app.py

import io
import os
import re
import glob
import textwrap
import base64
import sys
from datetime import datetime
from pathlib import Path
from contextlib import redirect_stdout

import streamlit as st
import pandas as pd
from PIL import Image
from reportlab.pdfgen import canvas
from reportlab.lib.pagesizes import letter
from reportlab.lib.utils import ImageReader
from reportlab.pdfbase import pdfmetrics
from reportlab.pdfbase.ttfonts import TTFont
import mistune
from gtts import gTTS

# --- Helper Functions ---

# 🗑️ Deletes a specified file and reruns the app.
def delete_asset(path):
    """Safely deletes a file if it exists and reruns the Streamlit app."""
    try:
        os.remove(path)
    except OSError as e:
        st.error(f"Error deleting file {path}: {e}")
    st.rerun()

# 📥 Gets text input from either a file upload or a text area.
def get_text_input(file_uploader_label, accepted_types, text_area_label):
    """
    Provides UI for uploading a text file or entering text manually.
    Returns the text content and a filename stem.
    """
    md_text = ""
    stem = datetime.now().strftime('%Y%m%d_%H%M%S')

    uploaded_file = st.file_uploader(file_uploader_label, type=accepted_types)
    if uploaded_file:
        md_text = uploaded_file.getvalue().decode("utf-8")
        stem = Path(uploaded_file.name).stem
    else:
        md_text = st.text_area(text_area_label, height=200, value="## Your Markdown Here\n\nEnter your markdown text, or upload a file.")

    # Convert markdown to plain text for processing
    renderer = mistune.HTMLRenderer()
    markdown = mistune.create_markdown(renderer=renderer)
    html = markdown(md_text or "")
    plain_text = re.sub(r'<[^>]+>', '', html)

    return plain_text, stem

# 🗣️ Generates an MP3 voice file from text using gTTS.
def generate_voice_file(text, lang, is_slow, filename):
    """
    Creates an MP3 from text, saves it, and provides it for playback and download.
    """
    if not text.strip():
        st.warning("No text to generate voice from.")
        return

    voice_file_path = f"{filename}.mp3"
    try:
        tts = gTTS(text=text, lang=lang, slow=is_slow)
        tts.save(voice_file_path)
        st.audio(voice_file_path)
        with open(voice_file_path, 'rb') as fp:
            st.download_button("📥 Download MP3", data=fp, file_name=voice_file_path, mime="audio/mpeg")
    except Exception as e:
        st.error(f"Failed to generate audio: {e}")

# 📄 Creates a PDF document from text and images.
def generate_pdf(text_content, images, pdf_params):
    """
    Generates a PDF buffer from text and a list of images based on specified parameters.
    """
    buf = io.BytesIO()
    
    # --- Register Custom Fonts ---
    for font_path in pdf_params.get('ttf_files', []):
        try:
            font_name = Path(font_path).stem
            pdfmetrics.registerFont(TTFont(font_name, font_path))
        except Exception as e:
            st.warning(f"Could not register font {font_path}: {e}")

    c = canvas.Canvas(buf)
    page_w, page_h = letter
    margin = 40
    gutter = 20
    col_w = (page_w - 2 * margin - (pdf_params['columns'] - 1) * gutter) / pdf_params['columns']
    
    # Use registered font name, which is the stem of the file path
    font_name_to_use = Path(pdf_params['font_family']).stem if ".ttf" in pdf_params['font_family'] else pdf_params['font_family']
    c.setFont(font_name_to_use, pdf_params['font_size'])
    
    line_height = pdf_params['font_size'] * 1.2
    # Estimate characters per line for wrapping
    # 0.6 is a common factor for average character width vs font size
    wrap_width = int(col_w / (pdf_params['font_size'] * 0.6)) if pdf_params['font_size'] > 0 else 50


    y = page_h - margin
    col_idx = 0

    # --- Render Text ---
    for paragraph in text_content.split("\n"):
        wrapped_lines = textwrap.wrap(paragraph, wrap_width) if paragraph.strip() else [""]
        for line in wrapped_lines:
            if y < margin:
                col_idx += 1
                if col_idx >= pdf_params['columns']:
                    c.showPage()
                    c.setFont(font_name_to_use, pdf_params['font_size'])
                    col_idx = 0
                y = page_h - margin
            
            x = margin + col_idx * (col_w + gutter)
            c.drawString(x, y, line)
            y -= line_height
        y -= line_height # Add extra space for paragraph breaks

    # --- Render Images ---
    for img_file in images:
        try:
            # Handle both file paths and uploaded file objects
            img = Image.open(img_file)
            w, h = img.size
            c.showPage()
            c.setPageSize((w, h))
            c.drawImage(ImageReader(img), 0, 0, w, h, preserveAspectRatio=True, mask='auto')
        except Exception as e:
            img_name = img_file.name if hasattr(img_file, 'name') else img_file
            st.warning(f"Could not process image {img_name}: {e}")
            continue

    c.save()
    buf.seek(0)
    return buf

# 🗂️ Displays a list of generated assets with download/delete options.
def show_asset_manager():
    """Scans for local files and displays them with management controls."""
    st.markdown("---")
    st.subheader("📂 Available Assets")
    assets = sorted(glob.glob("*.*"))
    if not assets:
        st.info("No assets generated yet.")
        return

    for asset_path in assets:
        # Avoid showing the script itself
        if asset_path.endswith('.py'):
            continue
            
        ext = Path(asset_path).suffix.lower()
        cols = st.columns([3, 1, 1])
        cols[0].write(f"`{asset_path}`")
        
        try:
            with open(asset_path, 'rb') as fp:
                file_bytes = fp.read()
            
            if ext == '.pdf':
                cols[1].download_button("📥", data=file_bytes, file_name=asset_path, mime="application/pdf", key=f"dl_{asset_path}")
            elif ext == '.mp3':
                cols[1].audio(file_bytes)
            elif ext in ['.png', '.jpg', '.jpeg']:
                 cols[1].image(file_bytes, width=60)
        except Exception as e:
            cols[1].error("Error")

        cols[2].button("🗑️", key=f"del_{asset_path}", on_click=delete_asset, args=(asset_path,))

# 🧩 Shows a demo of how to use the functions with file lists
def show_batch_processing_demo(pdf_params):
    """Finds local files and shows how to process them."""
    st.markdown("---")
    st.subheader("🧩 Batch Processing Demo")
    st.info("This section demonstrates how you could call the PDF generation function programmatically with lists of files.")

    md_files = glob.glob("*.md")
    img_files = glob.glob("*.png") + glob.glob("*.jpg")

    if not md_files or not img_files:
        st.warning("To run the demo, please ensure there is at least one `.md` file and one image (`.png`/`.jpg`) in the directory.")
        return

    st.write("Found the following files to use for the demo:")
    st.write(f"**Markdown file:** `{md_files[0]}`")
    st.write(f"**Image files:** `{', '.join(img_files)}`")

    if st.button("🧪 Run Demo with Above Files"):
        md_file_str = md_files[0]
        img_files_str = ",".join(img_files)
        
        # --- Example of programmatic execution ---
        # 1. Read the markdown file
        with open(md_file_str, 'r') as f:
            text_content = f.read()
        
        # 2. Open the image files (generate_pdf expects file-like objects or paths)
        image_objects = img_files # Pass paths directly

        # 3. Call the generator function
        pdf_buffer = generate_pdf(text_content, image_objects, pdf_params)
        
        # 4. Provide for download
        st.download_button(
            "⬇️ Download Batch Demo PDF",
            data=pdf_buffer,
            file_name="batch_demo_output.pdf",
            mime="application/pdf"
        )
        st.success("Batch demo PDF generated!")


# 🐍 Renders the entire UI and logic for the Python code interpreter.
def render_code_interpreter():
    """Sets up the UI and execution logic for the code interpreter tab."""
    st.header("🧪 Python Code Executor & Demo")

    # --- Nested Helper Functions for this Tab ---
    def extract_python_code(md_text):
        return re.findall(r"```python\s*(.*?)```", md_text, re.DOTALL)

    def execute_code(code_str):
        output_buffer = io.StringIO()
        try:
            # The exec function will have access to globally imported libraries
            exec_globals = {
                "st": st,
                "glob": glob,
                "base64": base64,
                "io": io,
                "canvas": canvas,
                "letter": letter
            }
            with redirect_stdout(output_buffer):
                exec(code_str, exec_globals)
            return output_buffer.getvalue(), None
        except Exception as e:
            return None, str(e)

    # --- Main Logic for the Tab ---
    DEFAULT_CODE = """
import streamlit as st
import glob
import base64
import io
from reportlab.pdfgen import canvas
from reportlab.lib.pagesizes import letter

st.title("📊 Enhanced Demo App")
st.markdown("This demo shows file galleries and base64 PDF downloads.")

# --- Image Gallery ---
with st.expander("🖼️ Show Image Files in Directory"):
    image_files = glob.glob("*.png") + glob.glob("*.jpg")
    if not image_files:
        st.write("No image files found.")
    else:
        st.image(image_files)

# --- PDF Gallery ---
with st.expander("📄 Show PDF Files in Directory"):
    pdf_files = glob.glob("*.pdf")
    if not pdf_files:
        st.write("No PDF files found.")
    else:
        st.write(pdf_files)

# --- PDF Generation and Download ---
if st.button("Generate Demo PDF & Download Link"):
    # 1. Create PDF in memory
    buffer = io.BytesIO()
    p = canvas.Canvas(buffer, pagesize=letter)
    p.drawString(100, 750, "This is a demo PDF generated from the code interpreter.")
    p.showPage()
    p.save()
    
    # 2. Encode PDF to Base64
    b64 = base64.b64encode(buffer.getvalue()).decode()
    
    # 3. Create download link
    href = f'<a href="data:application/pdf;base64,{b64}" download="demo_from_code.pdf">Download Generated PDF</a>'
    st.markdown(href, unsafe_allow_html=True)
    st.success("PDF generated! Click the link above to download.")
"""
    if 'code' not in st.session_state:
        st.session_state.code = DEFAULT_CODE

    uploaded_file = st.file_uploader("Upload .py or .md", type=['py', 'md'], key="code_uploader")
    if uploaded_file:
        file_content = uploaded_file.getvalue().decode()
        if uploaded_file.type == 'text/markdown':
            codes = extract_python_code(file_content)
            st.session_state.code = codes[0] if codes else ''
        else:
            st.session_state.code = file_content
    
    st.session_state.code = st.text_area("💻 Code Editor", value=st.session_state.code, height=400)

    c1, c2 = st.columns(2)
    if c1.button("▶️ Run Code", use_container_width=True):
        output, err = execute_code(st.session_state.code)
        st.subheader("Output")
        if err:
            st.error(err)
        if output: # Show output even if empty, to confirm it ran
            st.code(output, language='text')
        if not err:
            st.success("Executed successfully.")

    if c2.button("🗑️ Clear Code", use_container_width=True):
        st.session_state.code = ''
        st.rerun()

# --- Main App ---
def main():
    """Main function to run the Streamlit application."""
    st.set_page_config(page_title="PDF & Code Interpreter", layout="wide", page_icon="🚀")

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

    with tab1:
        st.header("📄 PDF Composer & Voice Generator 🚀")

        # --- Sidebar Controls ---
        st.sidebar.title("PDF Settings")
        
        # --- Dynamic Font Loading ---
        ttf_files = glob.glob("*.ttf")
        standard_fonts = ["Helvetica", "Times-Roman", "Courier"]
        available_fonts = ttf_files + standard_fonts
        default_font_index = 0 if ttf_files else 0 # Default to first ttf or Helvetica

        pdf_params = {
            'columns': st.sidebar.slider("Text columns", 1, 3, 1),
            'font_family': st.sidebar.selectbox("Font", available_fonts, index=default_font_index),
            'font_size': st.sidebar.slider("Font size", 6, 48, 12),
            'ttf_files': ttf_files
        }

        # --- Main UI ---
        plain_text, filename_stem = get_text_input(
            "Upload Markdown (.md)", ["md"], "Or enter markdown text directly"
        )

        st.subheader("🗣️ Voice Generation")
        languages = {"English (US)": "en", "English (UK)": "en-uk", "Spanish": "es"}
        voice_choice = st.selectbox("Voice Language", list(languages.keys()))
        slow_speech = st.checkbox("Slow Speech")

        if st.button("🔊 Generate Voice MP3"):
            generate_voice_file(plain_text, languages[voice_choice], slow_speech, filename_stem)

        st.subheader("🖼️ Image Upload")
        uploaded_images = st.file_uploader(
            "Upload Images for PDF", type=["png", "jpg", "jpeg"], accept_multiple_files=True
        )
        
        ordered_images = []
        if uploaded_images:
            df_imgs = pd.DataFrame([{"name": f.name, "order": i} for i, f in enumerate(uploaded_images)])
            edited_df = st.data_editor(df_imgs, use_container_width=True, key="img_order_editor")
            
            image_map = {f.name: f for f in uploaded_images}
            
            for _, row in edited_df.sort_values("order").iterrows():
                if row['name'] in image_map:
                    ordered_images.append(image_map[row['name']])

        st.subheader("🖋️ PDF Generation")
        if st.button("Generate PDF from UI"):
            if not plain_text.strip() and not ordered_images:
                 st.warning("Please provide some text or images to generate a PDF.")
            else:
                pdf_buffer = generate_pdf(plain_text, ordered_images, pdf_params)
                st.download_button(
                    "⬇️ Download PDF",
                    data=pdf_buffer,
                    file_name=f"{filename_stem}.pdf",
                    mime="application/pdf"
                )
        
        show_asset_manager()
        show_batch_processing_demo(pdf_params)

    with tab2:
        render_code_interpreter()

if __name__ == "__main__":
    main()