Makes it easy to create a command line interface for any function, method or classmethod..
Makes it easy to parse command line arguments for any function, method or classmethod.
You just finished writing an awesome piece of code and now comes the boring part: adding the command line parsing to actually use it …
So now you need to use the awesome, but very verbose, argparse module. For each argument of your entry point method you need to add a name, a help message and/or a default value. But wait… Your parameters are correctly named right!? And you have an awesome docstring for that method. There is probably a way of creating the ArgumentParser easily right?
Yes and it’s called parse_this!
parse_this contains a simple way to create a command line interface from an entire class. For that you will need to use the parse_class class decorator.
# script.py from __future__ import print_function from parse_this import Self, create_parser, parse_class @parse_class() class ParseMePlease(object): """This will be the description of the parser.""" @create_parser(Self, int) def __init__(self, foo, ham=1): """Get ready to be parsed! Args: foo: because naming stuff is hard ham: ham is good and it defaults to 1 """ self._foo = foo self._ham = ham @create_parser(Self, int, int) def do_stuff(self, bar, spam=1): """Can do incredible stuff with bar and spam. Args: bar: as in foobar, will be multiplied with everything else spam: goes well with eggs, spam, bacon, spam, sausage and spam Returns: Everything multiplied with each others """ return self._foo * self._ham * bar * spam if __name__ == "__main__": print(ParseMePlease.parser.call())
python script.py --help # Print a comprehensive help and usage message python script.py 2 do-stuff 2 >>> 4 python script.py 2 --ham 2 do-stuff 2 --spam 2 >>> 16
How does it work TL;DR version?
- You need to decorate the methods you want to be usable from the command line using create_parser.
- The __init__ method arguments and keyword arguments will be the arguments and options of the script command line i.e. the first arguments and options
- The other methods will be transformed into sub-command, again mapping the command line arguments and options to the method’s own arguments
- All you have to do for this to work is:
- Decorate your class with parse_class
- Decorate methods with create_parser making it aware of the type of the arguments. Using Self to designate the self parameter
- Document your class and method with properly formed docstring to get help and usage message
- Call <YourClass>.parser.call() and you are done!
If you feel like you may need more customization and details, please read on!
- If the __init__ method is decorated it will be considered the first, or top-level, parser this means that all arguments in your __init__ will be arguments pass right after invoking you script i.e. python script.py init_arg_1 init_arg_2 etc...
- The description of the top-level parser is taken from the class’s docstring or overwritten by the keyword argument description of parse_class.
- Each method decorated by create_parser will become a subparser of its own. The command name of the subparser is the same as the method name with _ replaced by -. ‘Private’ methods, whose name start with an _, do not have a subparser by default, as this would expose them to the outside. However if you want to expose them you can set the keyword argument parse_private=True in parse_class. If exposed their command name will not contain the leading - as this would be confusing for command parsing. Special methods, such as __str__, can be decorated as well. Their command name will be stripped of all _s resulting in command names such as str.
- When used in a parse_class decorated class create_parser can take an extra parameters name that will be used as the sub-command name. The same modifications are made to the name replacing _ with -
- When calling python script.py --help the help message for every parser will be displayed making easier to find what you are looking for
Arguments and types
Both parse_this and create_parser need a list of types to which arguments will be converted to. Any Python standard type can be used, two special values are used for the self and cls respectively Self and Class. There is no need to provide a type for keyword agurments since it is infered from the default value of the argument. If your method signature contains arg_with_default=12 parse_this expect an int where arg_with_default is.
If this is the containt of test.py:
from __future__ import print_function from parse_this import create_parser, Self class INeedParsing(object): """A class that clearly needs argument parsing!""" def __init__(self, an_argument): self._an_arg = an_argument @create_parser(Self, int, str, params_delim="--") def parse_me_if_you_can(self, an_int, a_string, an_other_int=12): """I dare you to parse me !!! Args: an_int -- int are pretty cool a_string -- string aren't that nice an_other_int -- guess what? I got a default value """ return a_string * an_int, an_other_int * self._an_arg if __name__ == "__main__": need_parsing = INeedParsing(2) print(INeedParsing.parse_me_if_you_can.parser.call(need_parsing))
The following would be the output of the command line python test.py --help:
usage: test.py [-h] [--an_other_int AN_OTHER_INT] an_int a_string I dare you to parse me !!! positional arguments: an_int int are pretty cool a_string string aren't that nice optional arguments: -h, --help show this help message and exit --an_other_int AN_OTHER_INT guess what? I got a default value
The method parse_me_if_you_can expect an int of the name an_int, a str of the name a_string and other int with the name an_other_int and a default value of 12. So does the parser !!! As displayed by the --help command.
Note: create_parser cannot decorate the __init__ method of a class unless the class is itself decorated with parse_class. A ParseThisError will be raised if you attempt to use the call method of such a parser.
The following would be the output of the command line python test.py 2 yes --default 4:
In order to get a help message generated automatically from the method docstring it needs to be in the specific format described below:
... @create_parser(Self, int, int, params_delim=<delimiter_chars>) def method(self, spam, ham): """<description> <blank_line> <arg_name><delimiter_chars><arg_help> <arg_name><delimiter_chars><arg_help> """ pass ...
- description: is a multiline description of the method used for the command line
- each line of argument help have the following component:
- arg_name: the same name as the argument of the method.
- delimiter_chars: one or more chars that separate the argument and its help message. Using whitespaces is not recommended as it could have an expected behavior with multiline help message.
- arg_help: is everything behind the delimiter_chars until the next argument, a blank line or the end of the docstring.
The delimiter_chars can be passed to both parse_this and create_parser as the keywords argument params_delim. It defaults to : since this is the convention I most often use.
If no docstring is specified a generic - not so useful - help message will be generated for the command line and arguments.
Using None as a default value and bool as flags
Using None as a default value is common practice in Python but for parse_this and create_parser to work properly the type of the argument which defaults to None needs to be specified. Otherwise a ParseThisError will be raised.
... @create_parser(str): def function(ham, spam=None): if spam is not None: return ham * spam return ham # Will raise ParseThisError: To use default value of 'None' you need to specify the type of the argument 'spam' for the method 'function' ...
But specifying the type of spam will allow create_parser to work properly
... @create_parser(str, int) def function(ham, spam=None): if spam is not None: return ham * spam return ham # Calling function.parser.call(args="yes".split()) -> 'yes' # Calling function.parser.call(args="yes --spam 3".split()) -> 'yesyesyes' ...
An other common practice is to use bools as flags or switches. All arguments of type bool, either typed direclty or inferred from the default value, will become optional arguments of the command line. A bool argument without default value will default to True as in the following example:
... @create_parser(str, bool) def function(ham, spam): if spam: return ham, spam return ham # Calling function.parser.call(args="yes".split()) -> 'yes', True # Calling function.parser.call(args="yes --spam".split()) -> 'yes' ...
Adding --spam to the arguments will act as a flag/switch setting spam to False. Note that spam as become optional and will be given the value True if --spam is not among the arguments to parse.
Arguments with a boolean default value will act in the same way i.e. acting as flag to change the default value:
... @create_parser(str) def function(ham, spam=False): if spam: return ham, spam return ham # Calling function.parser.call(args="yes".split()) -> 'yes' # Calling function.parser.call(args="yes --spam".split()) -> 'yes', True ...
Here everything works as intended and the default value for spam is False and passing --spam as an argument to be parsed will assign it True.
As a decorator create_parser will create an argument parser for the decorated function. A parser attribute will be added to the method and can be used to parse the command line argument.
from __future__ import print_function from parse_this import create_parser @create_parser(str, int) def concatenate_str(one, two=2): """Concatenates a string with itself a given number of times. Args: one: string to be concatenated with itself two: number of times the string is concatenated, defaults to 2 """ return one * two if __name__ == "__main__": print(concatenate_str.parser.call())
Calling this script from the command line as follow:
python script.py yes --two 3
will return 'yesyesyes' as expected and all the parsing have been done for you.
Note that the function can still be called as any other function from any python file. Also it is not possible to stack create_parser with any decorator that would modify the signature of the decorated function e.g. using functools.wraps.
As a function parse_this will handle the command line arguments directly.
from __future__ import print_function from parse_this import parse_this def concatenate_str(one, two=2): """Concatenates a string with itself a given number of times. Args: one: string to be concatenated with itself two: number of times the string is concatenated, defaults to 2 """ return one * two if __name__ == "__main__": print(parse_this(concatenate_str, [str, int]))
Calling this script with the same command line arguments yes --two 3 will also return 'yesyesyes' as expected.
In a similar fashion you can parse line arguments for classmethods:
class MyClass(object): ... @classmethod @create_parser(Class, int, str, params_delim="--") def parse_me_if_you_can(cls, an_int, a_string, default=12): """I dare you to parse me !!! Args: an_int -- int are pretty cool a_string -- string aren't that nice default -- guess what I got a default value """ return a_string * an_int, default * default ... MyClass.parse_me_if_you_can.parser.call(MyClass)
The output will be the same as using create_parser on a regular method. The only difference is the use of the special value Class to specify where the cls argument is used.
Notes: * The classmethod decorator is placed on top of the create_parser decorator in order for the method to still be a considered a class method. * A classmethod decorated with create_parser in a class decorated with parse_class will not be accessible through the class command line.
parse_this can be installed using the following command:
pip install parse_this
To check that everything is running fine you can run the following command:
python setup.py nosetests
- parse_this and create_parser are not able to be used on methods with *args and **kwargs
- A subsequent effect of the previous caveat is that create_parser cannot be stacked with other decorator that would alter the callable’s signature
- Classmethods cannot be access from the command line in a class decorated with parse_class
- When using create_parser on a method that has an argument with None as a default value its type must be past in the list of types. A ParseThisError will be raised otherwise.
parse_this is released under the MIT Licence. See the bundled LICENSE file for details.