app.py

import streamlit as st
import base64
from reportlab.lib.pagesizes import A4
from reportlab.platypus import SimpleDocTemplate, Paragraph, Spacer, Table, TableStyle
from reportlab.lib.styles import getSampleStyleSheet, ParagraphStyle
from reportlab.lib import colors
import pikepdf
import fpdf
import fitz  # pymupdf
import cv2
import numpy as np
from PIL import Image
import io
import os
import re

# Define the ML outline as a markdown string for multilevel content
ml_markdown = """# Cutting-Edge ML Outline

## Core ML Techniques
1. 🌟 **Mixture of Experts (MoE)**
   - Conditional computation techniques
   - Sparse gating mechanisms
   - Training specialized sub-models

2. 🔥 **Supervised Fine-Tuning (SFT) using PyTorch**
   - Loss function customization
   - Gradient accumulation strategies
   - Learning rate schedulers

3. 🤖 **Large Language Models (LLM) using Transformers**
   - Attention mechanisms
   - Tokenization strategies
   - Position encodings

## Training Methods
4. 📊 **Self-Rewarding Learning using NPS 0-10 and Verbatims**
   - Custom reward functions
   - Feedback categorization
   - Signal extraction from text

5. 👍 **Reinforcement Learning from Human Feedback (RLHF)**
   - Preference datasets
   - PPO implementation
   - KL divergence constraints

6. 🔗 **MergeKit: Merging Models to Same Embedding Space**
   - TIES merging
   - Task arithmetic
   - SLERP interpolation

## Optimization & Deployment
7. 📏 **DistillKit: Model Size Reduction with Spectrum Analysis**
   - Knowledge distillation
   - Quantization techniques
   - Model pruning strategies

8. 🧠 **Agentic RAG Agents using Document Inputs**
   - Vector database integration
   - Query planning
   - Self-reflection mechanisms

9. ⏳ **Longitudinal Data Summarization from Multiple Docs**
   - Multi-document compression
   - Timeline extraction
   - Entity tracking

## Knowledge Representation
10. 📑 **Knowledge Extraction using Markdown Knowledge Graphs**
    - Entity recognition
    - Relationship mapping
    - Hierarchical structuring

11. 🗺️ **Knowledge Mapping with Mermaid Diagrams**
    - Flowchart generation
    - Sequence diagram creation
    - State diagrams

12. 💻 **ML Code Generation with Streamlit/Gradio/HTML5+JS**
    - Code completion
    - Unit test generation
    - Documentation synthesis
"""

# Process multilevel markdown for PDF output
def markdown_to_pdf_content(markdown_text):
    """Convert markdown text to a format suitable for PDF generation"""
    import re
    
    # Convert markdown headers to styled text for PDF
    lines = markdown_text.strip().split('\n')
    pdf_content = []
    in_list_item = False
    current_item = None
    sub_items = []
    
    for line in lines:
        line = line.strip()
        if not line:
            continue
            
        if line.startswith('# '):
            # Main header - will be handled separately in the PDF generation
            pass
        elif line.startswith('## '):
            # Section header - add as a bold item
            if current_item and sub_items:
                # Store the previous item with its sub-items
                pdf_content.append([current_item, sub_items])
                sub_items = []
                current_item = None
                
            section = line.replace('## ', '').strip()
            pdf_content.append(f"<b>{section}</b>")
            in_list_item = False
        elif re.match(r'^\d+\.', line):
            # Numbered list item
            if current_item and sub_items:
                # Store the previous item with its sub-items
                pdf_content.append([current_item, sub_items])
                sub_items = []
            
            current_item = line.strip()
            in_list_item = True
        elif line.startswith('- ') and in_list_item:
            # Sub-item under a numbered list item
            sub_items.append(line.strip())
        else:
            # Regular line
            if not in_list_item:
                pdf_content.append(line.strip())
    
    # Add the last item if there is one
    if current_item and sub_items:
        pdf_content.append([current_item, sub_items])
    
    # Split the content for two columns
    mid_point = len(pdf_content) // 2
    left_column = pdf_content[:mid_point]
    right_column = pdf_content[mid_point:]
    
    return left_column, right_column

# Demo functions for PDF libraries
def demo_pikepdf():
    """Create a two-column PDF with the markdown outline using pikepdf"""
    # Process markdown content
    left_column, right_column = markdown_to_pdf_content(ml_markdown)
    
    # We'll use pymupdf (fitz) to create the content, then save with pikepdf
    doc = fitz.open()
    page = doc.new_page(width=842, height=595)  # A4 Landscape
    
    # Set up fonts and colors
    title_font = "helv-b"
    section_font = "helv-b"
    item_font = "helv-b"
    subitem_font = "helv"
    blue_color = (0, 0, 0.8)
    black_color = (0, 0, 0)
    
    # Add title
    page.insert_text((50, 40), "Cutting-Edge ML Outline (PikePDF Demo)", fontname=title_font, fontsize=16, color=blue_color)
    
    # First column
    x1, y1 = 50, 80
    current_y = y1
    
    for item in left_column:
        if isinstance(item, str) and item.startswith('<b>'):
            # Section header
            # Add extra space before sections (except the first one)
            if current_y > y1:
                current_y += 10
            
            text = item.replace('<b>', '').replace('</b>', '')
            page.insert_text((x1, current_y), text, fontname=section_font, fontsize=14, color=blue_color)
            current_y += 25
        elif isinstance(item, list):
            # Main item with sub-items
            main_item, sub_items = item
            page.insert_text((x1, current_y), main_item, fontname=item_font, fontsize=12, color=black_color)
            current_y += 20
            
            # Add sub-items
            for sub_item in sub_items:
                page.insert_text((x1 + 20, current_y), sub_item, fontname=subitem_font, fontsize=10, color=black_color)
                current_y += 15
            
            current_y += 5  # Extra space after a group
        else:
            # Regular item
            page.insert_text((x1, current_y), item, fontname=item_font, fontsize=12, color=black_color)
            current_y += 20
    
    # Second column
    x2, y2 = 450, 80
    current_y = y2
    
    for item in right_column:
        if isinstance(item, str) and item.startswith('<b>'):
            # Section header
            # Add extra space before sections (except the first one)
            if current_y > y2:
                current_y += 10
            
            text = item.replace('<b>', '').replace('</b>', '')
            page.insert_text((x2, current_y), text, fontname=section_font, fontsize=14, color=blue_color)
            current_y += 25
        elif isinstance(item, list):
            # Main item with sub-items
            main_item, sub_items = item
            page.insert_text((x2, current_y), main_item, fontname=item_font, fontsize=12, color=black_color)
            current_y += 20
            
            # Add sub-items
            for sub_item in sub_items:
                page.insert_text((x2 + 20, current_y), sub_item, fontname=subitem_font, fontsize=10, color=black_color)
                current_y += 15
            
            current_y += 5  # Extra space after a group
        else:
            # Regular item
            page.insert_text((x2, current_y), item, fontname=item_font, fontsize=12, color=black_color)
            current_y += 20
    
    # Draw a dividing line
    page.draw_line((421, 70), (421, 550))
    
    # Convert to pikepdf
    temp_buffer = io.BytesIO()
    doc.save(temp_buffer)
    temp_buffer.seek(0)
    
    pdf = pikepdf.Pdf.open(temp_buffer)
    
    # Save to buffer
    buffer = io.BytesIO()
    pdf.save(buffer)
    buffer.seek(0)
    return buffer.getvalue()

def demo_fpdf():
    """Create a two-column PDF with the markdown outline using FPDF"""
    # Process markdown content
    left_column, right_column = markdown_to_pdf_content(ml_markdown)
    
    pdf = fpdf.FPDF(orientation='L')  # Landscape
    pdf.add_page()
    
    # Set title
    pdf.set_font("Arial", 'B', size=16)
    pdf.set_text_color(0, 0, 128)  # Dark blue
    pdf.cell(0, 10, txt="Cutting-Edge ML Outline (FPDF Demo)", ln=True, align='C')
    pdf.ln(10)
    
    # Define coordinates for columns
    x_col1 = 20
    x_col2 = pdf.w / 2 + 10
    y_start = pdf.get_y()
    
    # Function to render a column
    def render_column(items, x_start, y_start):
        y_pos = y_start
        
        for item in items:
            if isinstance(item, str) and item.startswith('<b>'):
                # Section header
                text = item.replace('<b>', '').replace('</b>', '')
                pdf.set_font("Arial", 'B', size=14)
                pdf.set_text_color(0, 0, 128)  # Dark blue
                pdf.set_xy(x_start, y_pos)
                pdf.cell(0, 10, txt=text, ln=True)
                y_pos += 10
            elif isinstance(item, list):
                # Main item with sub-items
                main_item, sub_items = item
                
                # Main item
                pdf.set_font("Arial", 'B', size=11)
                pdf.set_text_color(0, 0, 0)  # Black
                pdf.set_xy(x_start, y_pos)
                pdf.multi_cell(180, 6, txt=main_item, align='L')
                main_height = pdf.get_y() - y_pos
                y_pos += main_height + 2
                
                # Sub-items
                pdf.set_font("Arial", size=10)
                for sub_item in sub_items:
                    pdf.set_xy(x_start + 10, y_pos)
                    pdf.multi_cell(170, 5, txt=sub_item, align='L')
                    sub_height = pdf.get_y() - y_pos
                    y_pos += sub_height + 1
                
                y_pos += 2  # Extra space after a group
            else:
                # Regular item
                pdf.set_font("Arial", 'B', size=11)
                pdf.set_text_color(0, 0, 0)  # Black
                pdf.set_xy(x_start, y_pos)
                pdf.multi_cell(180, 6, txt=item, align='L')
                item_height = pdf.get_y() - y_pos
                y_pos += item_height + 4
    
    # Render both columns
    render_column(left_column, x_col1, y_start)
    render_column(right_column, x_col2, y_start)
    
    # Draw a dividing line
    pdf.line(pdf.w/2, 30, pdf.w/2, 280)
    
    buffer = io.BytesIO()
    pdf.output(buffer)
    buffer.seek(0)
    return buffer.getvalue()

def demo_pymupdf():
    """Create a two-column PDF with the markdown outline using PyMuPDF"""
    # Process markdown content
    left_column, right_column = markdown_to_pdf_content(ml_markdown)
    
    doc = fitz.open()
    page = doc.new_page(width=842, height=595)  # A4 Landscape
    
    # Set up fonts and colors
    title_font = "helv-b"
    section_font = "helv-b"
    item_font = "helv-b"
    subitem_font = "helv"
    blue_color = (0, 0, 0.8)
    black_color = (0, 0, 0)
    
    # Add title
    page.insert_text((300, 40), "Cutting-Edge ML Outline (PyMuPDF Demo)", fontname=title_font, fontsize=16, color=blue_color)
    
    # First column
    x1, y1 = 50, 80
    current_y = y1
    
    for item in left_column:
        if isinstance(item, str) and item.startswith('<b>'):
            # Section header
            # Add extra space before sections (except the first one)
            if current_y > y1:
                current_y += 10
            
            text = item.replace('<b>', '').replace('</b>', '')
            page.insert_text((x1, current_y), text, fontname=section_font, fontsize=14, color=blue_color)
            current_y += 25
        elif isinstance(item, list):
            # Main item with sub-items
            main_item, sub_items = item
            page.insert_text((x1, current_y), main_item, fontname=item_font, fontsize=12, color=black_color)
            current_y += 20
            
            # Add sub-items
            for sub_item in sub_items:
                page.insert_text((x1 + 20, current_y), sub_item, fontname=subitem_font, fontsize=10, color=black_color)
                current_y += 15
            
            current_y += 5  # Extra space after a group
        else:
            # Regular item
            page.insert_text((x1, current_y), item, fontname=item_font, fontsize=12, color=black_color)
            current_y += 20
    
    # Second column
    x2, y2 = 450, 80
    current_y = y2
    
    for item in right_column:
        if isinstance(item, str) and item.startswith('<b>'):
            # Section header
            # Add extra space before sections (except the first one)
            if current_y > y2:
                current_y += 10
            
            text = item.replace('<b>', '').replace('</b>', '')
            page.insert_text((x2, current_y), text, fontname=section_font, fontsize=14, color=blue_color)
            current_y += 25
        elif isinstance(item, list):
            # Main item with sub-items
            main_item, sub_items = item
            page.insert_text((x2, current_y), main_item, fontname=item_font, fontsize=12, color=black_color)
            current_y += 20
            
            # Add sub-items
            for sub_item in sub_items:
                page.insert_text((x2 + 20, current_y), sub_item, fontname=subitem_font, fontsize=10, color=black_color)
                current_y += 15
            
            current_y += 5  # Extra space after a group
        else:
            # Regular item
            page.insert_text((x2, current_y), item, fontname=item_font, fontsize=12, color=black_color)
            current_y += 20
    
    # Draw a dividing line
    page.draw_line((421, 70), (421, 550))
    
    buffer = io.BytesIO()
    doc.save(buffer)
    buffer.seek(0)
    return buffer.getvalue()

# Demo function for image capture
def demo_image_capture():
    """Generate a demo image (fake capture) since we can't access the camera in this environment"""
    # Create a simple gradient image using numpy and PIL
    width, height = 640, 480
    
    # Create a gradient array
    x = np.linspace(0, 1, width)
    y = np.linspace(0, 1, height)
    xx, yy = np.meshgrid(x, y)
    gradient = (xx + yy) / 2
    
    # Convert to RGB image
    img_array = (gradient * 255).astype(np.uint8)
    rgb_array = np.stack([img_array, img_array//2, img_array*2], axis=2)
    
    # Create PIL Image
    img = Image.fromarray(rgb_array)
    
    # Add text to the image
    from PIL import ImageDraw, ImageFont
    draw = ImageDraw.Draw(img)
    try:
        font = ImageFont.truetype("arial.ttf", 30)
    except:
        font = ImageFont.load_default()
    
    draw.text((width//4, height//2), "OpenCV Demo Image", fill=(255, 255, 255), font=font)
    
    # Save to buffer
    buffer = io.BytesIO()
    img.save(buffer, format="JPEG")
    buffer.seek(0)
    return buffer.getvalue()

# Main PDF creation using ReportLab
def create_main_pdf(markdown_text):
    """Create a single-page landscape PDF with the outline in two columns"""
    from reportlab.platypus import Table, TableStyle, Paragraph, Spacer
    from reportlab.lib import pagesizes
    from reportlab.lib.styles import getSampleStyleSheet, ParagraphStyle
    
    # Process markdown content
    left_column, right_column = markdown_to_pdf_content(markdown_text)
    
    buffer = io.BytesIO()
    doc = SimpleDocTemplate(
        buffer, 
        pagesize=(A4[1], A4[0]),  # Landscape
        leftMargin=50,
        rightMargin=50,
        topMargin=50,
        bottomMargin=50
    )
    
    styles = getSampleStyleSheet()
    story = []
    
    # Create custom styles
    title_style = styles['Heading1']
    title_style.textColor = colors.darkblue
    title_style.alignment = 1  # Center alignment
    
    section_style = ParagraphStyle(
        'SectionStyle',
        parent=styles['Heading2'],
        textColor=colors.darkblue,
        spaceAfter=6
    )
    
    item_style = ParagraphStyle(
        'ItemStyle',
        parent=styles['Normal'],
        fontSize=11,
        leading=14,
        fontName='Helvetica-Bold'
    )
    
    subitem_style = ParagraphStyle(
        'SubItemStyle',
        parent=styles['Normal'],
        fontSize=10,
        leading=12,
        leftIndent=20
    )
    
    # Add title
    story.append(Paragraph("Cutting-Edge ML Outline (ReportLab)", title_style))
    story.append(Spacer(1, 20))
    
    # Prepare data for table
    left_cells = []
    for item in left_column:
        if isinstance(item, str) and item.startswith('<b>'):
            # Section header
            text = item.replace('<b>', '').replace('</b>', '')
            left_cells.append(Paragraph(text, section_style))
        elif isinstance(item, list):
            # Main item with sub-items
            main_item, sub_items = item
            left_cells.append(Paragraph(main_item, item_style))
            
            # Sub items
            for sub_item in sub_items:
                left_cells.append(Paragraph(sub_item, subitem_style))
        else:
            # Regular item
            left_cells.append(Paragraph(item, item_style))
    
    right_cells = []
    for item in right_column:
        if isinstance(item, str) and item.startswith('<b>'):
            # Section header
            text = item.replace('<b>', '').replace('</b>', '')
            right_cells.append(Paragraph(text, section_style))
        elif isinstance(item, list):
            # Main item with sub-items
            main_item, sub_items = item
            right_cells.append(Paragraph(main_item, item_style))
            
            # Sub items
            for sub_item in sub_items:
                right_cells.append(Paragraph(sub_item, subitem_style))
        else:
            # Regular item
            right_cells.append(Paragraph(item, item_style))
    
    # Make sure both columns have the same number of rows by adding empty cells
    max_cells = max(len(left_cells), len(right_cells))
    if len(left_cells) < max_cells:
        for i in range(max_cells - len(left_cells)):
            left_cells.append("")
    if len(right_cells) < max_cells:
        for i in range(max_cells - len(right_cells)):
            right_cells.append("")
    
    # Create table data (one row per cell)
    table_data = []
    for i in range(max_cells):
        table_data.append([left_cells[i], right_cells[i]])
    
    # Calculate column widths
    col_width = (A4[1] - 120) / 2.0  # Page width minus margins divided by 2
    
    # Create the table with the data
    table = Table(table_data, colWidths=[col_width, col_width])
    
    # Style the table
    table.setStyle(TableStyle([
        ('VALIGN', (0, 0), (-1, -1), 'TOP'),
        ('ALIGN', (0, 0), (0, -1), 'LEFT'),
        ('ALIGN', (1, 0), (1, -1), 'LEFT'),
        ('BACKGROUND', (0, 0), (-1, -1), colors.white),
        ('GRID', (0, 0), (-1, -1), 0.5, colors.white),
        ('LINEAFTER', (0, 0), (0, -1), 1, colors.grey),
    ]))
    
    story.append(table)
    
    doc.build(story)
    buffer.seek(0)
    return buffer.getvalue()

def get_binary_file_downloader_html(bin_data, file_label='File'):
    """Create a download link for binary data"""
    bin_str = base64.b64encode(bin_data).decode()
    href = f'<a href="data:application/octet-stream;base64,{bin_str}" download="{file_label}">Download {file_label}</a>'
    return href

# Streamlit UI
st.title("🚀 Cutting-Edge ML Outline Generator")

col1, col2 = st.columns(2)

with col1:
    st.header("📝 Markdown Outline")
    
    # Display the markdown content
    st.markdown(ml_markdown)
    
    # Create a download button for the markdown file
    st.download_button(
        label="Download Markdown",
        data=ml_markdown,
        file_name="ml_outline.md",
        mime="text/markdown"
    )
    
    # Show the markdown source code in an expandable section
    with st.expander("View Markdown Source"):
        st.code(ml_markdown, language="markdown")

with col2:
    st.header("📑 PDF Preview & Demos")
    
    # Library Demos
    st.subheader("Library Demos")
    if st.button("Run PDF Library Demos"):
        with st.spinner("Running demos..."):
            # Create tabs for each demo
            demo_tabs = st.tabs(["PikePDF", "FPDF", "PyMuPDF", "Image Demo"])
            
            with demo_tabs[0]:
                # pikepdf demo
                pike_pdf = demo_pikepdf()
                st.download_button("Download pikepdf Demo", pike_pdf, "pikepdf_demo.pdf")
                st.write("PikePDF demo created successfully!")
                st.info("This PDF contains the multilevel markdown outline in a two-column layout.")
            
            with demo_tabs[1]:
                # fpdf demo
                fpdf_pdf = demo_fpdf()
                st.download_button("Download fpdf Demo", fpdf_pdf, "fpdf_demo.pdf")
                st.write("FPDF demo created successfully!")
                st.info("This PDF contains the multilevel markdown outline in a two-column layout.")
            
            with demo_tabs[2]:
                # pymupdf demo
                pymupdf_pdf = demo_pymupdf()
                st.download_button("Download pymupdf Demo", pymupdf_pdf, "pymupdf_demo.pdf")
                st.write("PyMuPDF demo created successfully!")
                st.info("This PDF contains the multilevel markdown outline in a two-column layout.")
            
            with demo_tabs[3]:
                # Image demo
                img_data = demo_image_capture()
                st.image(img_data, caption="Demo Image (Camera simulation)")

    # Main PDF Generation
    st.subheader("Main Outline PDF")
    if st.button("Generate Main PDF"):
        with st.spinner("Generating PDF..."):
            try:
                pdf_bytes = create_main_pdf(ml_markdown)
                
                st.download_button(
                    label="Download Main PDF",
                    data=pdf_bytes,
                    file_name="ml_outline.pdf",
                    mime="application/pdf"
                )
                
                # Display the PDF in the app
                base64_pdf = base64.b64encode(pdf_bytes).decode('utf-8')
                pdf_display = f'''
                    <embed 
                        src="data:application/pdf;base64,{base64_pdf}" 
                        width="100%" 
                        height="400px" 
                        type="application/pdf">
                '''
                st.markdown(pdf_display, unsafe_allow_html=True)
                
                st.success("PDF generated successfully! The PDF displays the multilevel markdown outline in a two-column layout.")
            except Exception as e:
                st.error(f"Error generating PDF: {str(e)}")

    # Show the PDF rendering code in an expandable section
    with st.expander("View PDF Rendering Code"):
        st.code("""
# Process multilevel markdown for PDF output
def markdown_to_pdf_content(markdown_text):
    # Convert markdown headers to styled text for PDF
    lines = markdown_text.strip().split('\\n')
    pdf_content = []
    
    for line in lines:
        if line.startswith('# '):
            # Main header - will be handled separately
            pass
        elif line.startswith('## '):
            # Section header - add as a bold item
            section = line.replace('## ', '').strip()
            pdf_content.append(f"<b>{section}</b>")
        elif re.match(r'^\\d+\\.', line):
            # Numbered list item
            item = line.strip()
            pdf_content.append(item)
        elif line.startswith('- '):
            # Sub-item under a numbered list item
            sub_item = line.strip()
            pdf_content.append("  " + sub_item)
    
    # Split the content for two columns
    mid_point = len(pdf_content) // 2
    left_column = pdf_content[:mid_point]
    right_column = pdf_content[mid_point:]
    
    return left_column, right_column
        """, language="python")

# Add custom CSS for better appearance
st.markdown("""
<style>
    .stButton>button {
        background-color: #4CAF50;
        color: white;
        font-weight: bold;
    }
    .stTabs [data-baseweb="tab-list"] {
        gap: 2px;
    }
    .stTabs [data-baseweb="tab"] {
        height: 50px;
        white-space: pre-wrap;
        background-color: #f0f2f6;
        border-radius: 4px 4px 0px 0px;
        gap: 1px;
        padding-top: 10px;
        padding-bottom: 10px;
    }
</style>
""", unsafe_allow_html=True)