tff.federated_select

Sends selected values from a server database to clients.

tff.federated_select(
    client_keys, max_key, server_val, select_fn
)

Used in the tutorials
Sending Different Data To Particular Clients With tff.federated_select

Args
`client_keys`	`tff.CLIENTS`-placed one-dimensional fixed-size non-negative `int32` keys used to select values from `database` to load for each client.
`max_key`	A `tff.SERVER`-placed `int32` indicating the maximum value of any of `client_keys` (so that all client keys are in the range `[0, max_key]`, inclusive). Lower values may permit more optimizations.
`server_val`	`tff.SERVER`-placed value used as an input to `select_fn`.
`select_fn`	A `tff.Computation` which accepts unplaced `server_val` and a `int32` client key and returns a value to be sent to the client. `select_fn` should be deterministic (nonrandom).

Returns
`tff.CLIENTS`-placed sequences of values returned from `select_fn`. In each sequence, the order of values will match the order of keys in the corresponding `client_keys` tensor. For example, a client with keys `[1, 2, ...]` will receive a sequence of values `[select_fn(server_val, 1), select_fn(server_val, 2), ...]`.

Raises
`TypeError`	If `client_keys` is not of type `{int32[N]}@CLIENTS`, if `max_key` is not of type `int32@SERVER`, if `server_val` is not a server-placed value (`S@SERVER`), or if `select_fn` is not a function of type `<S, int32> -> RESULT`.

Args