Hint for using read_exif, read_xmp, read_iptc, read_comment

[RalfPeter]

Good morning, "For all those who have encountered difficulties using the functions read_exif, read_xmp, and read_iptc. From time to time, my script crashed with a runtime error without any further error message. I suspected the cause was in the coding of exiv2 and experimented with different images. I discovered that some images contained metadata that was not UTF-8 encoded, but rather ISO-8859-1. So I wrote the following routines (analogously, of course, for XMP, IPTC, and comments as well). Perhaps someone will find it helpfull:

    # -------------------------------------------------------------------------------------------
    def _log_unicode_decode_details(self, e: UnicodeDecodeError, verbose=False):
        error_details = [e.encoding, e.object, e.start, e.end, e.reason]
        # Übergabe an die log-Funktion
        if verbose:
            log('EXIF unicode_decode', f"Fehler im Foto {self.file.name}: Typ: {type(e).__name__}", *error_details, level=ERROR)

    # -------------------------------------------------------------------------------------------
    def _read_exif(self):
        # encoding standard is utf-8, but could be iso-8859-1
        try:
            data = self.image.read_exif(encoding=ENCODING_UTF)
        except UnicodeDecodeError as e:
            # Übergabe an die log-Funktion
            self._log_unicode_decode_details(e)
            try:
                data = self.image.read_exif(encoding=ENCODING_ISO)
            except UnicodeDecodeError as e:
                # Übergabe an die log-Funktion
                self._log_unicode_decode_details(e, verbose=True)
                raise UnicodeDecodeError

        return data

[LeoHsiao1]

Thank you for sharing. ISO-8859-1 is a special encoding format that uses all encoded values in the range 0x00 ~ 0xFF. Therefore, bytes in any encoded format can be decoded with ISO-8859-1 format, but the original meaning may be lost. For example:

• Suppose you encode a Chinese string:

  >>> data = '你好'.encode('gbk')

• It can not be decoded with utf-8:

  >>> data.decode('utf-8')
  UnicodeDecodeError: 'utf-8' codec can't decode byte 0xc4 in position 0: invalid continuation byte

• If you know that the data are encoded with gbk format, you can decode it with the same format:

  >>> data.decode('gbk')
  '你好'

• You can decode it with ISO-8859-1 format, which won't report errors, but you won't get the original Chinese characters.

  >>> data.decode('ISO-8859-1') 
  'ÄãºÃ'

In conclusion, ISO-8859-1 is not a one-size-fits-all encoding format. Users are advised to find the original encoding format of the data.

[LeoHsiao1]

I am closing this issue now. Hopefully others with the same problem will search for this issue.

[RalfPeter]

Thank you Leo, my only idea was that anybody with the same problem will have an advice how to use encoding with your pyexiv2. Thank you for your additional comment.