Core Concepts¶

DataPoint¶

A DataPoint is a single "piece" of data that may be related to other datapoints via being an input to or output of resolvers. Each key in a DataMap corresponds to a single datapoint. A datapoint in other contexts is sometimes called an "attribute", "property", or "field" -- they are essentially the nodes on the data dependency graph.

Note

A DataPoint is not necessarily a single value, it may be a series of data values, or a column of a dataset, or even a whole data table. The important part is that it is a unit of data that is derived in total from other units of data.

Resolver¶

Resolvers are python functions that establish relationships between datapoints. A resolver relates the input datapoints, which correspond to arguments to the function, to the datapoint represented by the output of the function. Resolvers work with lambda or regular python functions defined with def, as well as functions cached with functools.lru_cache.

DataMap¶

A DataMap is a declaration of all datapoints and associated dependency datapoints & resolvers in a system. A DataMap is declared as a regular python dict that must adhere to a particular schema--see DataMap Reference for more.