src/app.py

import os
from pathlib import Path
from urllib.parse import urlparse

import convertapi
import gradio as gr
from dotenv import load_dotenv

load_dotenv()

convertapi.api_credentials = os.getenv("CONVERTAPI_TOKEN")

if not convertapi.api_credentials:
    raise ValueError("CONVERTAPI_TOKEN is required")

# Configuration constants
MAX_FILE_SIZE = int(os.getenv("MAX_FILE_SIZE", 10 * 1024 * 1024))  # 10MB default
TEMP_DIR = os.getenv("TEMP_DIR", "temp")


# Get all supported formats from ConvertAPI
def get_supported_formats():
    try:
        # Common formats that ConvertAPI supports
        return [
            "pdf",
            "docx",
            "doc",
            "txt",
            "rtf",
            "odt",
            "html",
            "epub",
            "png",
            "jpg",
            "jpeg",
            "gif",
            "bmp",
            "tiff",
            "webp",
            "svg",
            "xlsx",
            "xls",
            "csv",
            "ods",
            "pptx",
            "ppt",
            "odp",
            "mp4",
            "avi",
            "mov",
            "wmv",
            "flv",
            "webm",
            "mp3",
            "wav",
            "flac",
            "aac",
            "ogg",
        ]
    except Exception:
        # Fallback to basic formats
        return ["pdf", "docx", "txt", "png", "jpg", "xlsx", "pptx"]


SUPPORTED_FORMATS = get_supported_formats()


# Function to convert a local file
def convert_local_file(local_file_path: str, output_format: str) -> str:
    """
    Convert a local document to a different format.

    Args:
        local_file_path (str): The path to the input file provided by Gradio.
        output_format (str): The format to convert the file to.

    Returns:
        str: The path to the converted file or an error message.
    """
    try:
        # Input validation
        if not local_file_path or not os.path.exists(local_file_path):
            error_msg = "Invalid or non-existent local file path."
            return error_msg

        if output_format not in SUPPORTED_FORMATS:
            error_msg = f"Unsupported output format: {output_format}"
            return error_msg

        # Check file size
        file_size = os.path.getsize(local_file_path)
        if file_size > MAX_FILE_SIZE:
            error_msg = f"File too large: {file_size / (1024 * 1024):.1f}MB (max {MAX_FILE_SIZE / (1024 * 1024):.0f}MB)"
            return error_msg

        # Convert the local file
        result = convertapi.convert(output_format, {"File": local_file_path})

        # Derive filename from local path
        input_filename = Path(local_file_path).stem

        # Save the converted file to a temporary location
        output_filename = f"{input_filename}.{output_format}"
        output_path = Path(TEMP_DIR) / output_filename
        output_path.parent.mkdir(exist_ok=True)

        result.file.save(str(output_path))

        # Return the file path string on success
        return str(output_path)

    except Exception as e:
        error_msg = f"Error converting file: {str(e)}"
        return error_msg


# Function to convert a file from a URL
def convert_from_url(document_url: str, output_format: str) -> str:
    """
    Convert a document from a URL to a different format.

    Args:
        document_url (str): The URL of the input file.
        output_format (str): The format to convert the file to.

    Returns:
        str: The path to the converted file or an error message.
    """
    try:
        # Input validation
        if not document_url or not (
            document_url.lower().startswith("http://")
            or document_url.lower().startswith("https://")
        ):
            error_msg = "Invalid or unsupported URL format."
            return error_msg

        if output_format not in SUPPORTED_FORMATS:
            error_msg = f"Unsupported output format: {output_format}"
            return error_msg

        # Convert the file from a URL
        result = convertapi.convert(output_format, {"File": document_url})

        # Attempt to derive a filename from the URL for saving
        try:
            url_path = urlparse(document_url).path
            input_filename = (
                Path(url_path).stem or "converted_file"
            )  # Fallback for complex URLs or no path
        except Exception:
            input_filename = "converted_file"

        # Save the converted file to a temporary location
        output_filename = f"{input_filename}.{output_format}"
        output_path = Path(TEMP_DIR) / output_filename
        output_path.parent.mkdir(exist_ok=True)

        result.file.save(str(output_path))

        # Return the file path string on success
        return str(output_path)

    except Exception as e:
        error_msg = f"Error converting file from URL: {str(e)}"
        return error_msg


# Create Gradio interfaces
local_file_demo = gr.Interface(
    fn=convert_local_file,
    inputs=[
        gr.File(label="Upload File"),
        gr.Dropdown(
            choices=SUPPORTED_FORMATS,
            label="Output Format",
            value="pdf",
        ),
    ],
    outputs=gr.File(label="Download Converted File"),
    description="Convert documents from your local file system to different formats.",
)

url_demo = gr.Interface(
    fn=convert_from_url,
    inputs=[
        gr.Textbox(
            label="Document URL",
            placeholder="Enter the URL of the document to convert (e.g., https://example.com/document.pdf)",
            value="",
        ),
        gr.Dropdown(
            choices=SUPPORTED_FORMATS,
            label="Output Format",
            value="pdf",
        ),
    ],
    outputs=gr.File(label="Download Converted File"),
    description="Download and convert documents directly from URLs to different formats.",
)

# Create tabbed interface
demo = gr.TabbedInterface(
    [local_file_demo, url_demo],
    ["Local Files", "Online Files"],
    title="Universal Document Converter",
)

if __name__ == "__main__":
    # Launch without MCP server to avoid initialization issues
    demo.launch(share=False, mcp_server=True, debug=True)