escapeURI

#!/usr/bin/env python3
#
# escapeURI: %-escape a string as needed in a URI.
# 2024-03-11: Written by Steven J. DeRose.
#
import sys
import urllib.parse

__metadata__ = {
    "title"        : "escapeURI",
    "description"  : "%-escape a string as needed in a URI.",
    "rightsHolder" : "Steven J. DeRose",
    "creator"      : "http://viaf.org/viaf/50334488",
    "type"         : "http://purl.org/dc/dcmitype/Software",
    "language"     : "Python 3.9",
    "created"      : "2024-03-11",
    "modified"     : "2024-03-11",
    "publisher"    : "http://github.com/sderose",
    "license"      : "https://creativecommons.org/licenses/by-sa/3.0/"
}
__version__ = __metadata__["modified"]

descr = """
=Name=
escapeURI: %-escape a string as needed in a URI.


=Description=

Converts a string(s) to UTF-8 and insert needed %-escapes for it to go ok
inside a URI.

Use --plus to also turn spaces into "+" signs.

Letters, digits, and '_.-~' are never quoted. Use --safe '...' to
make other characters (such as "/" for scheme and directory separators, ":",
"#" for fragment separator, "?" for queries, and so on.

==Usage==

    escapeURI [options] [string]
  OR
    escapeURI [options]

To escape a single string, put it on the command line (quoting if needed to
avoid shell issues).

If no such string is found there, then each line of STDIN will be read and escaped.

This uses Python's ''urllib.parse.quote_plus()'', q.v.


=See also=

My `unescapeURI`.


=History=

* 2024-03-11: Written by Steven J. DeRose.


=Rights=

Written 2024-03-11 by Steven J. DeRose. This work is dedicated to the public domain.

=Options=
"""


###############################################################################
# Main
#
def doit(s:str) -> str:
    if (args.plus):
        return urllib.parse.quote_plus(s, safe=args.safe)
    else:
        return urllib.parse.quote(s, safe=args.safe)

if __name__ == "__main__":
    import argparse

    def processOptions() -> argparse.Namespace:
        try:
            from BlockFormatter import BlockFormatter
            parser = argparse.ArgumentParser(
                description=descr, formatter_class=BlockFormatter)
        except ImportError:
            parser = argparse.ArgumentParser(description=descr)

        parser.add_argument(
            "--plus", action="store_true",
            help="Encode spaces as plus-signs.")
        parser.add_argument(
            "--safe", type=str, default="/",
            help="List of characters NOT to escape (default: '/').")
        parser.add_argument(
            "--version", action="version", version=__version__,
            help="Display version information, then exit.")

        parser.add_argument(
            "rest", type=str, nargs='...', #argparse.REMAINDER,
            help="A string to escape.")

        args0 = parser.parse_args()
        return(args0)

    args = processOptions()

    #print(repr(args.rest))

    if (args.rest):
        rest = "".join(args.rest)
        theBytes = bytes(rest.encode('utf-8'))
        print(doit(theBytes))
    else:
        while (True):
            rec = sys.stdin.readline()
            if (not rec): break
            theBytes = bytes(rec.encode('utf-8'))
            print(doit(theBytes))