+def parse_command(command):
+ r"""parse_command(command) -> (args, kwargs) :: Parse a command.
+
+ This function parses a command and split it into a list of parameters. It
+ has a similar to bash commandline parser. Spaces are the basic token
+ separator but you can group several tokens into one by using (single or
+ double) quotes. You can escape the quotes with a backslash (\' and \"),
+ express a backslash literal using a double backslash (\\), use special
+ meaning escaped sequences (like \a, \n, \r, \b, \v) and use unescaped
+ single quotes inside a double quoted token or vice-versa. A special escape
+ sequence is provided to express a NULL/None value: \N and it should appear
+ always as a separated token.
+
+ Additionally it accepts keyword arguments. When an (not-escaped) equal
+ sign (=) is found, the argument is considered a keyword, and the next
+ argument it's interpreted as its value.
+
+ This function returns a tuple containing a list and a dictionary. The
+ first has the positional arguments, the second, the keyword arguments.
+
+ There is no restriction about the order, a keyword argument can be
+ followed by a positional argument and vice-versa. All type of arguments
+ are grouped in the list/dict returned. The order of the positional
+ arguments is preserved and if there are multiple keyword arguments with
+ the same key, the last value is the winner (all other values are lost).
+
+ Examples:
+
+ >>> parse_command('hello world')
+ ([u'hello', u'world'], {})
+ >>> parse_command('hello planet=earth')
+ ([u'hello'], {'planet': u'earth'})
+ >>> parse_command('hello planet="third rock from the sun"')
+ ([u'hello'], {'planet': u'third rock from the sun'})
+ >>> parse_command(u' planet="third rock from the sun" hello ')
+ ([u'hello'], {'planet': u'third rock from the sun'})
+ >>> parse_command(u' planet="third rock from the sun" "hi, hello"'
+ '"how are you" ')
+ ([u'hi, hello', u'how are you'], {'planet': u'third rock from the sun'})
+ >>> parse_command(u'one two three "fourth number"=four')
+ ([u'one', u'two', u'three'], {'fourth number': u'four'})
+ >>> parse_command(u'one two three "fourth number=four"')
+ ([u'one', u'two', u'three', u'fourth number=four'], {})
+ >>> parse_command(u'one two three fourth\=four')
+ ([u'one', u'two', u'three', u'fourth=four'], {})
+ >>> parse_command(u'one two three fourth=four=five')
+ ([u'one', u'two', u'three'], {'fourth': u'four=five'})
+ >>> parse_command(ur'nice\nlong\n\ttext')
+ ([u'nice\nlong\n\ttext'], {})
+ >>> parse_command('=hello')
+ ([u'=hello'], {})
+ >>> parse_command(r'\thello')
+ ([u'\thello'], {})
+ >>> parse_command(r'\N')
+ ([None], {})
+ >>> parse_command(r'none=\N')
+ ([], {'none': None})
+ >>> parse_command(r'\N=none')
+ ([], {'\\N': 'none'})
+ >>> parse_command(r'Not\N')
+ ([u'Not\\N'], {})
+ >>> parse_command(r'\None')
+ ([u'\\None'], {})
+
+ This examples are syntax errors:
+ Missing quote: "hello world
+ Missing value: hello=
+ """
+ SEP, TOKEN, DQUOTE, SQUOTE, EQUAL = u' ', None, u'"', u"'", u'=' # states
+ separators = (u' ', u'\t', u'\v', u'\n') # token separators
+ escaped_chars = (u'a', u'n', u'r', u'b', u'v', u't') # escaped sequences
+ seq = []
+ dic = {}
+ buff = u''
+ escape = False
+ keyword = None
+ state = SEP
+ for n, c in enumerate(command):
+ # Escaped character
+ if escape:
+ for e in escaped_chars:
+ if c == e:
+ buff += eval(u'"\\' + e + u'"')
+ break
+ else:
+ if c == 'N':
+ buff += r'\N'
+ else:
+ buff += c
+ escape = False
+ continue
+ # Escaped sequence start
+ if c == u'\\':
+ escape = True
+ continue
+ # Looking for spaces
+ if state == SEP:
+ if c in separators:
+ continue
+ if buff and n != 2: # Not the first item (even if was a escape seq)
+ if c == EQUAL: # Keyword found
+ keyword = buff
+ buff = u''
+ continue
+ if buff == r'\N':
+ buff = None
+ if keyword is not None: # Value found
+ dic[str(keyword)] = buff
+ keyword = None
+ else: # Normal parameter found
+ seq.append(buff)
+ buff = u''
+ state = TOKEN
+ # Getting a token
+ if state == TOKEN:
+ if c == DQUOTE:
+ state = DQUOTE
+ continue
+ if c == SQUOTE:
+ state = SQUOTE
+ continue
+ # Check if a keyword is added
+ if c == EQUAL and keyword is None and buff:
+ keyword = buff
+ buff = u''
+ state = SEP
+ continue
+ if c in separators:
+ state = SEP
+ continue
+ buff += c
+ continue
+ # Inside a double quote
+ if state == DQUOTE:
+ if c == DQUOTE:
+ state = TOKEN
+ continue
+ buff += c
+ continue
+ # Inside a single quote
+ if state == SQUOTE:
+ if c == SQUOTE:
+ state = TOKEN
+ continue
+ buff += c
+ continue
+ assert 0, u'Unexpected state'
+ if state == DQUOTE or state == SQUOTE:
+ raise ParseError(command, u'missing closing quote (%s)' % state)
+ if not buff and keyword is not None:
+ raise ParseError(command,
+ u'keyword argument (%s) without value' % keyword)
+ if buff:
+ if buff == r'\N':
+ buff = None
+ if keyword is not None:
+ dic[str(keyword)] = buff
+ else:
+ seq.append(buff)
+ return (seq, dic)
+