The lmdb backend does not seem to work properly in real world tests. Excessive CPU usage and data corruption seems to occur. Do not use this package (and probably not the python-lmdb package either).
0.0.6
ShockDB now saves the serializers and compressors at file creation, so that using the correct serializers and compressors at file reading is guaranteed.
User defined serializers and compressors can now be passed to the open function.