首页 / 技术问答 / Python问答列表 / 问答详情

Python 阻止正则表达式的尾部消耗头部

我在一个正则表达式(在Python3.9中)遇到了麻烦，我已经挖掘了好几天，仍然被阻止.

下面是这样的表达方式:

^(?P<head>(?<=^).+?)??(?P<mid>(?:(?<=^)|[.])x+)?(?P<tail>[.][^/]+?)?(?<!^)$

这与以下内容匹配:

xx  # works, goes in mid
a.xx  # works, head "a", mid ".xx"
a.b.xx  # FAILS, head "a", tails consumes the rest, why?
a.xx.c  # works, head "a", mid ".xx", tail ".c"
a.b.xx.c  # FAILS, head "a", tails consumes the rest, why?
  • A.b.xx:我需要"A.B"作为头部，".xx"作为中间，没有 tail
  • A.b.xx.c:我需要"A.B"作为头部，".xx"作为中间，".C"作为尾部

我不明白为什么在某些情况下 tail 会吃掉中间部分，而在某些情况下不会，你知道吗？你知道我怎么才能总是把".c"放在 tail 吗？非常感谢！

你能用语言描述一下你的正则表达式应该匹配什么吗？

这适用于特定于视觉效果的文件序列模式，例如:

  • 描述.详细信息.[UDIM令牌或瓷砖编号].[帧编号].exr
  • description.detail. [UDIM令牌或瓦片编号].exr
  • Description.Detail.[帧编号].exr
  • 描述.Detail.u&lt；udim&gt；.@.exr
  • Description.Detail.u1001.@.exr
  • Description.detail.u1001.1001.exr
  • 描述.Detail.u&lt；udim&gt；.exr
  • Description.detail.U1001.exr
  • 描述.Detail.@.exr
  • Description.detail.1001.exr
  • U&lt；Udim&>.exr
  • U1001.exr
  • _
  • 1001.exr

所以我们实际上有这样的群体:

  • 可选标题，文件序列的一般描述
  • 分隔符作为点，但这些可以包括在头部多次
  • 作为标记或数字的UDIM部分，前面始终为"u"
  • 作为填充令牌或数字的帧编号部分
  • 尾部，这里是文件扩展名，但可以更多
  • 我们需要使用UDIM、帧或两者进行检测，它们始终按此顺序排列

我把它简化到重现问题行为所需的最低限度，"MID"组可以是UDIM或框架.

更多详细信息，实际正则表达式:

import os

from aenum import LowerStrEnum, auto


class GroupName(LowerStrEnum):
    """Regex group name."""

    FRAME_NUMBER = auto()
    FRAME_SEPARATOR = auto()
    FRAME_TOKEN = auto()
    UDIM_SEPARATOR = auto()
    UDIM_TILE = auto()
    UDIM_TOKEN = auto()

    def __str__(self) -> str:
        return self.value


# Separator for pattern sections.
SEPARATOR = r"[_.]"

# Start or separator.
START_OR_SEPARATOR = rf"(?<=^)|{SEPARATOR}"

# UDIM tile.
UDIM_TILE = rf"(?P<{GroupName.UDIM_TILE}>[\d#@]{{4}})"

# UDIM or UDIM token.
UDIM_TOKEN = rf"(?P<{GroupName.UDIM_TOKEN}><[Uu][Dd][Ii][Mm]>)"

# Padding: '#', '###', '@', '@@', '%d', '%03d', '$F', '$F4'
PADDING = r"#+|@+|%(?:\d+)?d|\$F(?:\d+)?"

# Frame range: '1-2' '1,2' '1, 2' '1-2, 3' '1-2,3' '1,2-3'
FRAME_RANGE = r"-?\d+(?:[-,(?:, )]+-?\d+)+(x\d)?"

# Frame token (range/padding) regex.
FRAME_TOKEN = (
    rf"(?P<{GroupName.FRAME_TOKEN}>"
    r"(?:"
    r"(?P<joined_embedded_frame_range>"
    rf"(?:{FRAME_RANGE})"
    r"|(?:-?\d+)"
    r")"
    rf"(?P<joined_padding>{PADDING})"
    r")"
    rf"|(?P<embedded_frame_range>{FRAME_RANGE})"
    rf"|(?P<padding>{PADDING})"
    r")"
)

# Base for all types of paths.
BASE = (
    r"^"
    # Head group.
    r"(?:"
    rf"(?P<directory>(?<=^).*{os.path.sep})?"
    rf"(?P<head>[^{os.path.sep}]+?)"
    r"(?!$)"  # Stop the head from consuming everything next.
    r")??"
    # UDIM group.
    rf"(?:"
    rf"(?P<{GroupName.UDIM_SEPARATOR}>{START_OR_SEPARATOR})"
    rf"[uU](?:{UDIM_TILE}|{UDIM_TOKEN})"
    r")?"
    # Frame sequence group.
    r"(?:"
    rf"(?P<{GroupName.FRAME_SEPARATOR}>{START_OR_SEPARATOR})"
    r"(?:"
    rf"{FRAME_TOKEN}"
    rf"|(?P<{GroupName.FRAME_NUMBER}>-?\d+)"
    r")"
    r")?"
    # Tail & ext group.
    r"(?:"
    rf"(?P<tail>{SEPARATOR}[^/]+)?"
    rf"(?P<ext>{SEPARATOR}[a-zA-Z]+)"
    r")?"
    rf"(?: \[*(?P<extended_frame_range>{FRAME_RANGE})\]*?)?"
    r"(?<!^)$"
)

^(?:(?P<directory>(?<=^).*/)?(?P<head>[^/]+?)(?!$))??(?:(?P<udim_separator>(?<=^)|[_.])[uU](?:(?P<udim_tile>[\d#@]{4})|(?P<udim_token><[Uu][Dd][Ii][Mm]>)))?(?:(?P<frame_separator>(?<=^)|[_.])(?:(?P<frame_token>(?:(?P<joined_embedded_frame_range>(?:-?\d+(?:[-,(?:, )]+-?\d+)+(x\d)?)|(?:-?\d+))(?P<joined_padding>#+|@+|%(?:\d+)?d|\$F(?:\d+)?))|(?P<embedded_frame_range>-?\d+(?:[-,(?:, )]+-?\d+)+(x\d)?)|(?P<padding>#+|@+|%(?:\d+)?d|\$F(?:\d+)?))|(?P<frame_number>-?\d+)))?(?:(?P<tail>[_.][^/]+)?(?P<ext>[_.][a-zA-Z]+))?(?: \[*(?P<extended_frame_range>-?\d+(?:[-,(?:, )]+-?\d+)+(x\d)?)\]*?)?(?<!^)$

打印以下字符串的分组字典:

tests/test_paths/udim_seq/head.1001
{'directory': 'tests/test_paths/udim_seq/',
 'embedded_frame_range': None,
 'ext': None,
 'extended_frame_range': None,
 'frame_number': '1001',
 'frame_separator': '.',
 'frame_token': None,
 'head': 'head',
 'joined_embedded_frame_range': None,
 'joined_padding': None,
 'padding': None,
 'tail': None,
 'udim_separator': None,
 'udim_tile': None,
 'udim_token': None}

tests/test_paths/udim_seq/book.open.mid.u1001.1-2#
{'directory': 'tests/test_paths/udim_seq/',
 'embedded_frame_range': None,
 'ext': None,
 'extended_frame_range': None,
 'frame_number': None,
 'frame_separator': '.',
 'frame_token': '1-2#',
 'head': 'book.open.mid',
 'joined_embedded_frame_range': '1-2',
 'joined_padding': '#',
 'padding': None,
 'tail': None,
 'udim_separator': '.',
 'udim_tile': '1001',
 'udim_token': None}

tests/test_paths/udim_seq/head.1001.exr
{'directory': 'tests/test_paths/udim_seq/',
 'embedded_frame_range': None,
 'ext': '.exr',
 'extended_frame_range': None,
 'frame_number': '1001',
 'frame_separator': '.',
 'frame_token': None,
 'head': 'head',
 'joined_embedded_frame_range': None,
 'joined_padding': None,
 'padding': None,
 'tail': None,
 'udim_separator': None,
 'udim_tile': None,
 'udim_token': None}

tests/test_paths/udim_seq/book.open.mid.1001.exr
{'directory': 'tests/test_paths/udim_seq/',
 'embedded_frame_range': None,
 'ext': '.exr',
 'extended_frame_range': None,
 'frame_number': None,
 'frame_separator': None,
 'frame_token': None,
 'head': 'book',
 'joined_embedded_frame_range': None,
 'joined_padding': None,
 'padding': None,
 'tail': '.open.mid.1001',
 'udim_separator': None,
 'udim_tile': None,
 'udim_token': None}

tests/test_paths/udim_seq/book.open.mid.u1001.1-2#.exr
{'directory': 'tests/test_paths/udim_seq/',
 'embedded_frame_range': None,
 'ext': '.exr',
 'extended_frame_range': None,
 'frame_number': None,
 'frame_separator': None,
 'frame_token': None,
 'head': 'book',
 'joined_embedded_frame_range': None,
 'joined_padding': None,
 'padding': None,
 'tail': '.open.mid.u1001.1-2#',
 'udim_separator': None,
 'udim_tile': None,
 'udim_token': None}

一旦头部包括分隔符， tail 就会吃掉一切...

推荐答案

您的正则表达式的问题出在这个组中:

(?P<head>[^/]+?)

因为它执行延迟匹配，所以当您处理字符串tests/test_paths/udim_seq/book.open.mid.1001.exr并且引擎在book之后到达.时，它将继续将字符串的其余部分与(?P<tail>[_.][^/]+)捕获组进行匹配.

您可以通过强制输入中存在UDIM模式或框架模式(或两者都有)来防止这种情况发生.这可以以正则表达式的形式完成，如下所示:

(?:A?B|A)

其中A表示UDIM模式，B表示帧模式.现在，由于这需要重复UDIM模式，您需要使用PCRE兼容的正则表达式引擎来引用表达式(使用P>name表示法)，以避免重复组名的问题(或者，在重复中更改组名，然后在后处理中合并值).在python中，你可以使用支持这个特性的regex库.

为了可读性，用空格写出您的正则表达式:

^
(?:(?P<directory>(?<=^).*/)?(?P<head>[^/]+?)(?!$))??
(?:
(?P<udim>
(?:(?P<udim_separator>(?<=^)|[_.])[uU](?:(?P<udim_tile>[\d#@]{4})|(?P<udim_token><[Uu][Dd][Ii][Mm]>)))
)?
(?:(?P<frame_separator>(?<=^)|[_.])(?:(?P<frame_token>(?:(?P<joined_embedded_frame_range>(?:-?\d+(?:[-,(?:, )]+-?\d+)+(x\d)?)|(?:-?\d+))(?P<joined_padding>\#+|@+|%(?:\d+)?d|\$F(?:\d+)?))|(?P<embedded_frame_range>-?\d+(?:[-,(?:, )]+-?\d+)+(x\d)?)|(?P<padding>\#+|@+|%(?:\d+)?d|\$F(?:\d+)?))|(?P<frame_number>-?\d+)))
|
(?P>udim)
)
(?:(?P<tail>[_.][^/]+)?(?P<ext>[_.][a-zA-Z]+))?
(?: \[*(?P<extended_frame_range>-?\d+(?:[-,(?:, )]+-?\d+)+(x\d)?)\]*?)?
(?<!^)
$

regex101上的Regex演示

在Python 中

import regex

pat = '''^
(?:(?P<directory>(?<=^).*/)?(?P<head>[^/]+?)(?!$))??
(?:
(?P<udim>
(?:(?P<udim_separator>(?<=^)|[_.])[uU](?:(?P<udim_tile>[\d#@]{4})|(?P<udim_token><[Uu][Dd][Ii][Mm]>)))
)?
(?:(?P<frame_separator>(?<=^)|[_.])(?:(?P<frame_token>(?:(?P<joined_embedded_frame_range>(?:-?\d+(?:[-,(?:, )]+-?\d+)+(x\d)?)|(?:-?\d+))(?P<joined_padding>\#+|@+|%(?:\d+)?d|\$F(?:\d+)?))|(?P<embedded_frame_range>-?\d+(?:[-,(?:, )]+-?\d+)+(x\d)?)|(?P<padding>\#+|@+|%(?:\d+)?d|\$F(?:\d+)?))|(?P<frame_number>-?\d+)))
|
(?P>udim)
)
(?:(?P<tail>[_.][^/]+)?(?P<ext>[_.][a-zA-Z]+))?
(?: \[*(?P<extended_frame_range>-?\d+(?:[-,(?:, )]+-?\d+)+(x\d)?)\]*?)?
(?<!^)
$'''

rgx = regex.compile(pat, re.X)

files = ['tests/test_paths/udim_seq/head.1001',
    'tests/test_paths/udim_seq/book.open.mid.u1001.1-2#',
    'tests/test_paths/udim_seq/head.1001.exr',
    'tests/test_paths/udim_seq/book.open.mid.1001.exr',
    'tests/test_paths/udim_seq/book.open.mid.u1001.1-2#.exr'
]

for f in files:
    print(f, json.dumps(rgx.match(f).groupdict(), indent=4), sep='\n')

输出:

tests/test_paths/udim_seq/head.1001
{
    "directory": "tests/test_paths/udim_seq/",
    "head": "head",
    "udim": null,
    "udim_separator": null,
    "udim_tile": null,
    "udim_token": null,
    "frame_separator": ".",
    "frame_token": null,
    "joined_embedded_frame_range": null,
    "joined_padding": null,
    "embedded_frame_range": null,
    "padding": null,
    "frame_number": "1001",
    "tail": null,
    "ext": null,
    "extended_frame_range": null
}
tests/test_paths/udim_seq/book.open.mid.u1001.1-2#
{
    "directory": "tests/test_paths/udim_seq/",
    "head": "book.open.mid",
    "udim": ".u1001",
    "udim_separator": ".",
    "udim_tile": "1001",
    "udim_token": null,
    "frame_separator": ".",
    "frame_token": "1-2#",
    "joined_embedded_frame_range": "1-2",
    "joined_padding": "#",
    "embedded_frame_range": null,
    "padding": null,
    "frame_number": null,
    "tail": null,
    "ext": null,
}
tests/test_paths/udim_seq/head.1001.exr
{
    "directory": "tests/test_paths/udim_seq/",
    "head": "head",
    "udim": null,
    "udim_separator": null,
    "udim_tile": null,
    "udim_token": null,
    "frame_separator": ".",
    "frame_token": null,
    "joined_embedded_frame_range": null,
    "joined_padding": null,
    "embedded_frame_range": null,
    "padding": null,
    "frame_number": "1001",
    "tail": null,
    "ext": ".exr",
    "extended_frame_range": null
}
tests/test_paths/udim_seq/book.open.mid.1001.exr
{
    "directory": "tests/test_paths/udim_seq/",
    "head": "book.open.mid",
    "udim": null,
    "udim_separator": null,
    "udim_tile": null,
    "udim_token": null,
    "frame_separator": ".",
    "frame_token": null,
    "joined_embedded_frame_range": null,
    "joined_padding": null,
    "embedded_frame_range": null,
    "padding": null,
    "frame_number": "1001",
    "tail": null,
    "ext": ".exr",
    "extended_frame_range": null
}
tests/test_paths/udim_seq/book.open.mid.u1001.1-2#.exr
{
    "directory": "tests/test_paths/udim_seq/",
    "head": "book.open.mid",
    "udim": ".u1001",
    "udim_separator": ".",
    "udim_tile": "1001",
    "udim_token": null,
    "frame_separator": ".",
    "frame_token": "1-2#",
    "joined_embedded_frame_range": "1-2",
    "joined_padding": "#",
    "embedded_frame_range": null,
    "padding": null,
    "frame_number": null,
    "tail": null,
    "ext": ".exr",
    "extended_frame_range": null
}

Python相关问答推荐

单击cookie按钮，但结果不一致

try 使用tensorFlow.keras.models时optree Import错误

合并其中一个具有重叠范围的两个框架的最佳方法是什么？

为什么我的(工作)代码(生成交互式情节)在将其放入函数中时不再工作？

使用matplotlib pcolormesh，如何停止从一行绘制的磁贴连接到上下行？

为什么dict(id=1，**{id：2})有时会引发KeyMessage：id而不是TypMessage？

根据网格和相机参数渲染深度

try 与gemini-pro进行多轮聊天时出错

根据不同列的值在收件箱中移动数据

如何在Windows上用Python提取名称中带有逗号的文件？

切片包括面具的第一个实例在内的眼镜的最佳方法是什么？

在Python argparse包中添加formatter_class MetavarTypeHelpFormatter时， - help不再工作""""

Pandas DataFrame中行之间的差异

所有列的滚动标准差，忽略NaN

如何禁用FastAPI应用程序的Swagger UI autodoc中的application/json？

Tkinter菜单自发添加额外项目

LocaleError：模块keras._' tf_keras. keras没有属性__internal_'''

找到相对于列表索引的当前最大值列表""

如果包含特定值，则筛选Groupby

在用于Python的Bokeh包中设置按钮的样式

实用课程推荐
相关教程推荐
休闲君评测 |  TOOLFK工具网 |  古诗文网 |  良许Linux教程网 |  Python之禅 |  红色石头 |  入党申请书 | 

© 2020 Hi LearnFk 意见反馈 免责声明 关于我们 桂ICP备11002319号-3 友链&广告位+QQ: 1963612630