pdf.py

from fastapi import FastAPI, UploadFile, File, HTTPException
from pydantic import BaseModel
import fitz  # PyMuPDF for PDF processing
import pdfplumber
from fpdf import FPDF
import json
import os

app = FastAPI()


class PDFContent(BaseModel):
    content: str


@app.post("/pdf-to-json")
async def pdf_to_json(file: UploadFile = File(...)):
    """Converts a PDF file to a structured JSON format."""
    try:
        pdf_data = {}
        with pdfplumber.open(file.file) as pdf:
            for i, page in enumerate(pdf.pages):
                extracted_text = page.extract_text()
                pdf_data[f"page_{i+1}"] = extracted_text if extracted_text else ""

        return {"filename": file.filename, "content": pdf_data}
    except Exception as e:
        raise HTTPException(status_code=500, detail=f"Error processing PDF: {str(e)}")


@app.post("/extract-text")
async def extract_text(file: UploadFile = File(...)):
    """Extracts raw text from a PDF file."""
    try:
        text = ""
        file_content = await file.read()
        with fitz.open(stream=file_content, filetype="pdf") as doc:
            for page in doc:
                text += page.get_text("text") + "\n"

        return {"filename": file.filename, "text": text}
    except Exception as e:
        raise HTTPException(status_code=500, detail=f"Error extracting text: {str(e)}")


@app.post("/generate-pdf")
async def generate_pdf(pdf_content: PDFContent):
    """Generates a PDF with the provided text content."""
    try:
        pdf = FPDF()
        pdf.add_page()
        pdf.set_font("Arial", size=12)
        pdf.multi_cell(0, 10, pdf_content.content)

        filename = "./generated_pdfs/generated.pdf"
        pdf.output(filename)

        return {"message": "PDF generated successfully", "file": filename}
    except Exception as e:
        raise HTTPException(status_code=500, detail=f"Error generating PDF: {str(e)}")


@app.get("/docs")
def api_docs():
    """Returns API documentation links."""
    return {"docs": "/docs", "redoc": "/redoc"}


if __name__ == "__main__":
    import uvicorn

    uvicorn.run(app, host="0.0.0.0", port=8000)