将 argparse 别名解析回原始命令 [英] Resolve argparse alias back to the original command

问题描述

我正在使用具有别名的子解析器/子命令.

I'm using a subparser/subcommand that has an alias.

我正在使用子解析器的 dest 选项来存储子命令的名称，以便我以后可以获取它.

I'm using the dest option for the subparser to store the name of the subcommand so I can get it later.

当前，如果子命令的名称是 reallyLongName 并且别名是 r(比如说)，则 dest 选项存储 reallyLongName 或 r 完全正确 - 我输入的任何内容都会被存储.这很烦人，因为我现在必须检查命令的名称或其任何别名以识别命令.

Currently if the subcommand's name is reallyLongName and the alias is r (say) then the dest option stores either reallyLongName or r exactly - whatever I typed in gets stored. This is annoying because I now have to check for the name of the command or any of its aliases in order to identify the command.

有没有办法让 argparse 将子命令的名称存储在 dest 字段中的某种单一的、规范的文本字符串中?

Is there a way to get argparse to store the subcommand's name in the dest field in some sort of single, canonical text string?

例如，给定以下代码:

import argparse

parser = argparse.ArgumentParser()

subparsers = parser.add_subparsers(dest='command', help='sub-command help')

parser_ag = subparsers.add_parser(  'mySubcommand',
                                    aliases=['m'],
                                    help='Subcommand help')

print(parser.parse_args('mySubcommand'.split()))

print(parser.parse_args('m'.split()))

产生以下输出:

Namespace(command='mySubcommand')
Namespace(command='m')

预期结果:command 对两者都有一个规范的值，例如:

Desired result: command has a single, canonical value for both, for example:

Namespace(command='mySubcommand')
Namespace(command='mySubcommand')

推荐答案

这出奇地难以挖掘.当您添加子解析器时，它会存储在父级 ._actions 属性中.从那里开始，它只是挖掘属性以获得您需要的东西.下面我创建字典以通过 dest 名称引用子解析器参数，然后添加一个函数，让我们将输入的参数重新映射到主参数名称.

This was surprisingly difficult to dig out. When you add a subparser, it gets stored in the parents ._actions attribute. From there it is just digging through attributes to get what you need. Below I create dictionaries to reference the subparser arguments by the dest name, and then added a function that lets us remap the inputted arguments to the primary argument name.

from collections import defaultdict

def get_subparser_aliases(parser, dest):
    out = defaultdict(list)
    prog_str = parser.prog
    dest_dict = {a.dest: a for a in parser._actions}
    try:
        choices = dest_dict.get(dest).choices
    except AttributeError:
        raise AttributeError(f'The parser "{parser}" has no subparser with a `dest` of "{dest}"')

    for k, v in choices.items():
        clean_v = v.prog.replace(prog_str, '', 1).strip()
        out[clean_v].append(k)
    return dict(out)

def remap_args(args, mapping, dest):
    setattr(args, dest, mapping.get(getattr(args, dest)))
    return args

使用您的示例，我们可以使用以下方法重新映射解析参数:

Using your example, we can remap the parse args using:

import argparse

parser = argparse.ArgumentParser()
subparsers = parser.add_subparsers(dest='command', help='sub-command help')
parser_ag = subparsers.add_parser('mySubcommand',
                                  aliases=['m'],
                                  help='Subcommand help')

args = parser.parse_args('m'.split())
mapping = get_subparser_aliases(parser, 'command')
remap_args(args, mapping, 'command')
print(args)
# prints:
Namespace(command='mySubcommand')

这是一个在多个子解析器级别工作的例子.我们有一个带有可选参数的解析器和一个子解析器.子解析器有 3 个可能的参数，最后一个调用另一个子解析器(一个子子解析器)，有 2 个可能的参数.

Here is an example of it at work with multiple subparser levels.. We have a parser with an optional argument and a subparser. The subparser has 3 possible arguments, the last of which invoke another subparser (a sub-subparser), with 2 possible arguments.

您可以检查顶级解析器或第一级子解析器以查看别名映射.

You can examine either the top level parser or the first level subparser to see alias mappings.

import argparse

parser = argparse.ArgumentParser()
parser.add_argument('--someoption', '-s', action='store_true')

subparser1 = parser.add_subparsers(help='sub-command help', dest='sub1')
parser_r = subparser1.add_parser('reallyLongName', aliases=['r'])
parser_r.add_argument('foo', type=int, help='foo help')
parser_s = subparser1.add_parser('otherReallyLong', aliases=['L'])
parser_s.add_argument('bar', choices='abc', help='bar help')
parser_z = subparser1.add_parser('otherOptions', aliases=['oo'])

subparser2 = parser_z.add_subparsers(help='sub-sub-command help', dest='sub2')
parser_x = subparser2.add_parser('xxx', aliases=['x'])
parser_x.add_argument('fizz', type=float, help='fizz help')
parser_y = subparser2.add_parser('yyy', aliases=['y'])
parser_y.add_argument('blip', help='blip help')

get_subparser_aliases(parser, 'sub1')
# returns:
{'reallyLongName': ['reallyLongName', 'r'],
 'otherReallyLong': ['otherReallyLong', 'L'],
 'otherOptions': ['otherOptions', 'oo']}

get_subparser_aliases(parser_z, 'sub2')
# returns:
{'xxx': ['xxx', 'x'], 'yyy': ['yyy', 'y']}

将其与上述函数结合使用，我们可以将收集到的参数重新映射到它们更长的名称.

Using this with the function above, we can remap the collected args to their longer names.

args = parser.parse_args('-s oo x 1.23'.split())
print(args)
# prints:
Namespace(fizz=1.23, someoption=True, sub1='oo', sub2='x')

for p, dest in zip((parser, parser_z), ('sub1', 'sub2')):
    mapping = get_subparser_aliases(p, dest)
    remap_args(args, mapping, dest)

print(args)
# prints:
Namespace(fizz=1.23, someoption=True, sub1='otherOptions', sub2='xxx')

这篇关于将 argparse 别名解析回原始命令的文章就介绍到这了，希望我们推荐的答案对大家有所帮助，也希望大家多多支持IT屋！