app.py

from fastapi import FastAPI, HTTPException, BackgroundTasks, UploadFile, File, Form
from fastapi.responses import JSONResponse
from pathlib import Path
import tempfile
import uuid
import os
import aiohttp
import pdfplumber
from PIL import Image
import subprocess
from bs4 import BeautifulSoup
import io
import docx
from docx.shared import Inches
import logging
import base64
import time

# Setup
API_KEY = os.getenv("PDF_API_KEY")
ZAI_SECRET_API_KEY = os.getenv("ZAI_SECRET_API_KEY")
app = FastAPI()
api_tokens = {"client-1": API_KEY}
MAX_PDF_SIZE_MB = 40

# Logging
logging.basicConfig(level=logging.INFO)
logger = logging.getLogger(__name__)

@app.get("/health")
def health():
    return {"status": "ok"}

@app.post("/convert-pdf")
async def convert_pdf_endpoint(
        background_tasks: BackgroundTasks,
        file: UploadFile = File(...),
        client_id: str = Form(...),
        token: str = Form(...),
        callback_url: str = Form(...),
        book_id : int = Form(...)
):
    # Auth
    if client_id not in api_tokens or api_tokens[client_id] != token:
        raise HTTPException(status_code=401, detail="Invalid API credentials")

    # Check file type
    if not file.filename.lower().endswith(".pdf"):
        raise HTTPException(status_code=400, detail="Only PDF files are supported")

    # Read file
    try:
        pdf_bytes = await file.read()
    except Exception as e:
        logger.error(f"Failed to read uploaded file: {e}")
        raise HTTPException(status_code=400, detail="Unable to read file")

    # Enforce size limit
    if len(pdf_bytes) > MAX_PDF_SIZE_MB * 1024 * 1024:
        raise HTTPException(status_code=413, detail=f"PDF too large (> {MAX_PDF_SIZE_MB} MB)")

    # Write to temp file
    try:
        tmp_dir = Path(tempfile.mkdtemp())
        tmp_pdf_path = tmp_dir / f"{uuid.uuid4()}.pdf"
        tmp_pdf_path.write_bytes(pdf_bytes)
    except Exception as e:
        logger.error(f"Failed to write PDF file: {e}")
        raise HTTPException(status_code=500, detail="Internal error writing PDF")

    # Background processing
    task_id = str(uuid.uuid4())
    background_tasks.add_task(handle_pdf_processing, str(tmp_pdf_path), callback_url, task_id, book_id)

    logger.info(f"Started task {task_id} for client {client_id}")
    return JSONResponse({"status": "processing", "task_id": task_id})


async def handle_pdf_processing(pdf_path: str, callback_url: str, task_id: str, book_id:int):
    try:
        with open(pdf_path, "rb") as f:
            html, _ = extract_pdf_to_html(f)
    except Exception as e:
        logger.error(f"PDF processing failed: {e}")
        html = f"<p>Error extracting PDF: {e}</p>"
    finally:
        try:
            Path(pdf_path).unlink(missing_ok=True)
        except Exception as e:
            logger.warning(f"Temp file cleanup failed: {e}")

    try:
        headers = { "x-api-key": ZAI_SECRET_API_KEY }
        async with aiohttp.ClientSession() as session:
            await session.post(callback_url, json={
                "task_id": task_id,
                "content": html,
                "book_id": book_id
            }, headers=headers )
        logger.info(f"Content Generated \n {html} ")
        logger.info(f"Callback sent for task {task_id}")
    except Exception as e:
        logger.error(f"Callback failed for task {task_id}: {e}")


def extract_text_from_image(image: Image.Image) -> str:
    temp_img_path = Path(tempfile.mktemp(suffix=".png"))
    image.save(temp_img_path)
    try:
        result = subprocess.run(
            ["latexocr", str(temp_img_path)],
            capture_output=True,
            text=True
        )
        if result.returncode == 0:
            return f"<pre>\\[{result.stdout.strip()}\\]</pre>"
        else:
            return ""
    except Exception as e:
        return f"<i>LaTeX-OCR error: {str(e)}</i>"
    finally:
        try:
            temp_img_path.unlink(missing_ok=True)
        except Exception:
            pass


def extract_pdf_to_html(file) -> tuple[str, docx.Document]:
    html_output = ""
    docx_output = docx.Document()
    toc = []

    with pdfplumber.open(file) as pdf:
        for page_num, page in enumerate(pdf.pages):
            start = time.time()
            page_title = f"Page {page_num + 1}"
            toc.append(f"<li><a href='#page{page_num+1}'>{page_title}</a></li>")
            html_output += f"<h2 id='page{page_num+1}'>{page_title}</h2>\n"
            docx_output.add_heading(page_title, level=2)

            text = page.extract_text()
            if text:
                for line in text.split("\n"):
                    html_output += f"<p>{line}</p>\n"
                    docx_output.add_paragraph(line)
            else:
                html_output += "<p><i>No text detected on this page.</i></p>"

            for img_obj in page.images:
                try:
                    x0, top, x1, bottom = img_obj["x0"], img_obj["top"], img_obj["x1"], img_obj["bottom"]
                    cropped = page.crop((x0, top, x1, bottom)).to_image(resolution=300).original
                except Exception:
                    continue

                math_html = extract_text_from_image(cropped)
                if math_html.strip():
                    html_output += f"<div>{math_html}</div>\n"
                    docx_output.add_paragraph(BeautifulSoup(math_html, "html.parser").text)

                buffer = io.BytesIO()
                cropped.save(buffer, format="PNG")
                buffer.seek(0)
                b64_img = base64.b64encode(buffer.read()).decode("utf-8")
                html_output += f'<img src="data:image/png;base64,{b64_img}" style="width:100%; margin: 1rem 0;" />\n'

                buffer.seek(0)
                try:
                    docx_output.add_picture(buffer, width=Inches(5))
                except Exception:
                    pass
            logger.info(f"Processed page {page_num + 1} in {time.time() - start:.2f}s")
    full_html = f"<ul>{''.join(toc)}</ul>\n" + html_output
    return full_html, docx_output