正则If-then、前瞻断言与括号应用：提取Python指定import语句需求

Got it, let's solve this problem of matching both styles of Python import statements using a single regex with grep. First, let's recap the two formats we need to cover:

1. Parenthesized imports (with optional newlines inside)
   Example:

   from mymodule import (
       func_a,
       ClassB,
       CONSTANT_C
   )
   

2. Line-continued imports (using \, no internal newlines without the line-continuation character)
   Example:

   from another_module import func_x, ClassY, \
        CONSTANT_Z
   

The Combined Regular Expression #

Here's a regex that handles both cases, optimized for use with grep (using Perl-compatible mode):

^from\s+([\w.]+)\s+import\s+(?:\((?:\s|\n)*([\w\s,]+)(?:\s|\n)*\)|([\w,]+(?:\\\n[\w, ]+)*))$

Breakdown of the Regex #

Let's break this down piece by piece so you can tweak it if needed:

• ^from\s+: Anchors to the start of a line, matches the literal from followed by one or more spaces.
• ([\w.]+): Captures the module name (supports dots for nested packages like my.package.module).
• \s+import\s+: Matches the import keyword surrounded by spaces.
• (?:...): A non-capturing group that wraps our two import style options, separated by | (OR):
  1. Parenthesized branch: \((?:\s|\n)*([\w\s,]+)(?:\s|\n)*\)
     • \(: Matches the opening parenthesis.
     • (?:\s|\n)*: Allows any number of spaces or newlines inside the parentheses (before, between, and after imports).
     • ([\w\s,]+): Captures the list of imports (words, spaces, commas).
     • \): Matches the closing parenthesis.
  2. Line-continued branch: ([\w,]+(?:\\\n[\w, ]+)*)
     • [\w,]+: Matches the first part of the import list (words and commas).
     • (?:\\\n[\w, ]+)*: Allows any number of line-continued segments (matches the \ followed by a newline, then more imports with spaces/commas).
• $: Anchors to the end of the line (works with grep's multiline mode when using -P).

Using with Grep #

To run this on your Python file and extract only the matching import statements, use this command:

grep -Pzo '^from\s+([\w.]+)\s+import\s+(?:\((?:\s|\n)*[\w\s,]+\s*\)|[\w,]+(?:\\\n[\w, ]+)*)$' your_script.py

Let's explain the flags:

• -P: Enables Perl-compatible regular expressions (required for non-capturing groups and proper newline handling).
• -z: Treats the input as a single NUL-terminated string, which lets the regex match across newlines (critical for parenthesized imports with internal line breaks).
• -o: Only outputs the parts of the file that match the regex, not the entire lines.

Quick Adjustment Tips #

• If you don't need to capture the module name specifically, you can replace ([\w.]+) with \S+ to simplify.
• For stricter matching (e.g., disallowing extra spaces around commas), tweak the [\w\s,]+ parts to something like [\w]+(?:,\s*[\w]+)*.

内容的提问来源于stack exchange，提问作者Brad Solomon