Most Spanish N43 files will use LATIN-1 encoding not pure ASCII

[Julien Moutte]

Norm43 files from BancSabadell will contain extended ASCII characters with accents. When using the traditional open with ASCII codecs those characters will not be decodable.

Also when converting to OFX, the UTF-8 charset should be used instead of ASCII to allow for those.

[Julien Moutte]

Extend argparse file type to support encoding option. Open incoming file using latin-1 instead of ascii and output file using UTF-8. Reflect the charset in the OFX header as well.

[wmj]

Thank you for all your contributions!