import io
import logging
import re

from fastapi import FastAPI, HTTPException
from fastapi.middleware.cors import CORSMiddleware
from fastapi.responses import JSONResponse, StreamingResponse
from pydantic import BaseModel, Field
from weasyprint import HTML

# Initialize logging
logging.basicConfig(level=logging.INFO)


class PdfRequest(BaseModel):
    html: str = Field(..., min_length=1, description="HTML content to convert to PDF")
    filename: str | None = None


app = FastAPI()

app.add_middleware(
    CORSMiddleware,
    allow_origins=["*"],
    allow_credentials=True,
    allow_methods=["*"],
    allow_headers=["*"],
)


# Compile the regular expression once
FILENAME_SANITIZE_REGEX = re.compile(r"[^a-zA-Z0-9_\-]")


def sanitize_filename(filename: str) -> str:
    """
    Sanitize the filename by replacing invalid characters with underscores.

    Args:
        filename (str): The original filename.

    Returns:
        str: The sanitized filename.
    """
    return FILENAME_SANITIZE_REGEX.sub("_", filename)


async def pdf_generator(byte_string: bytes):
    """
    Generator to yield PDF content in chunks.

    Args:
        byte_string (bytes): The PDF content as bytes.

    Yields:
        bytes: Chunks of the PDF content.
    """
    byte_stream = io.BytesIO(byte_string)
    chunk = byte_stream.read(4096)
    while chunk:
        yield chunk
        chunk = byte_stream.read(4096)


@app.post("/pdf")
async def pdf(body: PdfRequest):
    """
    Endpoint to convert HTML content to a PDF file.

    Args:
        body (PdfRequest): Request body containing HTML content and an optional filename.

    Returns:
        StreamingResponse: A streaming response with the generated PDF file.
    """
    logging.info("Received request to generate PDF")
    try:
        byte_string = HTML(string=body.html).write_pdf()
    except Exception as e:
        logging.error(f"Error generating PDF: {e}")
        raise HTTPException(status_code=400, detail="Invalid HTML input")

    filename = sanitize_filename(body.filename.strip() if body.filename else "hinpdof")

    headers = {
        "Content-Type": "application/pdf",
        "Content-Disposition": f'attachment; filename="{filename}.pdf"',
    }

    logging.info(f"PDF generated successfully: {filename}.pdf")
    return StreamingResponse(pdf_generator(byte_string), headers=headers)


@app.get("/health")
async def health_check():
    """
    Endpoint to check the health status of the application.

    Returns:
        JSONResponse: A JSON response with the status of the application.
    """
    return JSONResponse(content={"status": "ok"})