A wrapper to execute Hyperopt cost functions safely on a separate process, with timeout

hyperopt_wrap_cost.py

"""A wrapper to execute Hyperopt cost functions safely on a separate process, with timeout

This code is based off parts of https://github.com/hyperopt/hyperopt-sklearn,
which falls under the following license:

=======

Copyright (c) 2013, James Bergstra
All rights reserved.

Redistribution and use in source and binary forms, with or without
modification, are permitted provided that the following conditions are met:

    * Redistributions of source code must retain the above copyright
      notice, this list of conditions and the following disclaimer.
    * Redistributions in binary form must reproduce the above copyright
      notice, this list of conditions and the following disclaimer in the
      documentation and/or other materials provided with the distribution.
    * Neither the name of hyperopt-sklearn nor the names of its contributors
      may be used to endorse or promote products derived from this software
      without specific prior written permission.

THIS SOFTWARE IS PROVIDED BY THE COPYRIGHT HOLDERS ''AS IS'' AND ANY
EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE IMPLIED
WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE ARE
DISCLAIMED. IN NO EVENT SHALL THE COPYRIGHT HOLDERS BE LIABLE FOR ANY
DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL DAMAGES
(INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES;
LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION) HOWEVER CAUSED AND
ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT LIABILITY, OR TORT
(INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY OUT OF THE USE OF THIS
SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE.
"""
from __future__ import absolute_import

from multiprocessing import Process, Pipe
import sys
import time

import hyperopt
import numpy as np

PY2 = sys.version_info[0] == 2
int_types = (int, long) if PY2 else (int,)


def is_integer(obj):
    return isinstance(obj, int_types + (np.integer,))


def is_number(obj, check_complex=False):
    types = ((float, complex, np.number) if check_complex else
             (float, np.floating))
    return is_integer(obj) or isinstance(obj, types)


def get_vals(trial):
    """Determine hyperparameter values given a ``Trial`` object"""
    # based on hyperopt/base.py:Trials:argmin
    return dict((k, v[0]) for k, v in trial['misc']['vals'].items() if v)


def wrap_cost(cost_fn, timeout=None, iters=1, verbose=0):
    """Wrap cost function to execute trials safely on a separate process.

    Parameters
    ----------
    cost_fn : callable
        The cost function (aka. objective function) to wrap. It follows the
        same specifications as normal Hyperopt cost functions.
    timeout : int
        Time to wait for process to complete, in seconds. If this time is
        reached, the process is re-tried if there are remaining iterations,
        otherwise marked as a failure. If ``None``, wait indefinitely.
    iters : int
        Number of times to allow the trial to timeout before marking it as
        a failure due to timeout.
    verbose : int
        How verbose this function should be. 0 is not verbose, 1 is verbose.

    Example
    -------
    def objective(args):
        case, val = args
        return val**2 if case else val

    space = [hp.choice('case', [False, True]), hp.uniform('val', -1, 1)]
    safe_objective = wrap_cost(objective, timeout=2, iters=2, verbose=1)
    best = hyperopt.fmin(safe_objective, space, max_evals=100)

    Notes
    -----
    Based on code from https://github.com/hyperopt/hyperopt-sklearn
    """
    def _cost_fn(*args, **kwargs):
        _conn = kwargs.pop('_conn')
        try:
            t_start = time.time()
            rval = cost_fn(*args, **kwargs)
            t_done = time.time()

            if not isinstance(rval, dict):
                rval = dict(loss=rval)
            assert 'loss' in rval, "Returned dictionary must include loss"
            loss = rval['loss']
            assert is_number(loss), "Returned loss must be a number type"
            rval.setdefault('status', hyperopt.STATUS_OK if np.isfinite(loss)
                            else hyperopt.STATUS_FAIL)
            rval.setdefault('duration', t_done - t_start)
            rtype = 'return'

        except Exception as exc:
            rval = exc
            rtype = 'raise'

        # -- return the result to calling process
        _conn.send((rtype, rval))

    def wrapper(*args, **kwargs):
        for k in range(iters):
            conn1, conn2 = Pipe()
            kwargs['_conn'] = conn2
            th = Process(target=_cost_fn, args=args, kwargs=kwargs)
            th.start()
            if conn1.poll(timeout):
                fn_rval = conn1.recv()
                th.join()
            else:
                if verbose >= 1:
                    print("TRIAL TIMED OUT (%d/%d)" % (k+1, iters))
                th.terminate()
                th.join()
                continue

            assert fn_rval[0] in ('raise', 'return')
            if fn_rval[0] == 'raise':
                raise fn_rval[1]
            else:
                return fn_rval[1]

        return {'status': hyperopt.STATUS_FAIL,
                'failure': 'timeout'}

    return wrapper

Thanks! Will use.

I have been trying to adapt this method from a linux environement (where it is working fine !) to a mac OS environment (macOS Big Sur 11.5.2, python 3.9 Spyder 5.0.5), and i am getting this error :
'job exception: Can't pickle local object 'wrap_cost.._cost_fn'
The error is occurring when i run the line :
best = fmin(objective,timeout = 7200, space = space, algo =mix_algo, max_evals = 250, trials = trials)
from hyperopt package. Can't really debug it... Anyone has found the solution to this ?