Keras 3 backend selection before import: tf.keras aliasing and legacy API support

Goal

Determine how Keras 3 enforces backend selection prior to the first import and how the tf.keras aliasing layer preserves or drops legacy Keras 2 symbols when different backends are chosen.

Constraints

Backend must be chosen via KERAS_BACKEND or keras.config.set_backend() before any import keras. Switching backends later in the same process is unsupported. The tf.keras aliasing available in TensorFlow 2.16+ may not expose all removed or relocated Keras 2 symbols, and behavior varies with the TensorFlow/Keras release pair.

Questions

1. When KERAS_BACKEND is unset and tf.keras is imported in a Keras 3 environment using the TensorFlow backend, which legacy optimizer symbols remain resolvable through the aliasing layer?
2. If KERAS_BACKEND is set to "jax" before import, does the tf.keras aliasing mechanism still activate, and which Keras 2 optimizer symbols are accessible?
3. Across which Keras 3 release series and TensorFlow version combinations does the tf.keras aliasing behavior for removed legacy optimizer symbols become formally documented and deterministic?