app.py

# ---------------------------------------------------------------------------------------
# Imports and Options
# ---------------------------------------------------------------------------------------
import streamlit as st
import pandas as pd
import requests
import re
import fitz  # PyMuPDF
import io
import matplotlib.pyplot as plt
from PIL import Image
from transformers import AutoProcessor, AutoModelForVision2Seq
from docling_core.types.doc import DoclingDocument
from docling_core.types.doc.document import DocTagsDocument
import torch

import logging
logging.basicConfig(level=logging.INFO)

# ---------------------------------------------------------------------------------------
# API Configuration
# ---------------------------------------------------------------------------------------
API_URL = "https://api.stack-ai.com/inference/v0/run/2df89a6c-a4af-4576-880e-27058e498f02/67acad8b0603ba4631db38e7"
headers = {
    'Authorization': 'Bearer a9e4979e-cdbe-49ea-a193-53562a784805',
    'Content-Type': 'application/json'
}

# ---------------------------------------------------------------------------------------
# Survey Analysis Class
# ---------------------------------------------------------------------------------------
class SurveyAnalysis:
    def __init__(self, api_key=None):
        self.api_key = api_key

    def prepare_llm_input(self, survey_response, topics):
        # Create topic description string from user input
        topic_descriptions = "\n".join([f"- **{topic}**: {description}" for topic, description in topics.items()])
        
        llm_input = f"""
Your task is to review PDF docling and extract information related to the provided topics. Here are the topic descriptions:

{topic_descriptions}

**Instructions:**
- Extract and summarize the PDF focusing only on the provided topics.
- If a topic is not mentioned in the notes, it should not be included in the Topic_Summary.
- Use **exact quotes** from the original text for each point in your Topic_Summary.
- Exclude erroneous content.
- Do not add additional explanations or instructions.

**Format your response as follows:**
[Topic]
- "Exact quote"
- "Exact quote"
- "Exact quote"

**Meeting Notes:**
{survey_response}
"""
        return llm_input

    def query_api(self, payload):
        response = requests.post(API_URL, headers=headers, json=payload)
        return response.json()

    def extract_meeting_notes(self, response):
        output = response.get('outputs', {}).get('out-0', '')
        return output

    def process_dataframe(self, df, topics):
        results = []
        for _, row in df.iterrows():
            llm_input = self.prepare_llm_input(row['Document_Text'], topics)
            payload = {
                "user_id": "<USER or Conversation ID>",  
                "in-0": llm_input
            }
            response = self.query_api(payload)
            meeting_notes = self.extract_meeting_notes(response)
            results.append({
                'Document_Text': row['Document_Text'],
                'Topic_Summary': meeting_notes
            })
        
        result_df = pd.DataFrame(results)
        df = df.reset_index(drop=True)
        return pd.concat([df, result_df[['Topic_Summary']]], axis=1)

# ---------------------------------------------------------------------------------------
# Function to Extract Excerpts
# ---------------------------------------------------------------------------------------
def extract_excerpts(processed_df):
    new_rows = []
    
    for _, row in processed_df.iterrows():
        Topic_Summary = row['Topic_Summary']
        
        # Split the Topic_Summary by topic
        sections = re.split(r'\n(?=\[)', Topic_Summary)
        
        for section in sections:
            # Extract the topic
            topic_match = re.match(r'\[([^\]]+)\]', section)
            if topic_match:
                topic = topic_match.group(1)
                
                # Extract all excerpts within the section
                excerpts = re.findall(r'- "([^"]+)"', section)
                
                for excerpt in excerpts:
                    new_rows.append({
                        'Document_Text': row['Document_Text'],
                        'Topic_Summary': row['Topic_Summary'],
                        'Excerpt': excerpt,
                        'Topic': topic
                    })
    
    return pd.DataFrame(new_rows)

#------------------------------------------------------------------------
# Streamlit Configuration
#------------------------------------------------------------------------

# Set page configuration
st.set_page_config(
    page_title="Choose Your Own Adventure (Topic Extraction) PDF Analysis App", 
    page_icon=":bar_chart:", 
    layout="centered", 
    initial_sidebar_state="auto",
    menu_items={
        'Get Help': 'mailto:[email protected]',
        'About': "This app is built to support PDF analysis"
    }
)

#------------------------------------------------------------------------
# Sidebar
#------------------------------------------------------------------------

# Sidebar with image
with st.sidebar:
    # Set the desired width in pixels
    image_width = 300  
    # Define the path to the image
    # image_path = "steelcase_small.png"
    image_path = "mtss.ai_small.png"
    # Display the image
    st.image(image_path, width=image_width)

    # Additional sidebar content

    with st.expander("**MTSS.ai**", expanded=True):
        st.write("""
        - **Support**: Cheyne LeVesseur PhD  
        - **Email**: [email protected]
        """)
    st.divider()
    st.subheader('Instructions')

    Instructions = """
    - **Step 1**: Upload your PDF file.
    - **Step 2**: Review the processed text.
    - **Step 3**: Add your topics and descriptions of interest.
    - **Step 4**: Review the extracted excerpts and classifications, and topic distribution and frequency.
    - **Step 5**: Review bar charts of topics.
    - **Step 6**: Download the processed data as a CSV file.
    """
    st.markdown(Instructions)

# Load SmolDocling model using transformers
@st.cache_resource
def load_smol_docling():
    device = "cuda" if torch.cuda.is_available() else "cpu"
    processor = AutoProcessor.from_pretrained("ds4sd/SmolDocling-256M-preview")
    model = AutoModelForVision2Seq.from_pretrained(
        "ds4sd/SmolDocling-256M-preview",
        torch_dtype=torch.float32
    ).to(device)
    return model, processor

model, processor = load_smol_docling()

# # Convert PDF to images
# def convert_pdf_to_images(pdf_file):
#     images = []
#     doc = fitz.open(stream=pdf_file.read(), filetype="pdf")
#     for page_number in range(len(doc)):
#         page = doc.load_page(page_number)
#         pix = page.get_pixmap(dpi=300)  # Higher DPI for clarity
#         img_data = pix.tobytes("png")
#         image = Image.open(io.BytesIO(img_data))
#         images.append(image)
#     return images

# Improved PDF to image conversion
def convert_pdf_to_images(pdf_file, dpi=150, max_size=1600):
    images = []
    doc = fitz.open(stream=pdf_file.read(), filetype="pdf")
    for page_number in range(len(doc)):
        page = doc.load_page(page_number)
        pix = page.get_pixmap(dpi=dpi)
        img_data = pix.tobytes("png")
        image = Image.open(io.BytesIO(img_data)).convert("RGB")
        # Resize image to max dimension
        image.thumbnail((max_size, max_size), Image.LANCZOS)
        images.append(image)
    return images

# Extract structured markdown text using SmolDocling (transformers)
# def extract_markdown_from_image(image):
#     prompt_text = "Convert this page to docling."
#     device = "cuda" if torch.cuda.is_available() else "cpu"

#     # Prepare inputs
#     messages = [
#         {
#             "role": "user",
#             "content": [
#                 {"type": "image"},
#                 {"type": "text", "text": prompt_text}
#             ]
#         }
#     ]
#     prompt = processor.apply_chat_template(messages, add_generation_prompt=True)
#     inputs = processor(text=prompt, images=[image], return_tensors="pt").to(device)
    
#     # Generate outputs
#     generated_ids = model.generate(**inputs, max_new_tokens=1024)
#     prompt_length = inputs.input_ids.shape[1]
#     trimmed_generated_ids = generated_ids[:, prompt_length:]
#     doctags = processor.batch_decode(trimmed_generated_ids, skip_special_tokens=False)[0].lstrip()

#     # Clean the output
#     doctags = doctags.replace("<end_of_utterance>", "").strip()

#     # Populate document
#     doctags_doc = DocTagsDocument.from_doctags_and_image_pairs([doctags], [image])

#     # Create a docling document
#     doc = DoclingDocument(name="ExtractedDocument")
#     doc.load_from_doctags(doctags_doc)

#     # Export as markdown
#     markdown_text = doc.export_to_markdown()
#     return markdown_text

def extract_markdown_from_image(image):
    start_time = time.time()
    prompt_text = "Convert this page to docling."
    device = "cuda" if torch.cuda.is_available() else "cpu"

    messages = [{"role": "user", "content": [{"type": "image"}, {"type": "text", "text": prompt_text}]}]
    prompt = processor.apply_chat_template(messages, add_generation_prompt=True)
    inputs = processor(text=prompt, images=[image], return_tensors="pt").to(device)

    with torch.no_grad():  # <-- Crucial for speed
        generated_ids = model.generate(**inputs, max_new_tokens=1024)

    prompt_length = inputs.input_ids.shape[1]
    trimmed_generated_ids = generated_ids[:, prompt_length:]
    doctags = processor.batch_decode(trimmed_generated_ids, skip_special_tokens=False)[0].lstrip()
    doctags = doctags.replace("<end_of_utterance>", "").strip()

    doctags_doc = DocTagsDocument.from_doctags_and_image_pairs([doctags], [image])
    doc = DoclingDocument(name="ExtractedDocument")
    doc.load_from_doctags(doctags_doc)
    markdown_text = doc.export_to_markdown()
    processing_time = time.time() - start_time
    logging.info(f"Inference took {processing_time:.2f} seconds")
    return markdown_text

# Streamlit UI
st.title("Choose Your Own Adventure (Topic Extraction) PDF Analysis App")

uploaded_file = st.file_uploader("Upload PDF file", type=["pdf"])

if uploaded_file:
    with st.spinner("Processing PDF..."):
        images = convert_pdf_to_images(uploaded_file)

        markdown_texts = []
        for idx, image in enumerate(images):
            markdown_text = extract_markdown_from_image(image)
            markdown_texts.append(markdown_text)

        df = pd.DataFrame({'Document_Text': markdown_texts})

    st.success("PDF processed successfully!")

    # Check if extraction was successful
    if df.empty or df['Document_Text'].isnull().all():
        st.error("No meaningful text extracted from the PDF.")
        st.stop()

    st.markdown("### Extracted Markdown Preview")
    st.write(df.head())

    # ---------------------------------------------------------------------------------------
    # User Input for Topics
    # ---------------------------------------------------------------------------------------
    st.markdown("### Enter Topics and Descriptions")
    num_topics = st.number_input("Number of topics", min_value=1, max_value=10, value=1, step=1)

    topics = {}
    for i in range(num_topics):
        topic = st.text_input(f"Topic {i+1} Name", key=f"topic_{i}")
        description = st.text_area(f"Topic {i+1} Description", key=f"description_{i}")
        if topic and description:
            topics[topic] = description

    # Add a button to execute the analysis
    if st.button("Run Analysis"):
        if not topics:
            st.warning("Please enter at least one topic and description.")
            st.stop()

        # ---------------------------------------------------------------------------------------
        # Your existing SurveyAnalysis and extract_excerpts functions remain unchanged here:
        # ---------------------------------------------------------------------------------------
        analyzer = SurveyAnalysis()
        processed_df = analyzer.process_dataframe(df, topics)
        df_VIP_extracted = extract_excerpts(processed_df)

        required_columns = ['Document_Text', 'Topic_Summary', 'Excerpt', 'Topic']
        missing_columns = [col for col in required_columns if col not in df_VIP_extracted.columns]

        if missing_columns:
            st.error(f"Missing columns after processing: {missing_columns}")
            st.stop()

        df_VIP_extracted = df_VIP_extracted[required_columns]

        st.markdown("### Processed Meeting Notes")
        st.dataframe(df_VIP_extracted)

        st.write(f"**Number of meeting notes analyzed:** {len(df)}")
        st.write(f"**Number of excerpts extracted:** {len(df_VIP_extracted)}")

        # CSV download
        csv = df_VIP_extracted.to_csv(index=False)
        st.download_button(
            "Download data as CSV",
            data=csv,
            file_name='extracted_meeting_notes.csv',
            mime='text/csv'
        )

        # Topic distribution visualization
        topic_counts = df_VIP_extracted['Topic'].value_counts()
        frequency_table = pd.DataFrame({'Topic': topic_counts.index, 'Count': topic_counts.values})
        frequency_table['Percentage'] = (frequency_table['Count'] / frequency_table['Count'].sum() * 100).round(0)

        st.markdown("### Topic Distribution")
        st.dataframe(frequency_table)

        fig, ax = plt.subplots(figsize=(10, 5))
        ax.bar(frequency_table['Topic'], frequency_table['Count'], color='#3d9aa1')
        ax.set_ylabel('Count')
        ax.set_title('Frequency of Topics')
        st.pyplot(fig)

else:
    st.info("Please upload a PDF file to begin.")