response.py 文件源码

def encoding(self):
        """
        encoding of Response.content.

        if Response.encoding is None, encoding will be guessed
        by header or content or chardet if available.
        """
        if hasattr(self, '_encoding'):
            return self._encoding

        # content is unicode
        if isinstance(self.content, six.text_type):
            return 'unicode'

        # Try charset from content-type
        encoding = get_encoding_from_headers(self.headers)
        if encoding == 'ISO-8859-1':
            encoding = None

        # Try charset from content
        if not encoding and get_encodings_from_content:
            if six.PY3:
                encoding = get_encodings_from_content(utils.pretty_unicode(self.content[:100]))
            else:
                encoding = get_encodings_from_content(self.content)
            encoding = encoding and encoding[0] or None

        # Fallback to auto-detected encoding.
        if not encoding and chardet is not None:
            encoding = chardet.detect(self.content)['encoding']

        if encoding and encoding.lower() == 'gb2312':
            encoding = 'gb18030'

        self._encoding = encoding or 'utf-8'
        return self._encoding