Deterministic (seeded) enum values

This is a follow up issue for https://github.com/ardatan/graphql-tools/issues/2272 (where it is suggested to open a new issue).

**Is your feature request related to a problem? Please describe.**

Some parts of the mock data generation process use their own source of randomness and are difficult to work around. Specifically the generation of enum values.

**Describe the solution you'd like**

The API could accept a function that returns a random value which is used instead of Math.random(). That would allow the user to make the default behavior deterministic.

**Describe alternatives you've considered**

Providing custom data generators for every enum type. It's impractical for larger schemas and tedious to account for schema changes.
Generators for enum values could themselves be generated with a custom codgen plugin, but that also requires an understanding of the codegen process.

The issue names no files, tests, or entry points; start by locating enum value generation in the mock data generation path and inspect how randomness is sourced. Done means the API can accept a caller-provided random-value function for enum generation, with deterministic behavior demonstrated by tests.

Written by the indexing model from the issue text.