#!/usr/bin/env python2.7
# -*- coding: utf-8 -*-
"""
pdfparanoia - pdf watermark removal tool

This is the command-line client. It accepts pdf formatted data either through
stdin/piping or by referencing a file in argv[0].
"""


if __name__ == "__main__":
    import sys
    import pdfparanoia
    import argparse

    try:
        from StringIO import StringIO
    except ImportError:
        from io import StringIO, BytesIO

    ArgP = argparse.ArgumentParser(description="pdfparanoia is a PDF watermark removal library for academic papers. Some publishers include private information like institution names, personal names, ip addresses, timestamps and other identifying information in watermarks on each page.")
    ArgP.add_argument('in_pdf', nargs='?', type=argparse.FileType('rb'),
                        default='-') # argparse.FileType interprets "-" as Stdin.
    ArgP.add_argument("-o", "--output", type=argparse.FileType('wb'),
                        default=sys.stdout)
    ArgP.add_argument("-v", "--verbose", action="store_true", default=False,
                      help="Output more information, which may be sensitive or excessive.")
    ArgP.add_argument("-V", "--more-verbose", action="store_true", default=False,
                      help="Output even more information. Implies -v.")
    Args = ArgP.parse_args()

    verbose = 0
    if Args.verbose: verbose = 1
    if Args.more_verbose: verbose = 2

    # I really don't like having to read a file only to cast as StringIO, but seems
    # necessary to get reading from StdIn to play nicely with pdfparanoia.
    outputcontent = pdfparanoia.scrub(StringIO(Args.in_pdf.read()), verbose=verbose)
    Args.in_pdf.close()
    Args.output.write(outputcontent)
    if Args.output != sys.stdout:
        Args.output.close()