pdf2audio.py

from PyPDF2 import PdfReader, PdfWriter
from gtts import gTTS
import sys

# python3 main.py --pdf Documents/book.pdf --range 30:31 --output preface.mp3

_opts = [opt for opt in sys.argv[1:] if opt.startswith("-")]
args = [arg for arg in sys.argv[1:] if not arg.startswith("-")]

pdfFile, rangePages, audioFile = args

with open(pdfFile, "rb") as pdf:
    reader = PdfReader(pdf)
    info = reader.metadata
    print(info)
    text = ""
    startPage,endPage = rangePages.split(':')

    for i in range(int(startPage),int(endPage)):
        selected_page = reader.pages[i]
        text += selected_page.extract_text().replace("\n"," ")

    pytts = gTTS(text, lang='en', tld='us')
    pytts.save(audioFile)