app.py

import io
import os
import re
import glob
import textwrap
from datetime import datetime
from pathlib import Path
import streamlit as st
import pandas as pd
from PIL import Image
from reportlab.pdfgen import canvas
from reportlab.lib.pagesizes import letter, A4
from reportlab.lib.utils import ImageReader
from reportlab.platypus import SimpleDocTemplate, Paragraph, Spacer, Table, TableStyle, Image as ReportLabImage
from reportlab.lib.styles import getSampleStyleSheet, ParagraphStyle
from reportlab.lib import colors
from reportlab.pdfbase import pdfmetrics
from reportlab.pdfbase.ttfonts import TTFont
import mistune
import fitz
import edge_tts
import asyncio
import base64
from urllib.parse import quote

# Page config
st.set_page_config(page_title="PDF & Code Interpreter", layout="wide", page_icon="🚀")

def delete_asset(path):
    try:
        os.remove(path)
    except Exception as e:
        st.error(f"Error deleting file: {e}")
    st.rerun()

async def generate_audio(text, voice, filename):
    communicate = edge_tts.Communicate(text, voice)
    await communicate.save(filename)
    return filename

def clean_for_speech(text):
    text = text.replace("#", "")
    emoji_pattern = re.compile(
        r"[\U0001F300-\U0001F5FF"
        r"\U0001F600-\U0001F64F"
        r"\U0001F680-\U0001F6FF"
        r"\U0001F700-\U0001F77F"
        r"\U0001F780-\U0001F7FF"
        r"\U0001F800-\U0001F8FF"
        r"\U0001F900-\U0001F9FF"
        r"\U0001FA00-\U0001FA6F"
        r"\U0001FA70-\U0001FAFF"
        r"\u2600-\u26FF"
        r"\u2700-\u27BF]+", flags=re.UNICODE)
    return emoji_pattern.sub('', text)

def detect_and_convert_links(text):
    md_link_pattern = re.compile(r'\[(.*?)\]\((https?://[^\s\[\]()<>{}]+)\)')
    text = md_link_pattern.sub(r'<a href="\2" color="blue">\1</a>', text)
    url_pattern = re.compile(r'(?<!href=")(https?://[^\s<>{}]+)', re.IGNORECASE)
    text = url_pattern.sub(r'<a href="\1" color="blue">\1</a>', text)
    return text

def apply_emoji_font(text, emoji_font):
    tag_pattern = re.compile(r'(<[^>]+>)')
    segments = tag_pattern.split(text)
    result = []
    emoji_pattern = re.compile(
        r"([\U0001F300-\U0001F5FF"
        r"\U0001F600-\U0001F64F"
        r"\U0001F680-\U0001F6FF"
        r"\U0001F700-\U0001F77F"
        r"\U0001F780-\U0001F7FF"
        r"\U0001F800-\U0001F8FF"
        r"\U0001F900-\U0001F9FF"
        r"\U0001FAD0-\U0001FAD9"
        r"\U0001FA00-\U0001FA6F"
        r"\U0001FA70-\U0001FAFF"
        r"\u2600-\u26FF"
        r"\u2700-\u27BF]+)"
    )
    def replace_emoji(match):
        emoji = match.group(1)
        return f'<font face="{emoji_font}">{emoji}</font>'
    for segment in segments:
        if tag_pattern.match(segment):
            result.append(segment)
        else:
            parts = []
            last_pos = 0
            for match in emoji_pattern.finditer(segment):
                start, end = match.span()
                if last_pos < start:
                    parts.append(f'<font face="DejaVuSans">{segment[last_pos:start]}</font>')
                parts.append(replace_emoji(match))
                last_pos = end
            if last_pos < len(segment):
                parts.append(f'<font face="DejaVuSans">{segment[last_pos:]}</font>')
            result.append(''.join(parts))
    return ''.join(result)

def markdown_to_pdf_content(markdown_text, add_space_before_numbered, headings_to_fonts):
    lines = markdown_text.strip().split('\n')
    pdf_content = []
    number_pattern = re.compile(r'^\d+(\.\d+)*\.\s')
    heading_pattern = re.compile(r'^(#{1,4})\s+(.+)$')
    first_numbered_seen = False
    for line in lines:
        line = line.strip()
        if not line:
            continue
        if headings_to_fonts and line.startswith('#'):
            heading_match = heading_pattern.match(line)
            if heading_match:
                level = len(heading_match.group(1))
                heading_text = heading_match.group(2).strip()
                formatted_heading = f"<h{level}>{heading_text}</h{level}>"
                pdf_content.append(formatted_heading)
                continue
        is_numbered_line = number_pattern.match(line) is not None
        if add_space_before_numbered and is_numbered_line:
            if first_numbered_seen and not line.startswith("1."):
                pdf_content.append("")
            if not first_numbered_seen:
                first_numbered_seen = True
        line = detect_and_convert_links(line)
        line = re.sub(r'\*\*(.+?)\*\*', r'<b>\1</b>', line)
        line = re.sub(r'\*([^*]+?)\*', r'<b>\1</b>', line)
        pdf_content.append(line)
    total_lines = len(pdf_content)
    return pdf_content, total_lines

def create_pdf(markdown_texts, image_files, base_font_size=14, num_columns=2, add_space_before_numbered=True, headings_to_fonts=True, doc_title="Combined Document"):
    if not markdown_texts and not image_files:
        return None
    buffer = io.BytesIO()
    page_width = A4[0] * 2
    page_height = A4[1]
    doc = SimpleDocTemplate(
        buffer,
        pagesize=(page_width, page_height),
        leftMargin=36,
        rightMargin=36,
        topMargin=36,
        bottomMargin=36,
        title=doc_title
    )
    styles = getSampleStyleSheet()
    spacer_height = 10
    try:
        pdfmetrics.registerFont(TTFont("DejaVuSans", "DejaVuSans.ttf"))
        pdfmetrics.registerFont(TTFont("NotoEmoji-Bold", "NotoEmoji-Bold.ttf"))
    except Exception as e:
        st.error(f"Font registration error: {e}")
        return None
    story = []
    for markdown_text in markdown_texts:
        pdf_content, total_lines = markdown_to_pdf_content(markdown_text, add_space_before_numbered, headings_to_fonts)
        total_chars = sum(len(line) for line in pdf_content)
        hierarchy_weight = sum(1.5 if line.startswith("<b>") else 1 for line in pdf_content)
        longest_line_words = max(len(line.split()) for line in pdf_content) if pdf_content else 0
        content_density = total_lines * hierarchy_weight + total_chars / 50
        usable_height = page_height - 72 - spacer_height
        usable_width = page_width - 72
        avg_line_chars = total_chars / total_lines if total_lines > 0 else 50
        col_width = usable_width / num_columns
        min_font_size = 5
        max_font_size = 16
        lines_per_col = total_lines / num_columns if num_columns > 0 else total_lines
        target_height_per_line = usable_height / lines_per_col if lines_per_col > 0 else usable_height
        estimated_font_size = int(target_height_per_line / 1.5)
        adjusted_font_size = max(min_font_size, min(max_font_size, estimated_font_size))
        if avg_line_chars > col_width / adjusted_font_size * 10:
            adjusted_font_size = int(col_width / (avg_line_chars / 10))
            adjusted_font_size = max(min_font_size, adjusted_font_size)
        if longest_line_words > 17 or lines_per_col > 20:
            font_scale = min(17 / max(longest_line_words, 17), 60 / max(lines_per_col, 20))
            adjusted_font_size = max(min_font_size, int(base_font_size * font_scale))
        item_style = ParagraphStyle(
            'ItemStyle', parent=styles['Normal'], fontName="DejaVuSans",
            fontSize=adjusted_font_size, leading=adjusted_font_size * 1.15, spaceAfter=1,
            linkUnderline=True
        )
        numbered_bold_style = ParagraphStyle(
            'NumberedBoldStyle', parent=styles['Normal'], fontName="NotoEmoji-Bold",
            fontSize=adjusted_font_size, leading=adjusted_font_size * 1.15, spaceAfter=1,
            linkUnderline=True
        )
        section_style = ParagraphStyle(
            'SectionStyle', parent=styles['Heading2'], fontName="DejaVuSans",
            textColor=colors.darkblue, fontSize=adjusted_font_size * 1.1, leading=adjusted_font_size * 1.32, spaceAfter=2,
            linkUnderline=True
        )
        columns = [[] for _ in range(num_columns)]
        lines_per_column = total_lines / num_columns if num_columns > 0 else total_lines
        current_line_count = 0
        current_column = 0
        number_pattern = re.compile(r'^\d+(\.\d+)*\.\s')
        for item in pdf_content:
            if current_line_count >= lines_per_column and current_column < num_columns - 1:
                current_column += 1
                current_line_count = 0
            columns[current_column].append(item)
            current_line_count += 1
        column_cells = [[] for _ in range(num_columns)]
        for col_idx, column in enumerate(columns):
            for item in column:
                if isinstance(item, str):
                    heading_match = re.match(r'<h(\d)>(.*?)</h\1>', item) if headings_to_fonts else None
                    if heading_match:
                        level = int(heading_match.group(1))
                        heading_text = heading_match.group(2)
                        heading_style = ParagraphStyle(
                            f'Heading{level}Style',
                            parent=styles['Heading1'],
                            fontName="DejaVuSans",
                            textColor=colors.darkblue if level == 1 else (colors.black if level > 2 else colors.blue),
                            fontSize=adjusted_font_size * (1.6 - (level-1)*0.15),
                            leading=adjusted_font_size * (1.8 - (level-1)*0.15),
                            spaceAfter=4 - (level-1),
                            spaceBefore=6 - (level-1),
                            linkUnderline=True
                        )
                        column_cells[col_idx].append(Paragraph(apply_emoji_font(heading_text, "NotoEmoji-Bold"), heading_style))
                    elif item.startswith("<b>") and item.endswith("</b>"):
                        content = item[3:-4].strip()
                        if number_pattern.match(content):
                            column_cells[col_idx].append(Paragraph(apply_emoji_font(content, "NotoEmoji-Bold"), numbered_bold_style))
                        else:
                            column_cells[col_idx].append(Paragraph(apply_emoji_font(content, "NotoEmoji-Bold"), section_style))
                    else:
                        column_cells[col_idx].append(Paragraph(apply_emoji_font(item, "NotoEmoji-Bold"), item_style))
                else:
                    column_cells[col_idx].append(Paragraph(apply_emoji_font(str(item), "NotoEmoji-Bold"), item_style))
        max_cells = max(len(cells) for cells in column_cells) if column_cells else 0
        for cells in column_cells:
            cells.extend([Paragraph("", item_style)] * (max_cells - len(cells)))
        table_data = list(zip(*column_cells)) if column_cells else [[]]
        table = Table(table_data, colWidths=[col_width] * num_columns, hAlign='CENTER')
        table.setStyle(TableStyle([
            ('VALIGN', (0, 0), (-1, -1), 'TOP'),
            ('ALIGN', (0, 0), (-1, -1), 'LEFT'),
            ('BACKGROUND', (0, 0), (-1, -1), colors.white),
            ('GRID', (0, 0), (-1, -1), 0, colors.white),
            ('LINEAFTER', (0, 0), (num_columns-1, -1), 0.5, colors.grey),
            ('LEFTPADDING', (0, 0), (-1, -1), 2),
            ('RIGHTPADDING', (0, 0), (-1, -1), 2),
            ('TOPPADDING', (0, 0), (-1, -1), 1),
            ('BOTTOMPADDING', (0, 0), (-1, -1), 1),
        ]))
        story.append(Spacer(1, spacer_height))
        story.append(table)
        story.append(Spacer(1, spacer_height * 2))
    for img_path in image_files:
        try:
            img = Image.open(img_path)
            img_width, img_height = img.size
            page_width, page_height = A4
            scale = min((page_width - 40) / img_width, (page_height - 40) / img_height)
            new_width = img_width * scale
            new_height = img_height * scale
            story.append(ReportLabImage(img_path, width=new_width, height=new_height))
            story.append(Spacer(1, spacer_height))
        except Exception as e:
            st.warning(f"Could not process image {img_path}: {e}")
            continue
    doc.build(story)
    buffer.seek(0)
    return buffer.getvalue()

def pdf_to_image(pdf_bytes):
    if pdf_bytes is None:
        return None
    try:
        doc = fitz.open(stream=pdf_bytes, filetype="pdf")
        images = []
        for page in doc:
            pix = page.get_pixmap(matrix=fitz.Matrix(2.0, 2.0))
            img = Image.frombytes("RGB", [pix.width, pix.height], pix.samples)
            images.append(img)
        doc.close()
        return images
    except Exception as e:
        st.error(f"Failed to render PDF preview: {e}")
        return None

def get_video_html(video_path, width="100%"):
    try:
        video_url = f"data:video/mp4;base64,{base64.b64encode(open(video_path, 'rb').read()).decode()}"
        return f'''
        <video width="{width}" controls autoplay muted loop>
            <source src="{video_url}" type="video/mp4">
            Your browser does not support the video tag.
        </video>
        '''
    except Exception as e:
        st.warning(f"Could not load video {video_path}: {e}")
        return ""

def display_glossary_entity(k):
    search_urls = {
        "🚀🌌ArXiv": lambda k: f"https://arxiv.org/search/?query={quote(k)}&searchtype=all",
        "📖": lambda k: f"https://en.wikipedia.org/wiki/{quote(k)}",
        "🔍": lambda k: f"https://www.google.com/search?q={quote(k)}",
        "🎥": lambda k: f"https://www.youtube.com/results?search_query={quote(k)}",
    }
    links_md = ' '.join([f"[{emoji}]({url(k)})" for emoji, url in search_urls.items()])
    st.markdown(f"**{k}** <small>{links_md}</small>", unsafe_allow_html=True)

# Tabs setup
tab1, tab2 = st.tabs(["📄 PDF Composer", "🧪 Code Interpreter"])

with tab1:
    st.header("📄 PDF Composer & Voice Generator 🚀")
    # Sidebar PDF text settings
    columns = st.sidebar.slider("Text columns", 1, 3, 2)
    font_family = st.sidebar.selectbox("Font", ["Helvetica", "Times-Roman", "Courier", "DejaVuSans"])
    font_size = st.sidebar.slider("Font size", 6, 24, 14)
    # Markdown input
    md_file = st.file_uploader("Upload Markdown (.md)", type=["md"])
    if md_file:
        md_text = md_file.getvalue().decode("utf-8")
        stem = Path(md_file.name).stem
    else:
        md_text = st.text_area("Or enter markdown text directly", height=200)
        stem = datetime.now().strftime('%Y%m%d_%H%M%S')
    # Convert Markdown to plain text
    renderer = mistune.HTMLRenderer()
    markdown = mistune.create_markdown(renderer=renderer)
    html = markdown(md_text or "")
    plain_text = re.sub(r'<[^>]+>', '', html)
    # Voice settings
    languages = {"English (US)": "en", "English (UK)": "en-uk", "Spanish": "es"}
    voice_choice = st.selectbox("Voice Language", list(languages.keys()))
    voice_lang = languages[voice_choice]
    slow = st.checkbox("Slow Speech")
    VOICES = ["en-US-AriaNeural", "en-US-JennyNeural", "en-GB-SoniaNeural", "en-US-GuyNeural", "en-US-AnaNeural"]
    selected_voice = st.selectbox("Select Voice for TTS", options=VOICES, index=0)
    if st.button("🔊 Generate & Download Voice MP3 from Text"):
        if plain_text.strip():
            voice_file = f"{stem}_{selected_voice}.mp3"
            try:
                cleaned_text = clean_for_speech(plain_text)
                audio_file = asyncio.run(generate_audio(cleaned_text, selected_voice, voice_file))
                st.audio(audio_file)
                with open(audio_file, 'rb') as mp3:
                    st.download_button("📥 Download MP3", data=mp3, file_name=voice_file, mime="audio/mpeg")
            except Exception as e:
                st.error(f"Error generating voice: {e}")
        else:
            st.warning("No text to generate voice from.")
    # Image uploads and ordering
    imgs = st.file_uploader("Upload Images for PDF", type=["png", "jpg", "jpeg"], accept_multiple_files=True)
    ordered_images = []
    if imgs:
        df_imgs = pd.DataFrame([{"name": f.name, "order": i} for i, f in enumerate(imgs)])
        edited = st.data_editor(df_imgs, use_container_width=True, num_rows="dynamic")
        for _, row in edited.sort_values("order").iterrows():
            for f in imgs:
                if f.name == row['name']:
                    ordered_images.append(f)
                    break
    if st.button("🖋️ Generate PDF with Markdown & Images"):
        if not plain_text.strip() and not ordered_images:
            st.warning("Please provide some text or upload images to generate a PDF.")
        else:
            buf = io.BytesIO()
            c = canvas.Canvas(buf)
            if plain_text.strip():
                page_w, page_h = letter
                margin = 40
                gutter = 20
                col_w = (page_w - 2*margin - (columns-1)*gutter) / columns
                c.setFont(font_family, font_size)
                line_height = font_size * 1.2
                col = 0
                x = margin
                y = page_h - margin
                avg_char_width = font_size * 0.6
                wrap_width = int(col_w / avg_char_width) if avg_char_width > 0 else 100
                for paragraph in plain_text.split("\n"):
                    if not paragraph.strip():
                        y -= line_height
                        if y < margin:
                            col += 1
                            if col >= columns:
                                c.showPage()
                                c.setFont(font_family, font_size)
                                col = 0
                            x = margin + col*(col_w+gutter)
                            y = page_h - margin
                        continue
                    for line in textwrap.wrap(paragraph, wrap_width):
                        if y < margin:
                            col += 1
                            if col >= columns:
                                c.showPage()
                                c.setFont(font_family, font_size)
                                col = 0
                            x = margin + col*(col_w+gutter)
                            y = page_h - margin
                        c.drawString(x, y, line)
                        y -= line_height
                    y -= line_height
            for img_f in ordered_images:
                try:
                    img = Image.open(img_f)
                    w, h = img.size
                    c.showPage()
                    c.setPageSize((w, h))
                    c.drawImage(ImageReader(img), 0, 0, w, h, preserveAspectRatio=False)
                except Exception as e:
                    st.warning(f"Could not process image {img_f.name}: {e}")
                    continue
            c.save()
            buf.seek(0)
            pdf_name = f"{stem}.pdf"
            st.download_button("⬇️ Download PDF", data=buf, file_name=pdf_name, mime="application/pdf")
    st.markdown("---")
    st.subheader("📂 Available Assets")
    all_assets = glob.glob("*.*")
    excluded_extensions = ['.py', '.ttf', '.txt']
    excluded_files = ['README.md', 'index.html']
    assets = sorted([
        a for a in all_assets
        if not (a.lower().endswith(tuple(excluded_extensions)) or a in excluded_files)
        and a.lower().endswith(('.md', '.png', '.jpg', '.jpeg'))
    ])
    if 'selected_assets' not in st.session_state:
        st.session_state.selected_assets = []
    if not assets:
        st.info("No available assets found.")
    else:
        for a in assets:
            ext = a.split('.')[-1].lower()
            cols = st.columns([1, 3, 1, 1])
            with cols[0]:
                is_selected = st.checkbox("", key=f"select_{a}", value=a in st.session_state.selected_assets)
                if is_selected and a not in st.session_state.selected_assets:
                    st.session_state.selected_assets.append(a)
                elif not is_selected and a in st.session_state.selected_assets:
                    st.session_state.selected_assets.remove(a)
            cols[1].write(a)
            try:
                if ext == 'md':
                    with open(a, 'r', encoding='utf-8') as f:
                        cols[2].download_button("📥", data=f.read(), file_name=a, mime="text/markdown")
                elif ext in ['png', 'jpg', 'jpeg']:
                    with open(a, 'rb') as img_file:
                        cols[2].download_button("⬇️", data=img_file, file_name=a, mime=f"image/{ext}")
                cols[3].button("🗑️", key=f"del_{a}", on_click=delete_asset, args=(a,))
            except Exception as e:
                cols[3].error(f"Error handling file {a}: {e}")
        if st.button("📑 Generate PDF from Selected Assets"):
            if not st.session_state.selected_assets:
                st.warning("Please select at least one asset to generate a PDF.")
            else:
                markdown_texts = []
                image_files = []
                for a in st.session_state.selected_assets:
                    ext = a.split('.')[-1].lower()
                    if ext == 'md':
                        with open(a, 'r', encoding='utf-8') as f:
                            markdown_texts.append(f.read())
                    elif ext in ['png', 'jpg', 'jpeg']:
                        image_files.append(a)
                with st.spinner("Generating PDF from selected assets..."):
                    pdf_bytes = create_pdf(
                        markdown_texts=markdown_texts,
                        image_files=image_files,
                        base_font_size=14,
                        num_columns=2,
                        add_space_before_numbered=True,
                        headings_to_fonts=True,
                        doc_title="Combined_Selected_Assets"
                    )
                    if pdf_bytes:
                        pdf_images = pdf_to_image(pdf_bytes)
                        if pdf_images:
                            st.subheader("Preview of Generated PDF")
                            for i, img in enumerate(pdf_images):
                                st.image(img, caption=f"Page {i+1}", use_container_width=True)
                        prefix = datetime.now().strftime("%Y%m%d_%H%M%S")
                        st.download_button(
                            label="💾 Download Combined PDF",
                            data=pdf_bytes,
                            file_name=f"{prefix}_combined.pdf",
                            mime="application/pdf"
                        )
                    else:
                        st.error("Failed to generate PDF.")
    st.markdown("---")
    st.subheader("🖼 Image Gallery")
    image_files = glob.glob("*.png") + glob.glob("*.jpg") + glob.glob("*.jpeg")
    image_cols = st.slider("Gallery Columns 🖼", min_value=1, max_value=15, value=5, key="image_cols")
    if image_files:
        cols = st.columns(image_cols)
        for idx, image_file in enumerate(image_files):
            with cols[idx % image_cols]:
                try:
                    img = Image.open(image_file)
                    st.image(img, caption=image_file, use_container_width=True)
                    display_glossary_entity(os.path.splitext(image_file)[0])
                except Exception as e:
                    st.warning(f"Could not load image {image_file}: {e}")
    else:
        st.info("No images found in the current directory.")
    st.markdown("---")
    st.subheader("🎥 Video Gallery")
    video_files = glob.glob("*.mp4")
    video_cols = st.slider("Gallery Columns 🎬", min_value=1, max_value=5, value=3, key="video_cols")
    if video_files:
        cols = st.columns(video_cols)
        for idx, video_file in enumerate(video_files):
            with cols[idx % video_cols]:
                st.markdown(get_video_html(video_file, width="100%"), unsafe_allow_html=True)
                display_glossary_entity(os.path.splitext(video_file)[0])
    else:
        st.info("No videos found in the current directory.")

with tab2:
    st.header("🧪 Python Code Executor & Demo")
    import io, sys
    from contextlib import redirect_stdout
    DEFAULT_CODE = '''import streamlit as st
import random
st.title("📊 Demo App")
st.markdown("Random number and color demo")
col1, col2 = st.columns(2)
with col1:
    num = st.number_input("Number:", 1, 100, 10)
    mul = st.slider("Multiplier:", 1, 10, 2)
    if st.button("Calc"):
        st.write(num * mul)
with col2:
    color = st.color_picker("Pick color","#ff0000")
    st.markdown(f'<div style="background:{color};padding:10px;">Color</div>', unsafe_allow_html=True)
'''
    def extract_python_code(md: str) -> list:
        return re.findall(r"```python\s*(.*?)```", md, re.DOTALL)
    def execute_code(code: str) -> tuple:
        buf = io.StringIO(); local_vars = {}
        try:
            with redirect_stdout(buf):
                exec(code, {}, local_vars)
            return buf.getvalue(), None
        except Exception as e:
            return None, str(e)
    up = st.file_uploader("Upload .py or .md", type=['py', 'md'])
    if 'code' not in st.session_state:
        st.session_state.code = DEFAULT_CODE
    if up:
        text = up.getvalue().decode()
        if up.type == 'text/markdown':
            codes = extract_python_code(text)
            if codes:
                st.session_state.code = codes[0].strip()
            else:
                st.warning("No Python code block found in the markdown file.")
                st.session_state.code = ''
        else:
            st.session_state.code = text.strip()
        st.code(st.session_state.code, language='python')
    else:
        st.session_state.code = st.text_area("💻 Code Editor", value=st.session_state.code, height=400)
    c1, c2 = st.columns([1, 1])
    if c1.button("▶️ Run Code"):
        if st.session_state.code.strip():
            out, err = execute_code(st.session_state.code)
            if err:
                st.error(f"Execution Error:\n{err}")
            elif out:
                st.subheader("Output:")
                st.code(out)
            else:
                st.success("Executed with no standard output.")
        else:
            st.warning("No code to run.")
    if c2.button("🗑️ Clear Code"):
        st.session_state.code = ''
        st.rerun()