Support conversion from standard input

[yhatt]

This PR will support the conversion from standard input. It could chain process by pipe.

echo "# Markdown" | marp > output.html

By using curl, it will become that can convert Marp Markdown on network resource easily.

# Convert Marp Markdown on Gist
curl https://gist.githubusercontent.com/yhatt/a352e6087c2d5185113b9143f330fb68/raw/1cd8ff1f8b0c0cff39b9273e6ce471240d989de5/uncover.md \
  | marp -o uncover.pdf

The converted buffer came from stdin would output to stdout if you not specified --output option. You can continue processing by piping output to another tools.

For example, you can extract specified pages from converted PDF by using PDFtk server.

# Extract first 3 pages
marp input.md --pdf | pdftk - cat 1-3 output ./output.pdf

[yhatt]

By this change, we no longer restrict to output PDF into stdout. The default output path will decide by where came input from.