Generate a response
# Generate a response
def get_response(prompt, context):
combined_prompt = f"Context: {context}\n\nQuestion: {prompt}\nAnswer:"
response = model_pipeline(
combined_prompt,
max_new_tokens=150, # Ensure concise answers
num_return_sequences=1
)
# Extract only the answer part after the "Answer:" in the generated response
raw_response = response[0]["generated_text"]
answer_start = raw_response.find("Answer:") + len("Answer:")
answer = raw_response[answer_start:].strip()
return clean_text(answer)
# Streamlit App UI
st.title("Chat with PDF!!!")
# Sidebar for description
st.sidebar.title("Instructions")
st.sidebar.markdown("""
### How to Use This Application:
1. **Upload a PDF File**:
Use the file uploader to select and upload the PDF file you wish to analyze. The file should be in `.pdf` format.
2. **Text Extraction**:
The application extracts the text from the uploaded PDF using the `PyMuPDF` library (imported as `fitz`).
3. **Text Chunking**:
Large documents are divided into smaller chunks for better processing. Each chunk contains up to 1000 characters.
4. **Ask Questions**:
After the text is processed, type your question about the document.
5. **Model Response**:
The app sends the prompt and the relevant chunk to the model, which generates a response based on the content.
6. **Receive Insights**:
Get detailed insights and answers related to the PDF content.
""")
uploaded_file = st.file_uploader("Choose a PDF file", type=["pdf"])
if uploaded_file is not None:
# Extract and clean text from the uploaded PDF
pdf_text = extract_text_from_pdf(uploaded_file)
# Chunk the extracted text
text_chunks = chunk_text(pdf_text)
# Display the first chunk as a summary
st.subheader("PDF Content Summary:")
st.write(text_chunks[0]) # Display first chunk
# Input for user prompt
prompt = st.text_area(label="Ask a question based on the PDF content")
button = st.button("Ok")
if button:
if prompt:
# Select relevant chunk based on the question
relevant_chunk = None
for chunk in text_chunks:
if any(keyword.lower() in chunk.lower() for keyword in prompt.split()):
relevant_chunk = chunk
break
# If no relevant chunk was found, use the first chunk as a fallback
if not relevant_chunk:
relevant_chunk = text_chunks[0]
# Get response from the model
response = get_response(prompt, relevant_chunk)
st.markdown(f"**Answer:** {response}")
|