在bash中,python可以这样做吗?

时间:2012-10-20 23:11:07

标签: python shell path glob expansion

python可以通过像这样的短格式语法生成具有多子路径的路径吗?

vars=project/{DEBIAN,usr/{bin,usr/{applications,icons,share}},computer}

3 个答案:

答案 0 :(得分:1)

不,没有这样的内置速记。一种合理的pythonic方式是:

vars = ["project" + path for path in (
    ["/DEBIAN"] +
    ["/usr" + path for path in (
        ['/bin'] +
        ['/usr' +  path for path in [
            "/applications",
            "/icons",
            "/share"]
        ]
    )] +
    ['/computer']
)]

编辑:

您可以定义一个功能以使其更容易。这是一个懒惰的生成器版本:

def expand(base, paths):
    for path in paths:
        if type(path) == str:
            yield base + path
        else:
            for p in path:
                yield base + p

vars = expand("project", [
    "/debian",
    expand("/usr", [
        "/bin",
        expand("/usr", [
            "/applications",
            "/icons",
            "/share"
        ]),
        "/computer"
    ])
])

答案 1 :(得分:1)

通常我会建议使用像

这样的东西
def extend(base, *others):
    [base + o for o in others] if others else [base]

然后再做

 extend("project", "/debian", *extend("/usr", *(extend("/bin") + extend("/usr", "/applications", "/icons", "/share"))))

。但是,由于您更喜欢解析字符串,我尝试提供替代方法:

def commasplit(s):
    start = 0
    level = 0
    for i, c in enumerate(s):
        if c == '{':
            level += 1
        elif c == '}':
            level -= 1
        elif c == ',' and level == 0:
            yield s[start:i]
            start = i+1
    yield s[start:]

def split(s):
    import re
    found = False
    for m in re.finditer("(\{.*\})",s):
        found = True
        for p in commasplit(s[m.start() + 1:m.end() - 1]):
            for i in split(p):
                yield s[:m.start()] + i + s[m.end():]
    if not found:
        yield s

cs = "a,b,c,{d,e,f},g"
print list(commasplit(cs)) # -> seems to work

s = "project/{DEBIAN,usr/{bin,usr/{applications,icons,share}},computer}"
print s
for n, i in enumerate(split(s)): print n, i # ->  as well.

答案 2 :(得分:0)

谢谢。

我尝试编写代码。

我总是使用[“0,1,2,3,4,5,6,7,8,9”.split(“,”),[“a,b,c,d,e”.split (“,”)]]而不是正式列表。

因为它更快 - 没有移位按,显然,并且比[“x”,“y”,“z”]更有意义。

所以,我希望有一些字符串方法可以像bash一样,在代码中,在python中,很快...

Python语法:project,[DEBIAN,usr,[bin,usr,[applications,icons,share / winpath]],computer]

Bash语法:project / {DEBIAN,usr / {bin,usr / {applications,icons,share}},computer}

import re
def genpaths (multils, root=""):
    suc=[]
    for l in multils:
        if type(l) == type([]):
            suc.extend(genpaths (l, root+multils[multils.index(l)-1]+"/"))
        else: 
            suc.append( root+l)
    return filter(None, suc)

def easylist (s):
    s=eval("['''"+re.sub("(\]+)(''',''')?", "'''\\1,'''",re.sub("(''',''')?(\[+)","''',\\2'''",re.sub(",","''','''",s)))+"''']")
    return s

def bashlist (s):
    s=eval("['''"+re.sub("(\}+)(''',''')?", "'''\\1,'''",re.sub("(/)?(\{+)","''',\\2'''",re.sub(",","''','''",s))).replace("{","[").replace("}","]")+"''']")
    return s

def pythonGenPath (s):
    return genpaths(bashlist (s))

def bashGenPath (s):
    return genpaths(bashlist (s))

#testing
print pythonGenPath ("project,[DEBIAN,usr,[bin,usr,[applications,icons,share/winpath]],computer]")
print bashGenPath ("project/{DEBIAN,usr/{bin,usr/{applications,icons,share}},computer}")

结果是:

['project', 'project/DEBIAN', 'project/usr', 'project/usr/bin', 'project/usr/usr', 'project/usr/usr/applications', 'project/usr/usr/icons', 'project/usr/usr/share', 'project/computer']