app.py

import gradio as gr
from docling.document_converter import DocumentConverter
from llama_index.readers.docling import DoclingReader
import json
import tempfile
import os

def convert_document_docling(file, output_format):
    try:
        reader = DoclingReader(export_type=DoclingReader.ExportType.JSON if output_format == "JSON" else DoclingReader.ExportType.MARKDOWN)
        docs = reader.load_data(file_path=file.name)
        converted_text = docs[0].text
        temp_dir = tempfile.gettempdir()
        output_filename = os.path.splitext(os.path.basename(file.name))[0] + (".json" if output_format == "JSON" else ".md")
        output_path = os.path.join(temp_dir, output_filename)
        with open(output_path, 'w', encoding='utf-8') as f:
            f.write(converted_text)
        metadata = {
            "Filename": file.name,
            "File Size": f"{os.path.getsize(file.name) / 1024:.2f} KB",
            "Output Format": output_format,
            "Conversion Status": "Success",
            "Method": "llama-index-readers-docling"
        }
        return converted_text, metadata, output_path, gr.update(visible=True), "✅ Document converted successfully!"
    except Exception as e:
        error_metadata = {"Error": str(e), "Status": "Failed"}
        return "", error_metadata, None, gr.update(visible=False), "❌ Error during conversion"

def convert_document_original(file, output_format):
    try:
        converter = DocumentConverter()
        result = converter.convert(file.name)
        temp_dir = tempfile.gettempdir()
        if output_format == "Markdown":
            converted_text = result.document.export_to_markdown()
            file_extension = ".md"
        else:
            converted_text = result.document.export_to_json()
            file_extension = ".json"
        output_filename = os.path.splitext(os.path.basename(file.name))[0] + file_extension
        output_path = os.path.join(temp_dir, output_filename)
        with open(output_path, 'w', encoding='utf-8') as f:
            f.write(converted_text)
        metadata = {
            "Filename": file.name,
            "File Size": f"{os.path.getsize(file.name) / 1024:.2f} KB",
            "Output Format": output_format,
            "Conversion Status": "Success",
            "Method": "docling"
        }
        return converted_text, metadata, output_path, gr.update(visible=True), "✅ Document converted successfully!"
    except Exception as e:
        error_metadata = {"Error": str(e), "Status": "Failed"}
        return "", error_metadata, None, gr.update(visible=False), "❌ Error during conversion"


custom_css = """
:root {
    --primary-color: #2563eb;
    --secondary-color: #1e40af;
    --background-color: #1e1e1e;
    --card-background: #262626;
    --text-color: #ffffff;
    --border-radius: 10px;
}

body {
    background-color: var(--background-color);
    color: var(--text-color);
}

.container {
    max-width: 1200px;
    margin: 0 auto;
    padding: 2rem;
}

.gr-button {
    background: var(--primary-color) !important;
    border: none !important;
    color: white !important;
    padding: 10px 20px !important;
    border-radius: var(--border-radius) !important;
    transition: all 0.3s ease !important;
}

.gr-button:hover {
    background: var(--secondary-color) !important;
    transform: translateY(-2px);
    box-shadow: 0 4px 6px rgba(0, 0, 0, 0.1);
}

.gr-form {
    background-color: var(--card-background);
    padding: 2rem;
    border-radius: var(--border-radius);
    box-shadow: 0 4px 6px rgba(0, 0, 0, 0.1);
}

.gr-input, .gr-textbox {
    background-color: #333333 !important;
    border: 1px solid #404040 !important;
    color: var(--text-color) !important;
    border-radius: var(--border-radius) !important;
}

.gr-padded {
    padding: 1rem;
}

.gr-header {
    margin-bottom: 2rem;
    text-align: center;
}

.gr-subtitle {
    color: #9ca3af;
    font-size: 1.1rem;
    margin-bottom: 1.5rem;
}
"""

with gr.Blocks(css=custom_css) as app:
    gr.HTML(
        """
        <div class="gr-header">
            <h1 style='font-size: 2.5rem; color: #2563eb; margin-bottom: 1rem;'>📄 Docling Document Converter</h1>
            <p class="gr-subtitle">Transform your documents into Markdown or JSON format with ease</p>
        </div>
        """
    )
    with gr.Row():
        with gr.Column(scale=1):
            with gr.Group():
                gr.Markdown("### Input Settings")
                file_input = gr.File(
                    label="Upload Your Document",
                    file_types=[".doc", ".docx", ".pdf", ".txt"],
                    elem_classes="gr-input"
                )
                format_input = gr.Radio(
                    choices=["Markdown", "JSON"],
                    label="Output Format",
                    value="Markdown",
                    elem_classes="gr-input"
                )
                method_input = gr.Radio(
                    choices=["docling", "llama-index-readers-docling"],
                    label="Conversion Method",
                    value="docling",
                    elem_classes="gr-input"
                )
                convert_button = gr.Button(
                    "🔄 Convert Document",
                    variant="primary",
                    elem_classes=["gr-button"]
                )
            status_message = gr.Textbox(
                label="Status",
                interactive=False,
                visible=False,
                elem_classes="gr-padded"
            )
        with gr.Column(scale=2):
            with gr.Group():
                gr.Markdown("### Conversion Output")
                output_text = gr.Textbox(
                    label="Converted Content",
                    placeholder="The converted text will appear here...",
                    lines=15,
                    elem_classes="gr-textbox"
                )
                output_metadata = gr.JSON(
                    label="Document Metadata",
                    elem_classes="gr-input"
                )
                download_button = gr.File(
                    label="Download Converted File",
                    visible=False,
                    elem_classes="gr-padded"
                )

    with gr.Row():
        progress = gr.Slider(minimum=0, maximum=100, value=0, label="Progress", interactive=False)
    
    convert_button.click(
        fn=lambda file, format, method: convert_document_docling(file, format, progress) if method == "llama-index-readers-docling" else convert_document_original(file, format, progress),
        inputs=[file_input, format_input, method_input],
        outputs=[output_text, output_metadata, download_button, download_button, status_message, progress]
    )
    
    def update_progress(current, total):
        return (current / total) * 100
    
    def convert_document_docling(file, output_format, progress):
        try:
            reader = DoclingReader(export_type=DoclingReader.ExportType.JSON if output_format == "JSON" else DoclingReader.ExportType.MARKDOWN)
            docs = reader.load_data(file_path=file.name)
            converted_text = docs[0].text
            temp_dir = tempfile.gettempdir()
            output_filename = os.path.splitext(os.path.basename(file.name))[0] + (".json" if output_format == "JSON" else ".md")
            output_path = os.path.join(temp_dir, output_filename)
            with open(output_path, 'w', encoding='utf-8') as f:
                f.write(converted_text)
            metadata = {
                "Filename": file.name,
                "File Size": f"{os.path.getsize(file.name) / 1024:.2f} KB",
                "Output Format": output_format,
                "Conversion Status": "Success",
                "Method": "llama-index-readers-docling"
            }
            progress(100)  # Set progress to 100% once complete
            return converted_text, metadata, output_path, gr.update(visible=True), "✅ Document converted successfully!", 100
        except Exception as e:
            error_metadata = {"Error": str(e), "Status": "Failed"}
            return "", error_metadata, None, gr.update(visible=False), "❌ Error during conversion", 0
    
    def convert_document_original(file, output_format, progress):
        try:
            converter = DocumentConverter()
            result = converter.convert(file.name)
            temp_dir = tempfile.gettempdir()
            if output_format == "Markdown":
                converted_text = result.document.export_to_markdown()
                file_extension = ".md"
            else:
                converted_text = result.document.export_to_json()
                file_extension = ".json"
            output_filename = os.path.splitext(os.path.basename(file.name))[0] + file_extension
            output_path = os.path.join(temp_dir, output_filename)
            with open(output_path, 'w', encoding='utf-8') as f:
                f.write(converted_text)
            metadata = {
                "Filename": file.name,
                "File Size": f"{os.path.getsize(file.name) / 1024:.2f} KB",
                "Output Format": output_format,
                "Conversion Status": "Success",
                "Method": "docling"
            }
            progress(100)  # Set progress to 100% once complete
            return converted_text, metadata, output_path, gr.update(visible=True), "✅ Document converted successfully!", 100
        except Exception as e:
            error_metadata = {"Error": str(e), "Status": "Failed"}
            return "", error_metadata, None, gr.update(visible=False), "❌ Error during conversion", 0

app.launch(debug=True, share=True)