pdfToMarkdownServer.py

server.py

import uvicorn
import os
import tempfile
from fastapi import FastAPI, UploadFile, File, HTTPException
from fastapi.responses import PlainTextResponse

# MinerU imports
from magic_pdf.data.data_reader_writer import FileBasedDataWriter
from magic_pdf.data.dataset import PymuDocDataset
from magic_pdf.model.doc_analyze_by_custom_model import doc_analyze
from magic_pdf.config.enums import SupportedPdfParseMethod

app = FastAPI(title="PDFTOMARKDOWN SERVER")

@app.post("/convert", response_class=PlainTextResponse)
async def convert_pdf(file: UploadFile = File(...)):
    # Validate upload
    if file.content_type != "application/pdf":
        raise HTTPException(status_code=400, detail="Invalid file type. Please upload a PDF.")
    pdf_bytes = await file.read()

    # Create dataset from bytes
    ds = PymuDocDataset(pdf_bytes)

    # Choose parse method
    if ds.classify() == SupportedPdfParseMethod.OCR:
        infer = ds.apply(doc_analyze, ocr=True, lang='en')
    else:
        infer = ds.apply(doc_analyze, ocr=False, lang='en')

    # Process pipeline in temporary workspace
    with tempfile.TemporaryDirectory() as tmp:
        img_dir = os.path.join(tmp, "images")
        os.makedirs(img_dir, exist_ok=True)
        writer = FileBasedDataWriter(img_dir)

        # Generate markdown
        if ds.classify() == SupportedPdfParseMethod.OCR:
            pipe = infer.pipe_ocr_mode(writer, lang='en')
        else:
            pipe = infer.pipe_txt_mode(writer, lang='en')

        markdown = pipe.get_markdown("images")

    return markdown

if __name__ == "__main__":
    uvicorn.run("server:app", host="0.0.0.0", port=8080)