python-pcre2

A simple Python wrapper of PCRE2

Requirements

• Python 3.5+
• Cython
• libpcre2 + header file

Example

>>> import pcre2
>>> p = pcre2.PCRE2(r'(hello) (.+)'.encode())  # compile
>>> result = p.search('hello world'.encode())
>>> result.groups()
[b'hello', b'world']
>>> result.group(0)
b'hello world'
>>> result.group(1)
b'hello'
>>> result.group(2)
b'world'

For non-Latin word (\w) match, you should specify UTF and UCP option:

>>> p = pcre2.PCRE2(r'(北京)(\w+)广场'.encode(), pcre2.UTF | pcre2.UCP)
>>> result = p.search('我来到北京敏感词广场看日出东方'.encode())
>>> result.groups()
[b'\xe5\x8c\x97\xe4\xba\xac', b'\xe6\x95\x8f\xe6\x84\x9f\xe8\xaf\x8d']
>>> result.group(0).decode()
'北京敏感词广场'

Notes

• Pattern and Searched Text should be bytes type with UTF-8 charset.
• Matched result is also encoded with UTF-8

UTF-16/32 Support

If you want to match UTF-16 bytes (UTF-32 is similar), you could modify the settings in setup.py and then reinstall.

extra_compile_args=['-DPCRE2_CODE_UNIT_WIDTH=8']
  => extra_compile_args=['-DPCRE2_CODE_UNIT_WIDTH=16']

libraries=['pcre2-8']
  => libraries=['pcre2-16']