ChatGPT解决这个技术问题 Extra ChatGPT

Simple argparse example wanted: 1 argument, 3 results

The documentation for the argparse python module, while excellent I'm sure, is too much for my tiny beginner brain to grasp right now. I don't need to do math on the command line or meddle with formatting lines on the screen or change option characters. All I want to do is "If arg is A, do this, if B do that, if none of the above show help and quit".

then just check sys.argv for the argument you want...
Ever tried plac? It's an easy to use wrapper over argparse with great documentation.
it's not you. it's argparse. it's trying to take you on a journey to the stars and doesn't care where you were headed.
Crazy "pythonic" APIs again :/
Bless you matt wilkie, for standing up for tiny beginner brains everywhere.

R
Russia Must Remove Putin

Here's the way I do it with argparse (with multiple args):

parser = argparse.ArgumentParser(description='Description of your program')
parser.add_argument('-f','--foo', help='Description for foo argument', required=True)
parser.add_argument('-b','--bar', help='Description for bar argument', required=True)
args = vars(parser.parse_args())

args will be a dictionary containing the arguments:

if args['foo'] == 'Hello':
    # code here

if args['bar'] == 'World':
    # code here

In your case simply add only one argument.


as mentioned in my comment to the other answer, I'd like to keep argparse's automatic help formatting, but there doesn't seem to be an option to have an unamed argument (more likely I just don't understand it when I see it), e.g. one needs to do foo.py --action install or foo.py --action remove instead of simply foo.py install
@mattwilkie Then you have to define a positional argument like this: parser.add_argument('install', help='Install the app') (Notice you can't define a positional argument with required=True)
As a noob to argparse, this answer really helped because I did not know where to find the options after they were passed. In other words, I needed to understand how the args dict was generated as above.
Use the 'short form' when calling program directly from command line and the 'long form' when you run a program/command within a script. In that case it is more human readable with the long form and thus easier to follow the logic of the code/script.
Personally I find it cleaner to access arguments as args.foo and args.bar instead of the dictionary syntax. Either way is fine of course, but args is not actually a dictionary but an argparse.Namespace object.
m
mightypile

My understanding of the original question is two-fold. First, in terms of the simplest possible argparse example, I'm surprised that I haven't seen it here. Of course, to be dead-simple, it's also all overhead with little power, but it might get you started.

import argparse

parser = argparse.ArgumentParser()
parser.add_argument("a")
args = parser.parse_args()

if args.a == 'magic.name':
    print 'You nailed it!'

But this positional argument is now required. If you leave it out when invoking this program, you'll get an error about missing arguments. This leads me to the second part of the original question. Matt Wilkie seems to want a single optional argument without a named label (the --option labels). My suggestion would be to modify the code above as follows:

...
parser.add_argument("a", nargs='?', default="check_string_for_empty")
...
if args.a == 'check_string_for_empty':
    print 'I can tell that no argument was given and I can deal with that here.'
elif args.a == 'magic.name':
    print 'You nailed it!'
else:
    print args.a

There may well be a more elegant solution, but this works and is minimalist.


After some time reflecting, I conclude this question actually best answers the Q as asked and the predicament I was in at the time. The other excellent answers have garnered more than enough rep to prove their worth and can stand a little competition. :-)
@badnack: It's whatever you want it to be, whatever 'a' represents. If you expect one argument, a file name for example, it is what was entered as the file name on the command line. You could then do your own processing to determine whether it exists in the filesystem, but that is another Q&A.
@mightypile could you please tell me when exactly "positional argument" is useful? I added something like this parser.add_argument('n', nargs='?', default=5) in my code and when I run python3 test.py n 3, I get this error: usage: test.py [-h] [n] test.py: error: unrecognized arguments: 3 Thank you in advance!
@Milan A positional argument would be required at its given position and would not need to be explicitly named when invoked because it's required/expected/positional (ie test.py 3). You created a script with a single optional argument (and no positional arguments) and as expected, it read "n" as the first argument named "n" (so args.n == "n") and does not know what to do with the "3". See the docs
@mightypile I see...now I understood. Thank you so much for your time and for replying back :)
D
DMH

The argparse documentation is reasonably good but leaves out a few useful details which might not be obvious. (@Diego Navarro already mentioned some of this but I'll try to expand on his answer slightly.) Basic usage is as follows:

parser = argparse.ArgumentParser()
parser.add_argument('-f', '--my-foo', default='foobar')
parser.add_argument('-b', '--bar-value', default=3.14)
args = parser.parse_args()

The object you get back from parse_args() is a 'Namespace' object: An object whose member variables are named after your command-line arguments. The Namespace object is how you access your arguments and the values associated with them:

args = parser.parse_args()
print (args.my_foo)
print (args.bar_value)

(Note that argparse replaces '-' in your argument names with underscores when naming the variables.)

In many situations you may wish to use arguments simply as flags which take no value. You can add those in argparse like this:

parser.add_argument('--foo', action='store_true')
parser.add_argument('--no-foo', action='store_false')

The above will create variables named 'foo' with value True, and 'no_foo' with value False, respectively:

if (args.foo):
    print ("foo is true")

if (args.no_foo is False):
    print ("nofoo is false")

Note also that you can use the "required" option when adding an argument:

parser.add_argument('-o', '--output', required=True)

That way if you omit this argument at the command line argparse will tell you it's missing and stop execution of your script.

Finally, note that it's possible to create a dict structure of your arguments using the vars function, if that makes life easier for you.

args = parser.parse_args()
argsdict = vars(args)
print (argsdict['my_foo'])
print (argsdict['bar_value'])

As you can see, vars returns a dict with your argument names as keys and their values as, er, values.

There are lots of other options and things you can do, but this should cover the most essential, common usage scenarios.


What's the point of the '-f' and '-b'? Why can't you omit this?
It's pretty conventional to have both a 'short form' (one dash) and 'long form' (two dashes) version for each runtime option. You will see this, for example, in almost every standard Unix/Linux utility; do a man cp or man ls and you'll find that many options come in both flavors (e.g. -f, --force). There are probably widely varying reasons why people prefer one or the other, but in any case it's pretty standard to make both forms available in your program.
@DMH Why is that print(args.my_foo) works but print(args.f) gives an error: AttributeError: 'Namespace' object has no attribute 'f'? On the other hand, in add_argument, if I don't specify a named label (the --option labels) i.e. parser.add_argument('-f', default='foobar') then I can run print(args.f). Why? Thank you in advance!
M
Mark E. Haase

Matt is asking about positional parameters in argparse, and I agree that the Python documentation is lacking on this aspect. There's not a single, complete example in the ~20 odd pages that shows both parsing and using positional parameters.

None of the other answers here show a complete example of positional parameters, either, so here's a complete example:

# tested with python 2.7.1
import argparse

parser = argparse.ArgumentParser(description="An argparse example")

parser.add_argument('action', help='The action to take (e.g. install, remove, etc.)')
parser.add_argument('foo-bar', help='Hyphens are cumbersome in positional arguments')

args = parser.parse_args()

if args.action == "install":
    print("You asked for installation")
else:
    print("You asked for something other than installation")

# The following do not work:
# print(args.foo-bar)
# print(args.foo_bar)

# But this works:
print(getattr(args, 'foo-bar'))

The thing that threw me off is that argparse will convert the named argument "--foo-bar" into "foo_bar", but a positional parameter named "foo-bar" stays as "foo-bar", making it less obvious how to use it in your program.

Notice the two lines near the end of my example -- neither of those will work to get the value of the foo-bar positional param. The first one is obviously wrong (it's an arithmetic expression args.foo minus bar), but the second one doesn't work either:

AttributeError: 'Namespace' object has no attribute 'foo_bar'

If you want to use the foo-bar attribute, you must use getattr, as seen in the last line of my example. What's crazy is that if you tried to use dest=foo_bar to change the property name to something that's easier to access, you'd get a really bizarre error message:

ValueError: dest supplied twice for positional argument

Here's how the example above runs:

$ python test.py
usage: test.py [-h] action foo-bar
test.py: error: too few arguments

$ python test.py -h
usage: test.py [-h] action foo-bar

An argparse example

positional arguments:
  action      The action to take (e.g. install, remove, etc.)
  foo-bar     Hyphens are cumbersome in positional arguments

optional arguments:
  -h, --help  show this help message and exit

$ python test.py install foo
You asked for installation
foo

nargs='?' is the incantation for an "optional positional" as per stackoverflow.com/questions/4480075/…
The fact that a positional foo-bar is not transformed to foo_bar is addressed in bugs.python.org/issue15125.
I think an easier workaround for this bug is to just call the argument "foo_bar" instead of "foo-bar", then print args.foo_bar works. Since it is an positional argument you don't have to specify the name when calling the script, so it doesn't matter for the user.
@luator You're right, it is easy to rename the argument, but the author of the bug report makes a good case that this is still a misfeature because of the unnecessary cognitive load. When using argparse, one must pause and recall the different naming conventions for options and arguments. See bugs.python.org/msg164968.
@mehaase I totally agree that this is a misfeature that should be fixed. I just think renaming the argument is the easier and less confusing workaround than having to use getattr (it is also more flexible as it allows you to change an argument from optional to positional without having to change the code that uses the value).
J
Jonathan H

Yet another summary introduction, inspired by this post.

import argparse

# define functions, classes, etc.

# executes when your script is called from the command-line
if __name__ == "__main__":

    parser = argparse.ArgumentParser()
    #
    # define each option with: parser.add_argument
    #
    args = parser.parse_args() # automatically looks at sys.argv
    #
    # access results with: args.argumentName
    #

Arguments are defined with combinations of the following:

parser.add_argument( 'name', options... )              # positional argument
parser.add_argument( '-x', options... )                # single-char flag
parser.add_argument( '-x', '--long-name', options... ) # flag with long name

Common options are:

help: description for this arg when --help is used.

default: default value if the arg is omitted.

type: if you expect a float or int (otherwise is str).

dest: give a different name to a flag (e.g. '-x', '--long-name', dest='longName'). Note: by default --long-name is accessed with args.long_name

action: for special handling of certain arguments store_true, store_false: for boolean args '--foo', action='store_true' => args.foo == True store_const: to be used with option const '--foo', action='store_const', const=42 => args.foo == 42 count: for repeated options, as in ./myscript.py -vv '-v', action='count' => args.v == 2 append: for repeated options, as in ./myscript.py --foo 1 --foo 2 '--foo', action='append' => args.foo == ['1', '2']

store_true, store_false: for boolean args '--foo', action='store_true' => args.foo == True

store_const: to be used with option const '--foo', action='store_const', const=42 => args.foo == 42

count: for repeated options, as in ./myscript.py -vv '-v', action='count' => args.v == 2

append: for repeated options, as in ./myscript.py --foo 1 --foo 2 '--foo', action='append' => args.foo == ['1', '2']

required: if a flag is required, or a positional argument is not.

nargs: for a flag to capture N args ./myscript.py --foo a b => args.foo = ['a', 'b']

choices: to restrict possible inputs (specify as list of strings, or ints if type=int).


A
Alexey

Note the Argparse Tutorial in Python HOWTOs. It starts from most basic examples, like this one:

import argparse
parser = argparse.ArgumentParser()
parser.add_argument("square", type=int,
                    help="display a square of a given number")
args = parser.parse_args()
print(args.square**2)

and progresses to less basic ones.

There is an example with predefined choice for an option, like what is asked:

import argparse
parser = argparse.ArgumentParser()
parser.add_argument("square", type=int,
                    help="display a square of a given number")
parser.add_argument("-v", "--verbosity", type=int, choices=[0, 1, 2],
                    help="increase output verbosity")
args = parser.parse_args()
answer = args.square**2
if args.verbosity == 2:
    print("the square of {} equals {}".format(args.square, answer))
elif args.verbosity == 1:
    print("{}^2 == {}".format(args.square, answer))
else:
    print(answer)

It's nice to see that the docs have been updated. I assure you this wasn't the case when OP posted the question 5 years ago.
P
Peter L

Here's what I came up with in my learning project thanks mainly to @DMH...

Demo code:

import argparse

def main():
    parser = argparse.ArgumentParser()
    parser.add_argument('-f', '--flag', action='store_true', default=False)  # can 'store_false' for no-xxx flags
    parser.add_argument('-r', '--reqd', required=True)
    parser.add_argument('-o', '--opt', default='fallback')
    parser.add_argument('arg', nargs='*') # use '+' for 1 or more args (instead of 0 or more)
    parsed = parser.parse_args()
    # NOTE: args with '-' have it replaced with '_'
    print('Result:',  vars(parsed))
    print('parsed.reqd:', parsed.reqd)

if __name__ == "__main__":
    main()

This may have evolved and is available online: command-line.py

Script to give this code a workout: command-line-demo.sh


Finally a argparse example that makes sense
c
crifan

code file: argparseDemo.py

Simple: common case

name(abbr, full), with help import argparse argParser = argparse.ArgumentParser() argParser.add_argument("-n", "--name", help="your name") args = argParser.parse_args() print("args=%s" % args) print("args.name=%s" % args.name) call python argparseDemo.py -n Crifan python argparseDemo.py --name Crifan output: args=Namespace(name='Crifan') and args.name=Crifan

call python argparseDemo.py -n Crifan python argparseDemo.py --name Crifan

python argparseDemo.py -n Crifan

python argparseDemo.py --name Crifan

output: args=Namespace(name='Crifan') and args.name=Crifan

type argParser.add_argument("-a", "--age", type=int, help="your current age") print("type(args.age)=%s" % type(args.age)) call: python argparseDemo.py --age 30 output: type(args.age)= and args.age=30

call: python argparseDemo.py --age 30

output: type(args.age)= and args.age=30

required argParser.add_argument("-a", "--age", required=True, type=int, help="your current age") call: python argparseDemo.py output: error argparseDemo.py: error: the following arguments are required: -a/--age

call: python argparseDemo.py

output: error argparseDemo.py: error: the following arguments are required: -a/--age

default argParser.add_argument("-a", "--age", type=int, default=20, help="your current age. Default is 20") call: python argparseDemo.py output: args.age=20

call: python argparseDemo.py

output: args.age=20

choices argParser.add_argument("-f", "--love-fruit", choices=['apple', 'orange', 'banana'], help="your love fruits") call: python argparseDemo.py -f apple output: args=Namespace(love_fruit='apple') and args.love_fruit=apple

call: python argparseDemo.py -f apple

output: args=Namespace(love_fruit='apple') and args.love_fruit=apple

multi args argParser.add_argument("-f", "--love-fruit", nargs=2, help="your love fruits") call: python argparseDemo.py -f apple orange output: args.love_fruit=['apple', 'orange']

call: python argparseDemo.py -f apple orange

output: args.love_fruit=['apple', 'orange']

Detail

most simple: -x

code: import argparse argParser = argparse.ArgumentParser() argParser.add_argument("-a") # most simple -> got args.a, type is `str` args = argParser.parse_args() print("args.a=%s" % args.a)

usage = run in command line python argparseDemo.py -a 30 or: ./argparseDemo.py -a 30 makesure argparseDemo.py is executable if not, add it: chmod +x argparseDemo.py

or: ./argparseDemo.py -a 30 makesure argparseDemo.py is executable if not, add it: chmod +x argparseDemo.py

makesure argparseDemo.py is executable if not, add it: chmod +x argparseDemo.py

if not, add it: chmod +x argparseDemo.py

output args.a=30

Note default type is str argParser.add_argument("-a") == argParser.add_argument("-a", type=str) print("type(args.a)=%s" % type(args.a)) -> type(args.a)= args type is Namespace print("type(args)=%s" % type(args)) -> type(args)= args value is Namespace(a='30') print("args=%s" % args) -> args=Namespace(a='30') so we can call/use args.a

default type is str argParser.add_argument("-a") == argParser.add_argument("-a", type=str) print("type(args.a)=%s" % type(args.a)) -> type(args.a)=

argParser.add_argument("-a") == argParser.add_argument("-a", type=str)

print("type(args.a)=%s" % type(args.a)) -> type(args.a)=

args type is Namespace print("type(args)=%s" % type(args)) -> type(args)=

print("type(args)=%s" % type(args)) -> type(args)=

args value is Namespace(a='30') print("args=%s" % args) -> args=Namespace(a='30') so we can call/use args.a

print("args=%s" % args) -> args=Namespace(a='30')

so we can call/use args.a

parameter name

full parameter name: --xxx

code argParser.add_argument("-a", "--age")

usage python argparseDemo.py -a 30 or: python argparseDemo.py --age 30

python argparseDemo.py -a 30 or: python argparseDemo.py --age 30

or: python argparseDemo.py --age 30

get parsed value: args.age Note: NOT args.a, and NOT exist args.a

Note: NOT args.a, and NOT exist args.a

full parameter name with multiple words: --xxx-yyy

code argParser.add_argument("-a", "--current-age")

get parsed value: args.current_age

add help description: help

code argParser.add_argument("-a", help="your age") # with help

output use --help can see description  python argparseDemo.py --help usage: argparseDemo.py [-h] [-a A] optional arguments: -h, --help show this help message and exit -a A your age

use --help can see description  python argparseDemo.py --help usage: argparseDemo.py [-h] [-a A] optional arguments: -h, --help show this help message and exit -a A your age

designate parameter type: type

code argParser.add_argument("-a", type=int) # parsed arg is `int`, not default `str`

output print("type(args.a)=%s" % type(args.a)) -> type(args.a)= print("args=%s" % args) -> args=Namespace(a=30)

print("type(args.a)=%s" % type(args.a)) -> type(args.a)=

print("args=%s" % args) -> args=Namespace(a=30)

add default value: default

code argParser.add_argument("-a", type=int, default=20) # if not pass a, a use default value: 20

effect usage: python argparseDemo.py output: print("args.age=%s" % args.age) -> args=Namespace(a=20)

usage: python argparseDemo.py

output: print("args.age=%s" % args.age) -> args=Namespace(a=20)


q
quasoft

You could also use plac (a wrapper around argparse).

As a bonus it generates neat help instructions - see below.

Example script:

#!/usr/bin/env python3
def main(
    arg: ('Argument with two possible values', 'positional', None, None, ['A', 'B'])
):
    """General help for application"""
    if arg == 'A':
        print("Argument has value A")
    elif arg == 'B':
        print("Argument has value B")

if __name__ == '__main__':
    import plac
    plac.call(main)

Example output:

No arguments supplied - example.py:

usage: example.py [-h] {A,B}
example.py: error: the following arguments are required: arg

Unexpected argument supplied - example.py C:

usage: example.py [-h] {A,B}
example.py: error: argument arg: invalid choice: 'C' (choose from 'A', 'B')

Correct argument supplied - example.py A :

Argument has value A

Full help menu (generated automatically) - example.py -h:

usage: example.py [-h] {A,B}

General help for application

positional arguments:
  {A,B}       Argument with two possible values

optional arguments:
  -h, --help  show this help message and exit

Short explanation:

The name of the argument usually equals the parameter name (arg).

The tuple annotation after arg parameter has the following meaning:

Description (Argument with two possible values)

Type of argument - one of 'flag', 'option' or 'positional' (positional)

Abbreviation (None)

Type of argument value - eg. float, string (None)

Restricted set of choices (['A', 'B'])

Documentation:

To learn more about using plac check out its great documentation:

Plac: Parsing the Command Line the Easy Way


S
Sandy Chapman

To add to what others have stated:

I usually like to use the 'dest' parameter to specify a variable name and then use 'globals().update()' to put those variables in the global namespace.

Usage:

$ python script.py -i "Hello, World!"

Code:

...
parser.add_argument('-i', '--input', ..., dest='inputted_variable',...)
globals().update(vars(parser.parse_args()))
...
print(inputted_variable) # Prints "Hello, World!"

Internally argparse uses getattr and setattr to access values in the Namespace. That way it isn't bothered by oddly formed dest values.
a
aibrahim

I went through all the examples and answers and in a way or another they didn't address my need. So I will list her a scenario that I need more help and I hope this can explain the idea more.

Initial Problem

I need to develop a tool which is getting a file to process it and it needs some optional configuration file to be used to configure the tool.

so what I need is something like the following

mytool.py file.text -config config-file.json

The solution

Here is the solution code

import argparse

def main():
    parser = argparse.ArgumentParser(description='This example for a tool to process a file and configure the tool using a config file.')
    parser.add_argument('filename', help="Input file either text, image or video")
    # parser.add_argument('config_file', help="a JSON file to load the initial configuration ")
    # parser.add_argument('-c', '--config_file', help="a JSON file to load the initial configuration ", default='configFile.json', required=False)
    parser.add_argument('-c', '--config', default='configFile.json', dest='config_file', help="a JSON file to load the initial configuration " )
    parser.add_argument('-d', '--debug', action="store_true", help="Enable the debug mode for logging debug statements." )

    args = parser.parse_args()
    
    filename = args.filename
    configfile = args.config_file

    print("The file to be processed is", filename)
    print("The config file is", configfile)

    if args.debug:
        print("Debug mode enabled")
    else:
        print("Debug mode disabled")

    print("and all arguments are: ", args)

if __name__ == '__main__':
    main()

I will show the solution in multiple enhancements to show the idea

First Round: List the arguments

List all input as mandatory inputs so second argument will be

parser.add_argument('config_file', help="a JSON file to load the initial configuration ")

When we get the help command for this tool we find the following outcome

(base) > python .\argparser_example.py -h
usage: argparser_example.py [-h] filename config_file

This example for a tool to process a file and configure the tool using a config file.

positional arguments:
  filename     Input file either text, image or video
  config_file  a JSON file to load the initial configuration

optional arguments:
  -h, --help   show this help message and exit

and when I execute it as the following

(base) > python .\argparser_example.py filename.txt configfile.json

the outcome will be

The file to be processed is filename.txt
The config file is configfile.json
and all arguments are:  Namespace(config_file='configfile.json', filename='filename.txt')

But the config file should be optional, I removed it from the arguments

(base) > python .\argparser_example.py filename.txt

The outcome will be is:

usage: argparser_example.py [-h] filename config_file
argparser_example.py: error: the following arguments are required: c

Which means we have a problem in the tool

Second Round : Make it optimal

So to make it optional I modified the program as follows

    parser.add_argument('-c', '--config', help="a JSON file to load the initial configuration ", default='configFile.json', required=False)

The help outcome should be

usage: argparser_example.py [-h] [-c CONFIG] filename

This example for a tool to process a file and configure the tool using a config file.

positional arguments:
  filename              Input file either text, image or video

optional arguments:
  -h, --help            show this help message and exit
  -c CONFIG, --config CONFIG
                        a JSON file to load the initial configuration

so when I execute the program

(base) > python .\argparser_example.py filename.txt

the outcome will be

The file to be processed is filename.txt
The config file is configFile.json
and all arguments are:  Namespace(config_file='configFile.json', filename='filename.txt')

with arguments like

(base) > python .\argparser_example.py filename.txt --config_file anotherConfig.json

The outcome will be

The file to be processed is filename.txt
The config file is anotherConfig.json
and all arguments are:  Namespace(config_file='anotherConfig.json', filename='filename.txt')

Round 3: Enhancements

to change the flag name from --config_file to --config while we keep the variable name as is we modify the code to include dest='config_file' as the following:

parser.add_argument('-c', '--config', help="a JSON file to load the initial configuration ", default='configFile.json', dest='config_file')

and the command will be

(base) > python .\argparser_example.py filename.txt --config anotherConfig.json

To add the support for having a debug mode flag, we need to add a flag in the arguments to support a boolean debug flag. To implement it i added the following:

    parser.add_argument('-d', '--debug', action="store_true", help="Enable the debug mode for logging debug statements." )

the tool command will be:

(carnd-term1-38) > python .\argparser_example.py image.jpg -c imageConfig,json --debug

the outcome will be

The file to be processed is image.jpg
The config file is imageConfig,json
Debug mode enabled
and all arguments are:  Namespace(config_file='imageConfig,json', debug=True, filename='image.jpg')

p
pavol.kutaj

New to this, but combining Python with Powershell and using this template, being inspired by an in-depth and great Python Command Line Arguments – Real Python

There is a lot you can do within the init_argparse() and I am covering just the most simple scenario here.

import argparse use if __name__ == "__main__": main() pattern to execute from terminal parse arguments within the main() function that has no parameters as all define a init_argparse() function create a parser object by calling argparse.ArgumentParser() declare one or more argumnent with parser.add_argument("--") retur parser parse args by creating an args object by calling parser.parse_args() define a function proper with param1, param2, ... call function_proper with params being assigned as attributes of an args object e.g. `function_proper(param1=args.param1, param2=args.param2) within a shell call the module with named arguments: e.g. python foobar.py --param1="foo" --param2=="bar"

#file: foobar.py
import argparse

def function_proper(param1, param2):
    #CODE...

def init_argparse() -> argparse.ArgumentParser:
    parser = argparse.ArgumentParser()
    parser.add_argument("--param1")
    parser.add_argument("--param2")
    return parser


def main() -> None:
    parser = init_argparse()
    args = parser.parse_args()
    function_proper(param1=args.param1, param2=args.param2)


if __name__ == "__main__":
    main()
>>> python .\foobar.py --param1="foo" --param2=="bar"

H
Hutch

A really simple way to use argparse and amend the '-h'/ '--help' switches to display your own personal code help instructions is to set the default help to False, you can also add as many additional .add_arguments as you like:

import argparse

parser = argparse.ArgumentParser(add_help=False)

parser.add_argument('-h', '--help', action='help',
                help='To run this script please provide two arguments')
parser.parse_args()

Run: python test.py -h

Output:

usage: test.py [-h]

optional arguments:
  -h, --help  To run this script please provide two arguments

h
halfer

Since you have not clarified wheather the arguments 'A' and 'B' are positional or optional, I'll make a mix of both.

Positional arguments are required by default. If not giving one will throw 'Few arguments given' which is not the case for the optional arguments going by their name. This program will take a number and return its square by default, if the cube option is used it shall return its cube.

import argparse
parser = argparse.ArgumentParser('number-game')
parser.add_argument(
    "number",
    type=int,
    help="enter a number"
   )
parser.add_argument(
   "-c", "--choice",
   choices=['square','cube'],
   help="choose what you need to do with the number"
)

# all the results will be parsed by the parser and stored in args
args = parser.parse_args()

# if square is selected return the square, same for cube 
if args.c == 'square':
    print("{} is the result".format(args.number**2))
elif args.c == 'cube':
    print("{} is the result".format(args.number**3))
else:
    print("{} is not changed".format(args.number))

usage

$python3 script.py 4 -c square
16

Here the optional arguments are taking value, if you just wanted to use it like a flag you can too. So by using -s for square and -c for cube we change the behaviour, by adding action = "store_true". It is changed to true only when used.

parser.add_argument(
    "-s", "--square",
    help="returns the square of number",
    action="store_true"
    )
parser.add_argument(
    "-c", "--cube",
    help="returns the cube of number",
    action="store_true"
    )

so the conditional block can be changed to,

if args.s:
    print("{} is the result".format(args.number**2))
elif args.c:
    print("{} is the result".format(args.number**3))
else:
    print("{} is not changed".format(args.number))

usage

$python3 script.py 4 -c
64

A
Ash Lander

As an addition to existing answers, if you are lazy enough, it is possible to use code generation tool called protoargs. It generates arguments parser from the configuration. For python it uses argparse.

Configuration with optional A and B:

syntax = "proto2";
message protoargs
{
    optional string A     = 1; // A param description
    optional string B     = 2; // B param description
}//protoargs

Configuration with required A and B:

syntax = "proto2";
message protoargs
{
    required string A     = 1; // A param description
    required string B     = 2; // B param description
}//protoargs

Configuration with positional A and B:

syntax = "proto2";
message protoargs
{
    required string A     = 1; // A param description
    required string B     = 2; // B param description
}//protoargs
message protoargs_links
{
}//protoargs_links

Now all you should run is:

python ./protoargs.py -i test.proto -o . --py

And use it (it is possible to take other examples here):

import sys
import test_pa

class ArgsParser:
    program = "test"
    description = "Simple A and B parser test."

    def parse(self, argv):
        self.config = test_pa.parse(self.program, self.description, argv)

    def usage(self):
        return test_pa.usage(self.program, self.description)

if __name__ == "__main__":
    parser = ArgsParser()

    if len(sys.argv) == 1:
        print(parser.usage())
    else:
        parser.parse(sys.argv[1:])

        if parser.config.A:
            print(parser.config.A)

        if parser.config.B:
            print(parser.config.B)

If you want more - change configuration, regenerate parser, use an updated parser.config.

UPD: As mentioned in rules, I must specify that this is my own project


g
gouxute

The simplest answer!

P.S. the one who wrote the document of argparse is foolish

python code:

import argparse
parser = argparse.ArgumentParser(description='')
parser.add_argument('--o_dct_fname',type=str)
parser.add_argument('--tp',type=str)
parser.add_argument('--new_res_set',type=int)
args = parser.parse_args()
o_dct_fname = args.o_dct_fname
tp = args.tp
new_res_set = args.new_res_set

running code

python produce_result.py --o_dct_fname o_dct --tp father_child --new_res_set 1

This answer doesn't add anything new/different than existing answers.