collections designed for ML usecases

 The package provices two classes called ConfigDict and FrozenConfigDict that
 are "dict-like" data structures with dot access to nested elements. Together,
 they are supposed to be used as a main way of expressing configurations of
 experiments and models.
 .
 Features:
  * Dot-based access to fields.
  * Locking mechanism to prevent spelling mistakes.
  * Lazy computation.
  * FrozenConfigDict() class which is immutable and hashable.
  * Type safety.
  * "Did you mean" functionality.
  * Human readable printing (with valid references and cycles), using
    valid YAML format.
  * Fields can be passed as keyword arguments using the ** operator.
 .
 There are two exceptions to the strong type-safety of the ConfigDict.
 int values can be passed in to fields of type float. In such a case, the
 value is type-converted to a float before being stored. Similarly,
 all string types (including Unicode strings) can be stored in fields
 of type str or unicode.