We Keep Coding

Selected options activate right subparsers

I have to invoke my script in this way.
script.py multiple --ways aa bb -as abab -bs bebe
In this case -as abab reffers to "aa" option from --ways parameter and -bs bebe to "bb" option.
And all chosed options should affect what "fulfill" method will be used.
If we chose 'aa' and 'bb' there should only be options '-as' and '-bs' not '-cs'.
import sys
from argparse import ArgumentParser
def fulfill_aa_parser(aa_parser):
aa_parser.add_argument('--ass', '-as', type=str, required=True, choices=['abababa'])
def fulfill_bb_parser(aa_parser):
aa_parser.add_argument('--bass', '-bs', type=str, required=True, choices=['bebebe'])
def fulfill_cc_parser(aa_parser):
aa_parser.add_argument('--cass', '-cs', type=str, required=True, choices=['cycycyc'])
def fulfill_multiple_parser(multiple_parser):
multiple_parser.add_argument('--ways', '-w', type=str, choices=['aa','bb', 'cc'], nargs='+', required=True)
def main(argv):
parser = ArgumentParser(description='TEST CASE')
subparsers = parser.add_subparsers(dest='type')
multiple_parser = subparsers.add_parser(
'multiple'
)
aabbparsers = multiple_parser.add_subparsers()
aa_parser = aabbparsers.add_parser('aa')
bb_parser = aabbparsers.add_parser('bb')
cc_parser = aabbparsers.add_parser('cc')
fulfill_multiple_parser(multiple_parser)
fulfill_aa_parser(aa_parser)
fulfill_bb_parser(bb_parser)
fulfill_cc_parser(cc_parser)
args = parser.parse_args(argv)
if args.type is None:
parser.print_help()
return
if __name__ == '__main__':
main(sys.argv[1:])
Parsing this in this way:
fulfill_aa_parser(multiple_parser)
fulfill_bb_parser(multiple_parser)
fulfill_cc_parser(multiple_parser)
will lead to parser always asking for '-as', '-bs' ,'-cs' and options in '--ways' will not affect this
EDIT : \
This is it looks when there is some thought put to it.
Just simply pass parser to this function
def fulfill_apple_argparser(parser):
parser.add_argument("--apple_argument")
def fulfill_banana_argparser(parser):
parser.add_argument("--banana_argument")
def fulfill_peach_argparser(parser):
parser.add_argument("--peach_argument")
def many_fruits_parse(parser, progs=None, list_of_fruits=('apple', 'banana', 'peach')):
progs = progs or []
if len(list_of_fruits) == 0 or parser in progs:
return
fulfill = {'apple': fulfill_apple_argparser, 'banana': fulfill_banana_argparser,
'peach': fulfill_peach_argparser}
subparsers = parser.add_subparsers(title='subparser', dest=parser.prog)
progs.append(parser)
for fruit in list_of_fruits:
secondary = [x for x in list_of_fruits if x != fruit]
fruit_parser = subparsers.add_parser(fruit, help=fruit)
fulfill[fruit](fruit_parser)
many_fruits_parse(fruit_parser, progs, secondary)

add_subparsers creates a special kind of positional argument, one that uses the add_parser command to create choices. Once a valid choice is provided, parsing is passed to that parser.
With
script.py multiple --ways aa bb -as abab -bs bebe
parser passes the task to multiple_parser. The --ways optional then gets 2 values
Namespace(ways=['aa','bb'])
Neither of those strings is used as a value for aabbparsers, and multiple_parser doesn't know what to do with '-as` or '-bs', and (I expect) will raise an error.
With:
script.py multiple aa -as abab
parsing is passed from parser to multiple_parser to aa_parser, which in turn handles '-as abab', producing (I think)
Namespace(as='abab')
Nesting as you do with multiple and aa is the only way to use multiple subparsers. You can't have two subparses 'in-parallel' (e.g. 'aa' and 'bb').
Especially when testing it's a good idea to provide a dest to the add_subparsers command. It gives information on which subparsers is being invoked.

Check for certain arguments and assign values to an arbitrary variable

I'm writing a program and need to look for one of two arguments set on the command line, and save a value to a single variable based on which one is set.
If I call the program like this:
myprogram -a --foo 123
I want the variable action set to 'a value'. Call it like this:
myprogram -b --foo 123
And action should be set to 'another value'. Call it with neither:
myprogram -c --foo 123
And it should exit with usage info.
Obviously I can do this with some if statements after the fact:
import argparse
parser = argparse.ArgumentParser()
parser.add_argument('-a', action='count')
parser.add_argument('-b', action='count')
parser.add_argument('--foo')
....
args = parser.parse_args()
if args.a == 0 and args.b == 0:
parser.print_usage()
sys.exit(1)
if args.a > 0:
action = 'a value'
elif args.b > 0:
action = 'another value'
But I'm wondering if argparse can do it for me with less code. From what I've seen in the documentation it's not possible, but Python is very very new to me. Thanks.

Look at action='store_const'.
In [240]: parser=argparse.ArgumentParser()
In [241]: parser.add_argument('-a', dest='action',
action='store_const', const='a value')
In [242]: parser.add_argument('-b', dest='action',
action='store_const', const='another value')
In [243]: parser.add_argument('--foo')
In [244]: parser.parse_args('-a --foo 123'.split())
Out[244]: Namespace(action='a value', foo='123')
In [245]: parser.parse_args('-b --foo 123'.split())
Out[245]: Namespace(action='another value', foo='123')
In [246]: parser.parse_args('-c --foo 123'.split())
usage: ipython3 [-h] [-a] [-b] [--foo FOO]
ipython3: error: unrecognized arguments: -c
SystemExit: 2
So args.action will have a value' orb value' depending on the argument. Note both -a and -b store to the same dest.
I left -c undefined, so it uses the normal undefined exit with usage. that could be refined.
Defining a c like this would let you do your own exit:
In [247]: parser.add_argument('-c', dest='action', action='store_const', const='exit')
In [248]: args=parser.parse_args('-c --foo 123'.split())
In [249]: if args.action=='exit':parser.print_usage()
usage: ipython3 [-h] [-a] [-b] [--foo FOO] [-c]
If you used action='store_true' instead of 'count' for your -a and -b, you could simplify the if tree to:
if args.a:
action = 'a value'
elif args.b:
action = 'another value'
else:
parser.print_usage()
sys.exit(1)

I used hapulj's answer, but still ended up with an if statement to check if neither was set. Then I found the ArgumentParser.add_mutually_exclusive_group() function, and ended up with this, which works perfectly.
import argparse
parser = argparse.ArgumentParser()
actiongroup = parser.add_mutually_exclusive_group(required=True)
actiongroup.add_argument('-a', action='store_const', dest='action', const='a value')
actiongroup.add_argument('-b', action='store_const', dest='action', const='another value')
parser.add_argument('--foo')
....
args = parser.parse_args()
Now, arguments -a and -b can't be omitted, and both can't be specified at the same time.

Take multiple patterns from cli. argparse Python3

I have a python version of grep, that I am building for an assignment. I want my python module to take multiple patterns from the command line, just like grep. However no matter what I do, I keep getting conflicts with my 'debug' argument.
This is what it currently looks like from command line (with -h):
pgreper.py [-h] [--debug] pattern
At the moment I can only search with one pattern:
cat input.txt | ./pgreper.py "were"
I would like to be able to search the input.txt file like so, with multiple patterns:
cat input.txt | ./pgreper.py "were" "you"
However when I try and do this, I get the following error:
pgreper.py: error: unrecognized argument: you
I know that it is related to the fact I generate a pattern to search by reading sys.argv[1]. How would I go about editing my script, to allow it to take multiple patterns from sys.argv, without affecting the optional arguments I have implemented?
Many thanks :)
ps Please ignore my comments, thanks.
#!/usr/bin/python3
import sys
import re
import time
import datetime
import inspect
import argparse
parser = argparse.ArgumentParser(description='Python Grep.')
parser.add_argument('--debug', default='debug', action='store_true', help='Print debug messages')
parser.add_argument('pattern', type=str, help='Pattern for pgrepping')
args = parser.parse_args()
class CodeTrace(object):
def __init__(self, line, pattern):
self.line = line
self.pattern = pattern
# #staticmethod
def trace(self, line, pattern):
# Creating Timestamp
ts = time.time()
# Formatting Timestamp
ts = datetime.datetime.fromtimestamp(ts).strftime('[%Y-%m-%d %H:%M:%S:%f]')
stack = inspect.stack()
# Retrieve calling class information
the_class = stack[1][0].f_locals["self"].__class__
# Retrieve calling method information
the_method = stack[1][0].f_code.co_name
the_variables = stack[1][0].f_code.co_varnames
# Formats the contents of the debug trace into a readable format,
# Any parameters passed to the method and the return value, are included in the debug trace
debug_trace = ("{} {}.{}.{} {} {} ".format(ts, str(the_class), the_method, the_variables, pattern, line))
# Send out the debug trace as a standard error output
sys.stderr.write(debug_trace + "\n")
class Grepper(object):
def __init__(self, pattern):
self.pattern = pattern
# #CodeTrace.trace()
def matchline(self, pattern):
regex = re.compile(self.pattern)
for line in sys.stdin:
if regex.search(line):
sys.stdout.write(line)
if args.debug != 'debug':
(CodeTrace(line, pattern).trace(line, pattern))
def main():
pattern = str(sys.argv[1])
print(sys.argv)
Grepper(pattern).matchline(pattern)
if __name__ == "__main__":
main()

You can tell argparse to expect 1 or more arguments, using the nargs keyword argument:
parser.add_argument('patterns', type=str, nargs='+', help='Pattern(s) for pgrepping')
Here + means 1 or more. You can then combine these patterns:
pattern = '|'.join(['(?:{})'.format(p) for p in args.patterns])
and pass that to your grepper. The patterns are combined with | after first being placed in a non-capturing group ((?:...)) to make sure each pattern is treated as distinct.
I'd place all argument parsing in the main() function here:
def main():
parser = argparse.ArgumentParser(description='Python Grep.')
parser.add_argument('--debug', action='store_true', help='Print debug messages')
parser.add_argument('pattern', type=str, nargs='+', help='Pattern(s) for pgrepping')
args = parser.parse_args()
pattern = '|'.join(['(?:{})'.format(p) for p in args.pattern])
Grepper(pattern, args.debug).matchline()
I also removed the default for the --debug option; using store_true means it'll default to False; you can then simply test for args.debug being true or not.
You don't need to pass in pattern twice to Grepper(); you can simply use self.pattern in the matchline method, throughout. Instead, I'd pass in args.debug to Grepper() as well (no need for it to be a global).
Quick demo of what the argument parsing look like, including the help message:
>>> import argparse
>>> parser = argparse.ArgumentParser(description='Python Grep.')
>>> parser.add_argument('--debug', action='store_true', help='Print debug messages')
_StoreTrueAction(option_strings=['--debug'], dest='debug', nargs=0, const=True, default=False, type=None, choices=None, help='Print debug messages', metavar=None)
>>> parser.add_argument('pattern', type=str, nargs='+', help='Pattern(s) for pgrepping')
_StoreAction(option_strings=[], dest='pattern', nargs='+', const=None, default=None, type=<type 'str'>, choices=None, help='Pattern(s) for pgrepping', metavar=None)
>>> parser.print_help()
usage: [-h] [--debug] pattern [pattern ...]
Python Grep.
positional arguments:
pattern Pattern(s) for pgrepping
optional arguments:
-h, --help show this help message and exit
--debug Print debug messages
>>> parser.parse_args(['where'])
Namespace(debug=False, pattern=['where'])
>>> parser.parse_args(['were'])
Namespace(debug=False, pattern=['were'])
>>> parser.parse_args(['were', 'you'])
Namespace(debug=False, pattern=['were', 'you'])
>>> parser.parse_args(['--debug', 'were', 'you'])
Namespace(debug=True, pattern=['were', 'you'])
The pattern then looks like this:
>>> args = parser.parse_args(['were', 'you'])
>>> args.pattern
['were', 'you']
>>> pattern = '|'.join(['(?:{})'.format(p) for p in args.pattern])
>>> pattern
'(?:were)|(?:you)'
If instead you wanted all patterns to match, you'll need to alter Grepper() to take multiple patterns and test all those patterns. Use the all() function to make that efficient (only test as many patterns as is required):
def main():
parser = argparse.ArgumentParser(description='Python Grep.')
parser.add_argument('--debug', action='store_true', help='Print debug messages')
parser.add_argument('pattern', type=str, nargs='+', help='Pattern(s) for pgrepping')
args = parser.parse_args()
Grepper(args.pattern, args.debug).matchline()
and the Grepper class becomes:
class Grepper(object):
def __init__(self, patterns, debug=False):
self.patterns = [re.compile(p) for p in patterns]
self.debug = debug
def matchline(self, debug):
for line in sys.stdin:
if all(p.search(line) for p in self.patterns):
sys.stdout.write(line)
if self.debug:
CodeTrace(line, self.patterns).trace(line)
with appropriate adjustments for the CodeTrace class.

Your argparse argument configuration were not correct. This is really now how you configure the argument. Check the Python documentation for argparse as there is a very good example in there.
the format should always be yourscript.py -aARGUMENTVAL -bARGUMENTVAL ...etc. the -a and -b styles are important.
Your code is edited to have a better application of argparse module below. See if this works better (without action argument for debug):
import sys
import re
import time
import datetime
import inspect
import argparse
parser = argparse.ArgumentParser(description='Python Grep.')
parser.add_argument('-p', '--pattern', type=str, help='Pattern for pgrepping')
parser.add_argument('-d','--debug', type=str, default="false", help='Print debug messages')
args = vars(parser.parse_args());
class CodeTrace(object):
def __init__(self, line, pattern):
self.line = line
self.pattern = pattern
# #staticmethod
def trace(self, line, pattern):
# Creating Timestamp
ts = time.time()
# Formatting Timestamp
ts = datetime.datetime.fromtimestamp(ts).strftime('[%Y-%m-%d %H:%M:%S:%f]')
stack = inspect.stack()
# Retrieve calling class information
the_class = stack[1][0].f_locals["self"].__class__
# Retrieve calling method information
the_method = stack[1][0].f_code.co_name
the_variables = stack[1][0].f_code.co_varnames
# Formats the contents of the debug trace into a readable format,
# Any parameters passed to the method and the return value, are included in the debug trace
debug_trace = ("{} {}.{}.{} {} {} ".format(ts, str(the_class), the_method, the_variables, pattern, line))
# Send out the debug trace as a standard error output
sys.stderr.write(debug_trace + "\n")
class Grepper(object):
def __init__(self, pattern):
self.pattern = pattern
# #CodeTrace.trace()
def matchline(self, pattern):
regex = re.compile(self.pattern)
for line in sys.stdin:
if regex.search(line):
sys.stdout.write(line)
if args.debug != 'debug':
(CodeTrace(line, pattern).trace(line, pattern))
def main():
pattern = str(args['pattern'])
print(sys.argv)
Grepper(pattern).matchline(pattern)
if __name__ == "__main__":
main()
You can supply comma delimited string to separate patters `-p"were,you". Use python's powerful string functions for that
pattern = ((args['pattern']).replace(" ", "")).split(",");
the above will give you a list of patterns to look for?

Python argparse value range help message appearance

I have an argument for a program that is an integer from 1-100 and I just don't like the way that it shows up in the -h help message when using argparse (it literally lists 0, 1, 2, 3, 4, 5,... etc)
Is there any way to change this or have it represented in another way?
Thanks
EDIT:
Here is the code for those who asked:
norse = parser.add_argument_group('Norse')
norse.add_argument('-n', '--norse', required=False, help='Run the Norse IPViking scan.', action='store_true')
norse.add_argument('--threshold', required=False, type=int, choices=range(0,101), help='Threshold (0-100) denoting at what threat level to provide additional data on an IP \
address. Default is 49.', default=49)

Use the metavar parameter of add_argument().
For example:
norse = parser.add_argument_group('Norse')
norse.add_argument('-n', '--norse', required=False, help='Run the Norse IPViking scan.', action='store_true')
norse.add_argument('--threshold', required=False, type=int, choices=range(0,101),
metavar="[0-100]",
help='Threshold (0-100) denoting at what threat level to provide additional data on an IP \
address. Default is 49.', default=49)
Test:
from argparse import ArgumentParser
norse = ArgumentParser()
norse.add_argument('-n', '--norse', required=False, help='Run the Norse IPViking scan.', action='store_true')
norse.add_argument('--threshold', required=False, type=int, choices=range(0,101), metavar="[0-100]", help='Threshold (0-100) denoting at what threat level to provide additional data on an IP address. Default is 49.', default=49)
norse.print_help()
Results
usage: -c [-h] [-n] [--threshold [0-100]]
optional arguments:
-h, --help show this help message and exit
-n, --norse Run the Norse IPViking scan.
--threshold [0-100] Threshold (0-100) denoting at what threat level to
provide additional data on an IP address. Default is
49.

You can customize action, e.g:
#!/usr/bin/env python
import argparse
class Range(argparse.Action):
def __init__(self, minimum=None, maximum=None, *args, **kwargs):
self.min = minimum
self.max = maximum
kwargs["metavar"] = "[%d-%d]" % (self.min, self.max)
super(Range, self).__init__(*args, **kwargs)
def __call__(self, parser, namespace, value, option_string=None):
if not (self.min <= value <= self.max):
msg = 'invalid choice: %r (choose from [%d-%d])' % \
(value, self.min, self.max)
raise argparse.ArgumentError(self, msg)
setattr(namespace, self.dest, value)
norse = argparse.ArgumentParser('Norse')
norse.add_argument('--threshold', required=False, type=int, min=0, max=100,
action=Range,
help='Threshold [%(min)d-%(max)d] denoting at what threat \
level to provide additional data on an IP address. \
Default is %(default)s.', default=49)
args = norse.parse_args()
print args
Test it:
~: user$ ./test.py --threshold 10
Namespace(threshold=10)
~: user$ ./test.py --threshold -1
usage: Norse [-h] [--threshold [0-100]]
Norse: error: argument --threshold: invalid choice: -1 (choose from [0-100])
~: user$ ./test.py -h
usage: Norse [-h] [--threshold [0-100]]
optional arguments:
-h, --help show this help message and exit
--threshold [0-100] Threshold [0-100] denoting at what threat level to
provide additional data on an IP address. Default is
49.

With a custom type, it is easier to control the error message (via the ArgumentTypeError). I still need the metavar to control the usage display.
import argparse
def range_type(astr, min=0, max=101):
value = int(astr)
if min<= value <= max:
return value
else:
raise argparse.ArgumentTypeError('value not in range %s-%s'%(min,max))
parser = argparse.ArgumentParser()
norse = parser.add_argument_group('Norse')
...
norse.add_argument('--range', type=range_type,
help='Value in range: Default is %(default)s.',
default=49, metavar='[0-101]')
parser.print_help()
print parser.parse_args()
producing:
2244:~/mypy$ python2.7 stack25295487.py --ran 102
usage: stack25295487.py [-h] [-n] [--threshold [0:101]] [--range [0-101]]
optional arguments:
-h, --help show this help message and exit
Norse:
...
--range [0-101] Value in range: Default is 49.
usage: stack25295487.py [-h] [-n] [--threshold [0:101]] [--range [0-101]]
stack25295487.py: error: argument --range: value not in range 0-101
I could use functools.partial to customize the range values:
type=partial(range_type, min=10, max=90)

Here are a couple ways you can do it instead
def parseCommandArgs():
parser = argparse.ArgumentParser()
parser.add_argument('-i', dest='myDest', choices=range(1,101), type=int, required=True, metavar='INT[1,100]', help='my help message')
return parser.parse_args()
You can also use action instead, which I highly recommend since it allows more customization
def verify():
class Validity(argparse.Action):
def __call__(self, parser, namespace, values, option_string=None):
if values < 1 or values > 100:
# do something
pass
return Validity
def parseCommandArgs():
parser = argparse.ArgumentParser()
parser.add_argument('-i', dest='myDest', required=True, metavar='INT[1,100]', help='my help message', action=verify())
return parser.parse_args()

I have written a library for this: argparse-range
>>> from argparse import ArgumentParser
>>> from argparse_range import range_action
>>> parser = ArgumentParser()
>>> parser.add_argument("rangedarg", action=range_action(0, 10), help="An argument")
>>> args = parser.parse_args(["0"])
>>> args.rangedarg
0
>>> parser.parse_args(["20"])
Traceback (most recent call last):
....
argparse.ArgumentTypeError: Invalid choice: 20 (must be in range 0..=10)
Helptext is added transparently:
foo.py --help
usage: foo.py [-h] rangedarg
positional arguments:
rangedarg An argument (must be in range 0..=10)
optional arguments:
-h, --help show this help message and exit
Other features include:
Handles default, metavar, and nargs consistently with default argparse handling
Type inference, with explicit overrides using the type argument

Verbose level with argparse and multiple -v options

I'd like to be able to specify different verbose level, by adding more -v options to the command line. For example:
$ myprogram.py
$ myprogram.py -v
$ myprogram.py -vv
$ myprogram.py -v -v -v
would lead to verbose=0, verbose=1, verbose=2, and verbose=3 respectively. How can I achieve that using argparse?
Optionally, it could be great to also be able to specify it like
$ myprogram -v 2

argparse supports action='count':
import argparse
parser = argparse.ArgumentParser()
parser.add_argument('-v', '--verbose', action='count', default=0)
for c in ['', '-v', '-v -v', '-vv', '-vv -v', '-v -v --verbose -vvvv']:
print(parser.parse_args(c.split()))
Output:
Namespace(verbose=0)
Namespace(verbose=1)
Namespace(verbose=2)
Namespace(verbose=2)
Namespace(verbose=3)
Namespace(verbose=7)
The only very minor niggle is you have to explicitly set default=0 if you want no -v arguments to give you a verbosity level of 0 rather than None.

You could do this with nargs='?' (to accept 0 or 1 arguments after the -v flag) and a custom action (to process the 0 or 1 arguments):
import sys
import argparse
class VAction(argparse.Action):
def __init__(self, option_strings, dest, nargs=None, const=None,
default=None, type=None, choices=None, required=False,
help=None, metavar=None):
super(VAction, self).__init__(option_strings, dest, nargs, const,
default, type, choices, required,
help, metavar)
self.values = 0
def __call__(self, parser, args, values, option_string=None):
# print('values: {v!r}'.format(v=values))
if values is None:
self.values += 1
else:
try:
self.values = int(values)
except ValueError:
self.values = values.count('v')+1
setattr(args, self.dest, self.values)
# test from the command line
parser = argparse.ArgumentParser()
parser.add_argument('-v', nargs='?', action=VAction, dest='verbose')
args = parser.parse_args()
print('{} --> {}'.format(sys.argv[1:], args))
print('-'*80)
for test in ['-v', '-v -v', '-v -v -v', '-vv', '-vvv', '-v 2']:
parser = argparse.ArgumentParser()
parser.add_argument('-v', nargs='?', action=VAction, dest='verbose')
args=parser.parse_args([test])
print('{:10} --> {}'.format(test, args))
Running script.py -v -v from the command line yields
['-v', '-v'] --> Namespace(verbose=2)
--------------------------------------------------------------------------------
-v --> Namespace(verbose=1)
-v -v --> Namespace(verbose=2)
-v -v -v --> Namespace(verbose=3)
-vv --> Namespace(verbose=2)
-vvv --> Namespace(verbose=3)
-v 2 --> Namespace(verbose=2)
Uncomment the print statement to see better what the VAction is doing.

You could handle the first part of your question with append_const. Otherwise, you're probably stuck writing a custom action, as suggested in the fine answer by unutbu.
import argparse
ap = argparse.ArgumentParser()
ap.add_argument('-v', action = 'append_const', const = 1)
for c in ['', '-v', '-v -v', '-vv', '-vv -v']:
opt = ap.parse_args(c.split())
opt.v = 0 if opt.v is None else sum(opt.v)
print opt
Output:
Namespace(v=0)
Namespace(v=1)
Namespace(v=2)
Namespace(v=2)
Namespace(v=3)

Here's my take on this that doesn't use any new classes, works in both Python 2 and 3 and supports relative adjustments from the default using "-v"/"--verbose" and "-q"/"--quiet", but it doesn't support using numbers e.g. "-v 2":
#!/usr/bin/env python
import argparse
import logging
import sys
LOG_LEVELS = ["DEBUG", "INFO", "WARNING", "ERROR", "CRITICAL"]
DEFAULT_LOG_LEVEL = "INFO"
def main(argv):
parser = argparse.ArgumentParser()
parser.add_argument(
"--verbose", "-v",
dest="log_level",
action="append_const",
const=-1,
)
parser.add_argument(
"--quiet", "-q",
dest="log_level",
action="append_const",
const=1,
)
args = parser.parse_args(argv[1:])
log_level = LOG_LEVELS.index(DEFAULT_LOG_LEVEL)
# For each "-q" and "-v" flag, adjust the logging verbosity accordingly
# making sure to clamp off the value from 0 to 4, inclusive of both
for adjustment in args.log_level or ():
log_level = min(len(LOG_LEVELS) - 1, max(log_level + adjustment, 0))
log_level_name = LOG_LEVELS[log_level]
print(log_level_name)
logging.getLogger().setLevel(log_level_name)
if __name__ == "__main__":
main(sys.argv)
Example:
$ python2 verbosity.py -vvv
DEBUG
$ python3 verbosity.py -vvv -q
INFO
$ python2 verbosity.py -qqq -vvv -q
WARNING
$ python2 verbosity.py -qqq
CRITICAL

Expanding on unutbu's answer, here's a custom action including handling of a --quiet/-q combination. This is tested in Python3. Using it in Python >=2.7 should be no big deal.
class ActionVerbose(argparse.Action):
def __call__(self, parser, args, values, option_string=None):
#print(parser, args, values, option_string)
# Obtain previously set value in case this option call is incr/decr only
if args.verbose == None:
base = 0
else:
base = args.verbose
# One incr/decr is determined in name of option in use (--quiet/-q/-v/--verbose)
option_string = option_string.lstrip('-')
if option_string[0] == 'q':
incr = -1
elif option_string[0] == 'v':
incr = 1
else:
raise argparse.ArgumentError(self,
'Option string for verbosity must start with v(erbose) or q(uiet)')
# Determine if option only or values provided
if values==None:
values = base + incr
else:
# Values might be an absolute integer verbosity level or more 'q'/'v' combinations
try:
values = int(values)
except ValueError:
values = values.lower()
if not re.match('^[vq]+$', values):
raise argparse.ArgumentError(self,
"Option string for -v/-q must contain only further 'v'/'q' letters")
values = base + incr + values.count('v') - values.count('q')
setattr(args, self.dest, values)
#classmethod
def add_to_parser(cls,
parser, dest='verbose', default=0,
help_detail='(0:errors, 1:info, 2:debug)'):
parser.add_argument('--verbose', nargs='?', action=ActionVerbose, dest=dest, metavar='level',
default=default,
help='Increase or set level of verbosity {}'.format(help_detail))
parser.add_argument('-v', nargs='?', action=ActionVerbose, dest=dest, metavar='level',
help='Increase or set level of verbosity')
parser.add_argument('--quiet', nargs='?', action=ActionVerbose, dest=dest, metavar='level',
help='Decrease or set level of verbosity')
parser.add_argument('-q', nargs='?', action=ActionVerbose, dest=dest, metavar='level',
help='Decrease or set level of verbosity')
There's a convenience class method which can be used to set up all four option handlers for --verbose, -v, -q, --quiet. Use it like this:
parser = argparse.ArgumentParser()
ActionVerbose.add_to_parser(parser, default=defaults['verbose'])
# add more arguments here with: parser.add_argument(...)
args = parser.parse_args()
When using a script having these arguments you can do:
./script -vvvvvv -v 4 -v 0 -v -vvv --verbose --quiet 2 -v qqvvqvv
With this command line args.verbose would be 4.
Any -v/-q/--verbose/--quiet with a given number is a hard, absolute set of args.verbose to that given number (=verbosity level).
Any -v/--verbose without a number is an increment of that level.
Any -q/--quiet without a number is a decrement of that level.
Any -v/-q may immediately be followed up with more v/q letters, the resulting level is the old level + sum(count('v')) - sum(count('q'))
Overall default is 0
The custom action should be fairly easy to modify in case you want a different behaviour. For example, some people prefer that any --quiet resets the level to 0, or even to -1. For this, dremove the nargs from the add_argument of -q and --quiet, and also hardcode to set value = 0 if option_string[0] == 'q'.
Proper parser errors are nicely printed if usage is wrong:
./script -vvvvvv -v 4 -v 0 -v -vvv --verbose --quiet 2 -v qqvvqvav
usage: script [-h] [--verbose [level]]
[-v [level]] [--quiet [level]] [-q [level]]
script: error: argument -v: Option string for -v/-q must contain only further 'v'/'q' letters

argparse supports the append action which lets you specify multiple arguments. Check http://docs.python.org/library/argparse.html, search for "append".

Your first proposed method would be more likely to confuse. Different option names for different levels of verbosity, or one verbose flag optionally followed by a numeric indicator of the level of verbosity is less likely to confuse a user and would allow more flexibility in assigning verbosity levels.

I've come up with an alternative; while it doesn't exactly match OP's request, it fulfilled my requirements and I thought it worth sharing.
Use a mutually exclusive group to either count the number of short options or store the integer value of a long option.
import argparse
parser = argparse.ArgumentParser()
verbosity_group = parser.add_mutually_exclusive_group()
verbosity_group.add_argument(
'-v',
action='count',
dest='verbosity',
help='Turn on verbose output. Use more to turn up the verbosity level'
)
verbosity_group.add_argument(
'--verbose',
action='store',
type=int,
metavar='N',
dest='verbosity',
help='Set verbosity level to `N`'
)
parser.set_defaults(
verbosity=0
)
parser.parse_args()
parser.parse_args([])
# Namespace(verbosity=0)
parser.parse_args(['-v', '-vv'])
# Namespace(verbosity=3)
parser.parse_args(['--verbose=4'])
# Namespace(verbosity=4)
parser.parse_args(['--verbose'])
# error: argument --verbose: expected one argument
As you can see, it allows you to "stack" single char options and allows you to use the long option name to set the value explicitly. The downside is that you cannot use the long option as a switch (the last example generates an exception.)