Performance

The writer/encoder supports extra Python types in addition to standard JSON compatible types (int, float, bool, str, list and dict). This comes at a cost of heavy Python side type check.

The performance can be further improved if some features are not used.

Plain Container

If the target data only contains standard Python list and dict, disable custom container check.

1
2
3

from msglc.config import configure

configure(enable_custom_container=False)

No Other List Types

By default, sets will be sorted and packed as plain lists, tuples will be packed as plain lists, numpy.ndarray can be conditionally converted to plain lists or directly encoded as binary blob using numpy's own encoder. Those can be disabled if the target data does not contain those objects.

1
2
3

from msglc.config import configure

configure(has_numpy=False, has_set=False, has_tuple=False)