semantique.processor.types.DTYPE_MAPPING

semantique.processor.types.DTYPE_MAPPING = {'M': ['datetime'], 'O': ['ordinal', 'nominal'], 'U': ['ordinal', 'nominal'], 'b': ['binary'], 'f': ['continuous'], 'i': ['continuous', 'discrete', 'ordinal', 'nominal'], 'u': ['continuous', 'discrete', 'ordinal', 'nominal']}

Mapping between numpy data types and semantique value types.

A semantique value type describes what kind of data an array contains. It differs from the very technical, computer-oriented numpy dtype categorization, which contains e.g. int, float, etc. Instead, the semantique value type describes data on a more general, statistical level. Each numpy data type can be mapped to one or semantique value types.

In this mapping, numpy data types are expressed by their .kind property. Each numpy data type may be mapped to more than one semantique value type, meaning that it can represent any of these value types.

Type:

dict