app.py

from fastapi import FastAPI, UploadFile, File
from fastapi.responses import StreamingResponse
import fitz  # PyMuPDF
import pytesseract
from PIL import Image
from io import BytesIO

app = FastAPI()

@app.get("/")
def home():
    return {
        "message": "OCR API is running",
        "endpoint": "/ocr",
        "method": "POST",
        "upload": "PNG, JPG, JPEG, PDF",
        "output": "Searchable PDF"
    }


def ocr_image_to_pdf(image_bytes: bytes):
    img = Image.open(BytesIO(image_bytes))
    pdf_bytes = pytesseract.image_to_pdf_or_hocr(img, extension='pdf')
    return pdf_bytes


def ocr_pdf_to_searchable(pdf_bytes: bytes):
    original_pdf = fitz.open(stream=pdf_bytes, filetype="pdf")
    ocr_output = fitz.open()

    for page_num in range(len(original_pdf)):
        page = original_pdf.load_page(page_num)
        pix = page.get_pixmap(dpi=300)
        img_bytes = pix.tobytes("png")

        ocr_pdf_page = pytesseract.image_to_pdf_or_hocr(img_bytes, extension="pdf")
        ocr_output.insert_pdf(fitz.open("pdf", ocr_pdf_page))

    output_bytes = ocr_output.tobytes()
    ocr_output.close()
    return output_bytes


@app.post("/ocr")
async def upload_and_ocr(file: UploadFile = File(...)):
    file_bytes = await file.read()
    filename = file.filename.lower()

    if filename.endswith((".png", ".jpg", ".jpeg")):
        pdf_bytes = ocr_image_to_pdf(file_bytes)
    elif filename.endswith(".pdf"):
        pdf_bytes = ocr_pdf_to_searchable(file_bytes)
    else:
        return {"error": "Unsupported file type"}

    return StreamingResponse(
        BytesIO(pdf_bytes),
        media_type="application/pdf",
        headers={
            "Content-Disposition": f"attachment; filename=searchable_{file.filename}.pdf"
        }
    )