JPEG-LS near-lossless doesn't allow error bounds greater than 9

[Messier-16]

Especially when compressing uint16 images, it is important for me to be able to allow an error bound of more than 9, but for some reason, line 92 in _jpegls.pyx restricts the max value to 9: int32_t near_lossless = _default_value(level, 0, 0, 9)

[cgohlke]

Thank you for reporting the issue. The upper limit will be removed in the next version.