Real embeddings are learned representations. This course uses small word-count vectors so you can inspect every dimension and test the surrounding pipeline without a model call. That simplification teaches input ordering and indexing; it cannot establish semantic search quality for a production corpus.
Call a fresh MockClient.embeddings.create with case.texts and model fixture-vector. Return each data item embedding in input order. These four-dimensional word-count fixtures are not learned semantic vectors.
The notebook executes your Python against an original local mock. Use the supplied case rather than copying visible values. The hidden checks change identities, revisions and boundary inputs; no provider endpoint or external tool is contacted.
An example implementation of the function contract.