Move argparse object setup out of pdf2txt.py main to separate function.

[wlbentley]

As preparation for implementing Sphinx documentation, create a
separate maketheparser() function that builds and returns the
argparse.ArgumentParser() object.

Move import argparse out of main to the top of the file.

Add add_help=True to the parser.

The updated script parsed the samples/simple1.pdf file correctly,
as a basic test.

[tataganesh]

If this has been tested, I can merge it, since the main was already using Argparse.

[wlbentley]

1. Running pdf2txt.py outputs the short help, as expected, and throws an error for the missing filename:

pdf2txt.py: error: the following arguments are required: files

2. Running pdf2txt.py -h outputs the full help, as expected, without errors.
3. Running pdf2txt.py samples/simple1.pdf works as expected.
4. Running pdf2txt.py -o tempout samples/simple1.pdf places the output as text in the file tempout, as expected.

The automated tests in tests/test_tools_pdf2txt.py don't seem to test any more options than that.

To the best of my knowledge, the Argparse code now under the maketheparser() function was not altered while moving it outside of main().

So, I think it's ready to merge.

[tataganesh]

Looks good!