Communication

class optunity.communication.EvalManager(max_vectorized=100, replacements={})

Bases: object

Constructs an EvalManager object.

Parameters:

• max_vectorized (int) – the maximum size of a vector evaluation larger vectorizations will be chunked
• replacements (dict) – a mapping of original:replacement keyword names

add_to_queue(**kwargs)

cv

flush_queue()

get(number)

max_vectorized

pipe_eval(**kwargs)

pmap(f, *args)

Performs vector evaluations through pipes.

Parameters:

• f (callable) – the objective function (piped_function_eval)
• args (iterables) – function arguments

The vector evaluation is sent in chunks of size self.max_vectorized.

processed_semaphore

queue

queue_lock

replacements

Keys that must be replaced: keys are current values, values are what must be sent through the pipe.

semaphore

vectorized

optunity.communication.accept_server_connection(server_socket)

optunity.communication.json_decode(data)

Decodes given JSON string and returns its data.

>>> orig = {1: 2, 'a': [1,2]}
>>> data = json_decode(json_encode(orig))
>>> data[str(1)]
2
>>> data['a']
[1, 2]

optunity.communication.json_encode(data)

Encodes given data in a JSON string.

optunity.communication.make_piped_function(mgr)

optunity.communication.open_server_socket()

optunity.communication.open_socket(port, host='localhost')

Opens a socket to host:port and reconfigures internal channels.

optunity.communication.receive()

Reads data from channel.

optunity.communication.send(data)

Writes data to channel and flushes.