Generate Values within Stated Bounds

Task

Write generate_values_within_bounds(rng, kind, shape, lower, upper).

The function receives one already-created NumPy Generator. It must use that generator for exactly one request and return a new dictionary with exactly these keys:

values, evidence

values is a new NumPy array with the requested shape. evidence is a new dictionary with exactly these keys:

kind, shape, dtype, lower, upper, minimum, maximum

The two supported kind values define the complete generation rule:

• For "uniform", lower and upper are finite numbers with lower < upper. Generate float64 values from the half-open interval [lower, upper). The lower bound is included and the upper bound is excluded.
• For "integers", lower and upper are integers with lower < upper. Generate int64 values from the half-open interval [lower, upper).

shape must be a non-empty tuple or list of positive integers. The returned evidence["shape"] is a tuple, and evidence["dtype"] is the string form of the returned array's dtype. Store the requested bounds in lower and upper. Store the smallest and largest generated values in minimum and maximum as ordinary Python numbers. The generated values must satisfy the stated bounds.

Do not create or seed another generator. Do not change shape, lower, upper, or any other argument. A valid call advances the supplied generator by the one request described above. Invalid arguments must raise ValueError before requesting values.

Example

The exact values depend on the generator state. The shape, dtype, interval, and range evidence are the reusable result of the function.

An integer request uses the same interface but a different local rule:

Every value is an integer in {0, 1, 2, 3}. The value 4 is not allowed.

Your implementation

You may import NumPy as np. Do not print, create a generator, seed global state, or ask for input.