[FEATURE] Corpus for input values

[Stranger6667]

Is your feature request related to a problem? Please describe

Sometimes the user wants to test a large set of specific values to ensure corner cases are always covered.

Describe the solution you'd like

A way to load such a corpus conveniently. The corpus format is not yet clear but at first glance, it could be done on the Open API parameters basis, i.e. they are named so it could be just a mapping param => [variant_A, variant_B, ...]. It would also correspond to a CSV file structure, where each column will have values for each parameter.

However, it is unclear how to deal with other parameters, e.g. inside components - the user may want to cover some specific properties of objects nested inside components.

Possible Python API:

from schemathesis.corpus import Entry


schema.corpus.add(
    Entry.from_path(
        path="#/components/User/name",
        values=["a", "b", "c"],
        filter=...  # Use values only in some specific API operation
    ),
    Entry.named(
        # All `user_id` parameters
        name="user_id",
        values=["42", "43", "44"]
    ),
)

CLI:

st run --corpus-path=...

However, it is unclear how such a file should be structured.

Describe alternatives you've considered

Currently, it is not super convenient to do so. The closest equivalent is the add_examples hook, but it requires constructing examples manually. Alternatively, the user may modify their API schema which is not feasible for a large set of values.

Additional context

Discussed privately in Discord - maybe JMeter's input values could be a source of inspiration for this.