Example #1
0
    def __init__(self, experiment, max_idle_time=None):
        """Initialize a producer.

        :param experiment: Manager of this experiment, provides convenient
           interface for interacting with the database.
        """
        log.debug("Creating Producer object.")
        self.experiment = experiment
        self.space = experiment.space
        if self.space is None:
            raise RuntimeError(
                "Experiment object provided to Producer has not yet completed"
                " initialization.")
        self.algorithm = experiment.algorithms
        self.algorithm.algorithm.max_trials = experiment.max_trials
        if max_idle_time is None:
            max_idle_time = orion.core.config.worker.max_idle_time
        self.max_idle_time = max_idle_time
        self.strategy = experiment.producer['strategy']
        self.naive_algorithm = None
        # TODO: Move trials_history into PrimaryAlgo during the refactoring of Algorithm with
        #       Strategist and Scheduler.
        self.trials_history = TrialsHistory()
        self.params_hashes = set()
        self.naive_trials_history = None
        self.failure_count = 0
Example #2
0
def test_history_contains_new_child():
    """Verify that __contains__ return True for a new child"""
    trials_history = TrialsHistory()
    new_child = DummyTrial(1, [])
    assert new_child not in trials_history
    trials_history.update([new_child])
    assert new_child in trials_history
Example #3
0
def test_discarded_children():
    """Verify that ancestors of new children are discarded from history"""
    trials_history = TrialsHistory()
    trials = [DummyTrial(i, []) for i in range(3)]
    trials_history.update(trials)
    assert trials_history.children == [0, 1, 2]

    trials_history.update(trials)
    assert trials_history.children == [0, 1, 2]

    trials = [DummyTrial(i, [trials[i % 3].id]) for i in range(3, 6)]
    trials_history.update(trials)
    assert trials_history.children == [3, 4, 5]
Example #4
0
def test_added_children_without_ancestors():
    """Verify that children are added to history"""
    trials_history = TrialsHistory()
    trials_history.update([DummyTrial(i, []) for i in range(3)])
    assert trials_history.children == [0, 1, 2]
    trials_history.update([DummyTrial(i, []) for i in range(3, 6)])
    assert trials_history.children == [0, 1, 2, 3, 4, 5]
Example #5
0
def test_discarded_shared_children():
    """Verify that only ancestors are removed and not all past children"""
    trials_history = TrialsHistory()
    trials = [DummyTrial(i, []) for i in range(3)]
    trials_history.update(trials)
    assert trials_history.children == [0, 1, 2]

    trials = [DummyTrial(i, [0]) for i in range(3, 6)]
    trials_history.update(trials)
    assert trials_history.children == [1, 2, 3, 4, 5]
Example #6
0
def test_added_children_with_ancestors():
    """Verify that children with ancestors are added to history"""
    trials_history = TrialsHistory()
    trials = [DummyTrial(i, []) for i in range(3)]
    trials_history.update(trials)
    assert trials_history.children == [0, 1, 2]

    trials = [DummyTrial(i, [trials[i % 3].id]) for i in range(3, 6)]
    trials_history.update(trials)
    assert len(set(trials_history.children) & set([3, 4, 5])) == 3
Example #7
0
def test_discarded_duplicate_children():
    """Verify that duplicate children are not added twice"""
    trials_history = TrialsHistory()
    trials = [DummyTrial(i, []) for i in range(3)]
    trials_history.update(trials)
    assert trials_history.children == [0, 1, 2]

    trials = [DummyTrial(i, [trials[i].id]) for i in range(3)]
    assert all(trial.id == trial.parents[0] for trial in trials)
    trials_history.update(trials)
    assert trials_history.children == [0, 1, 2]
Example #8
0
def test_history_contains_old_child():
    """Verify that __contains__ return True for a new child"""
    trials_history = TrialsHistory()
    old_child = DummyTrial(1, [])
    trials_history.update([old_child])
    new_child = DummyTrial(2, [old_child.id])
    assert new_child not in trials_history
    trials_history.update([new_child])
    assert old_child.id not in trials_history.children
    assert old_child in trials_history
    assert new_child.id in trials_history.children
    assert new_child in trials_history
Example #9
0
class Producer(object):
    """Produce suggested sets of problem's parameter space to try out.

    It uses an `Experiment` object to poll for not yet observed trials which
    have been already evaluated and to register new suggestions (points of
    the parameter `Space`) to be evaluated.

    """
    def __init__(self, experiment, max_attempts=100):
        """Initialize a producer.

        :param experiment: Manager of this experiment, provides convenient
           interface for interacting with the database.
        """
        log.debug("Creating Producer object.")
        self.experiment = experiment
        self.space = experiment.space
        if self.space is None:
            raise RuntimeError(
                "Experiment object provided to Producer has not yet completed"
                " initialization.")
        self.algorithm = experiment.algorithms
        self.max_attempts = max_attempts
        self.strategy = experiment.producer['strategy']
        self.naive_algorithm = None
        # TODO: Move trials_history into PrimaryAlgo during the refactoring of Algorithm with
        #       Strategist and Scheduler.
        self.trials_history = TrialsHistory()
        self.naive_trials_history = None

    @property
    def pool_size(self):
        """Pool-size of the experiment"""
        return self.experiment.pool_size

    def produce(self):
        """Create and register new trials."""
        sampled_points = 0
        n_attempts = 0

        while sampled_points < self.pool_size and n_attempts < self.max_attempts:
            n_attempts += 1
            log.debug("### Algorithm suggests new points.")

            new_points = self.naive_algorithm.suggest(self.pool_size)
            # Sync state of original algo so that state continues evolving.
            self.algorithm.set_state(self.naive_algorithm.state_dict)

            for new_point in new_points:
                log.debug("#### Convert point to `Trial` object.")
                new_trial = format_trials.tuple_to_trial(new_point, self.space)
                try:
                    new_trial.parents = self.naive_trials_history.children
                    log.debug("#### Register new trial to database: %s",
                              new_trial)
                    self.experiment.register_trial(new_trial)
                    sampled_points += 1
                except DuplicateKeyError:
                    log.debug(
                        "#### Duplicate sample. Updating algo to produce new ones."
                    )
                    self.update()
                    break

        if n_attempts >= self.max_attempts:
            raise RuntimeError(
                "Looks like the algorithm keeps suggesting trial configurations"
                "that already exist in the database. Could be that you reached "
                "a point of convergence and the algorithm cannot find anything "
                "better. Or... something is broken. Try increasing `max_attempts` "
                "or please report this error on "
                "https://github.com/epistimio/orion/issues if something looks "
                "wrong.")

    def update(self):
        """Pull all trials to update model with completed ones and naive model with non completed
        ones.
        """
        trials = self.experiment.fetch_trials({})
        self._update_algorithm(
            [trial for trial in trials if trial.status == 'completed'])
        self._update_naive_algorithm(
            [trial for trial in trials if trial.status != 'completed'])

    def _update_algorithm(self, completed_trials):
        """Pull newest completed trials to update local model."""
        log.debug("### Fetch completed trials to observe:")

        new_completed_trials = []
        for trial in completed_trials:
            if trial not in self.trials_history:
                new_completed_trials.append(trial)

        log.debug("### %s", new_completed_trials)

        if new_completed_trials:
            log.debug("### Convert them to list of points and their results.")
            points = list(
                map(
                    lambda trial: format_trials.trial_to_tuple(
                        trial, self.space), new_completed_trials))
            results = list(
                map(format_trials.get_trial_results, new_completed_trials))

            log.debug("### Observe them.")
            self.trials_history.update(new_completed_trials)
            self.algorithm.observe(points, results)
            self.strategy.observe(points, results)

    def _produce_lies(self, incomplete_trials):
        """Add fake objective results to incomplete trials

        Then register the trials in the db
        """
        log.debug("### Fetch active trials to observe:")
        lying_trials = []
        log.debug("### %s", incomplete_trials)

        for trial in incomplete_trials:
            log.debug(
                "### Use defined ParallelStrategy to assign them fake results."
            )
            lying_result = self.strategy.lie(trial)
            if lying_result is not None:
                lying_trial = copy.deepcopy(trial)
                lying_trial.results.append(lying_result)
                lying_trials.append(lying_trial)
                log.debug("### Register lie to database: %s", lying_trial)
                lying_trial.parents = self.trials_history.children
                try:
                    self.experiment.register_lie(lying_trial)
                except DuplicateKeyError:
                    log.debug(
                        "#### Duplicate lie. No need to register a duplicate in DB."
                    )

        return lying_trials

    def _update_naive_algorithm(self, incomplete_trials):
        """Pull all non completed trials to update naive model."""
        self.naive_algorithm = copy.deepcopy(self.algorithm)
        self.naive_trials_history = copy.deepcopy(self.trials_history)
        log.debug("### Create fake trials to observe:")
        lying_trials = self._produce_lies(incomplete_trials)
        log.debug("### %s", lying_trials)
        if lying_trials:
            log.debug("### Convert them to list of points and their results.")
            points = list(
                map(
                    lambda trial: format_trials.trial_to_tuple(
                        trial, self.space), lying_trials))
            results = list(map(format_trials.get_trial_results, lying_trials))

            log.debug("### Observe them.")
            self.naive_trials_history.update(lying_trials)
            self.naive_algorithm.observe(points, results)
Example #10
0
class Producer(object):
    """Produce suggested sets of problem's parameter space to try out.

    It uses an `Experiment` object to poll for not yet observed trials which
    have been already evaluated and to register new suggestions (points of
    the parameter `Space`) to be evaluated.

    """
    def __init__(self, experiment, max_idle_time=None):
        """Initialize a producer.

        :param experiment: Manager of this experiment, provides convenient
           interface for interacting with the database.
        """
        log.debug("Creating Producer object.")
        self.experiment = experiment
        self.space = experiment.space
        if self.space is None:
            raise RuntimeError(
                "Experiment object provided to Producer has not yet completed"
                " initialization.")
        self.algorithm = experiment.algorithms
        self.algorithm.algorithm.max_trials = experiment.max_trials
        if max_idle_time is None:
            max_idle_time = orion.core.config.worker.max_idle_time
        self.max_idle_time = max_idle_time
        self.strategy = experiment.producer['strategy']
        self.naive_algorithm = None
        # TODO: Move trials_history into PrimaryAlgo during the refactoring of Algorithm with
        #       Strategist and Scheduler.
        self.trials_history = TrialsHistory()
        self.params_hashes = set()
        self.naive_trials_history = None
        self.failure_count = 0

    @property
    def pool_size(self):
        """Pool-size of the experiment"""
        return self.experiment.pool_size

    def backoff(self):
        """Wait some time and update algorithm."""
        waiting_time = max(0, random.gauss(1, 0.2))
        log.info('Waiting %d seconds', waiting_time)
        time.sleep(waiting_time)
        log.info('Updating algorithm.')
        self.update()
        self.failure_count += 1

    def _sample_guard(self, start):
        """Check that the time taken sampling is less than max_idle_time"""
        if time.time() - start > self.max_idle_time:
            raise SampleTimeout(
                "Algorithm could not sample new points in less than {} seconds."
                "Failed to sample points {} times".format(
                    self.max_idle_time, self.failure_count))

    def produce(self):
        """Create and register new trials."""
        sampled_points = 0

        # reset the number of time we failed to sample points
        self.failure_count = 0
        start = time.time()

        while (sampled_points < self.pool_size and
               not (self.experiment.is_done or self.naive_algorithm.is_done)):
            self._sample_guard(start)

            log.debug("### Algorithm suggests new points.")
            new_points = self.naive_algorithm.suggest(self.pool_size)

            # Sync state of original algo so that state continues evolving.
            self.algorithm.set_state(self.naive_algorithm.state_dict)

            if new_points is None:
                if self.algorithm.is_done:
                    return

                raise WaitingForTrials(
                    'Algo does not have more trials to sample.'
                    'Waiting for current trials to finish')

            for new_point in new_points:
                sampled_points += self.register_trials(new_point)

    def register_trials(self, new_point):
        """Register a new set of sampled parameters into the DB
        guaranteeing their uniqueness

        Parameters
        ----------
        new_point: tuple
            tuple of values representing the hyperparameters values

        """
        # FIXME: Relying on DB to guarantee uniqueness
        # when the trial history will be held by that algo we can move that logic out of the DB

        log.debug("#### Convert point to `Trial` object.")
        new_trial = format_trials.tuple_to_trial(new_point, self.space)

        try:
            self._prevalidate_trial(new_trial)
            new_trial.parents = self.naive_trials_history.children
            log.debug("#### Register new trial to database: %s", new_trial)
            self.experiment.register_trial(new_trial)
            self._update_params_hashes([new_trial])
            return 1

        except DuplicateKeyError:
            log.debug("#### Duplicate sample.")
            self.backoff()
            return 0

    def _prevalidate_trial(self, new_trial):
        """Verify if trial is not in parent history"""
        if Trial.compute_trial_hash(
                new_trial, ignore_experiment=True) in self.params_hashes:
            raise DuplicateKeyError

    def _update_params_hashes(self, trials):
        """Register locally all param hashes of trials"""
        for trial in trials:
            self.params_hashes.add(
                Trial.compute_trial_hash(trial,
                                         ignore_experiment=True,
                                         ignore_lie=True))

    def update(self):
        """Pull all trials to update model with completed ones and naive model with non completed
        ones.
        """
        trials = self.experiment.fetch_trials(with_evc_tree=True)

        self._update_algorithm(
            [trial for trial in trials if trial.status == 'completed'])
        self._update_naive_algorithm(
            [trial for trial in trials if trial.status != 'completed'])

    def _update_algorithm(self, completed_trials):
        """Pull newest completed trials to update local model."""
        log.debug("### Fetch completed trials to observe:")

        new_completed_trials = []
        for trial in completed_trials:
            if trial not in self.trials_history:
                new_completed_trials.append(trial)

        log.debug("### %s", new_completed_trials)

        if new_completed_trials:
            log.debug("### Convert them to list of points and their results.")
            points = list(
                map(
                    lambda trial: format_trials.trial_to_tuple(
                        trial, self.space), new_completed_trials))
            results = list(
                map(format_trials.get_trial_results, new_completed_trials))

            log.debug("### Observe them.")
            self.trials_history.update(new_completed_trials)
            self.algorithm.observe(points, results)
            self.strategy.observe(points, results)
            self._update_params_hashes(new_completed_trials)

    def _produce_lies(self, incomplete_trials):
        """Add fake objective results to incomplete trials

        Then register the trials in the db
        """
        log.debug("### Fetch active trials to observe:")
        lying_trials = []
        log.debug("### %s", incomplete_trials)

        for trial in incomplete_trials:
            log.debug(
                "### Use defined ParallelStrategy to assign them fake results."
            )
            lying_result = self.strategy.lie(trial)
            if lying_result is not None:
                lying_trial = copy.deepcopy(trial)
                lying_trial.results.append(lying_result)
                lying_trials.append(lying_trial)
                log.debug("### Register lie to database: %s", lying_trial)
                lying_trial.parents = self.trials_history.children
                try:
                    self.experiment.register_lie(lying_trial)
                except DuplicateKeyError:
                    log.debug(
                        "#### Duplicate lie. No need to register a duplicate in DB."
                    )

        return lying_trials

    def _update_naive_algorithm(self, incomplete_trials):
        """Pull all non completed trials to update naive model."""
        self.naive_algorithm = copy.deepcopy(self.algorithm)
        self.naive_trials_history = copy.deepcopy(self.trials_history)
        log.debug("### Create fake trials to observe:")
        lying_trials = self._produce_lies(incomplete_trials)
        log.debug("### %s", lying_trials)
        if lying_trials:
            log.debug("### Convert them to list of points and their results.")
            points = list(
                map(
                    lambda trial: format_trials.trial_to_tuple(
                        trial, self.space), lying_trials))
            results = list(map(format_trials.get_trial_results, lying_trials))

            log.debug("### Observe them.")
            self.naive_trials_history.update(lying_trials)
            self.naive_algorithm.observe(points, results)
            self._update_params_hashes(lying_trials)
Example #11
0
class Producer(object):
    """Produce suggested sets of problem's parameter space to try out.

    It uses an `Experiment` object to poll for not yet observed trials which
    have been already evaluated and to register new suggestions (points of
    the parameter `Space`) to be evaluated.

    """
    def __init__(self, experiment, max_idle_time=None):
        """Initialize a producer.

        :param experiment: Manager of this experiment, provides convenient
           interface for interacting with the database.
        """
        log.debug("Creating Producer object.")
        self.experiment = experiment
        self.space = experiment.space
        if self.space is None:
            raise RuntimeError(
                "Experiment object provided to Producer has not yet completed"
                " initialization.")
        self.algorithm = experiment.algorithms
        if max_idle_time is None:
            max_idle_time = orion.core.config.worker.max_idle_time
        self.max_idle_time = max_idle_time
        self.strategy = experiment.producer['strategy']
        self.naive_algorithm = None
        # TODO: Move trials_history into PrimaryAlgo during the refactoring of Algorithm with
        #       Strategist and Scheduler.
        self.trials_history = TrialsHistory()
        self.naive_trials_history = None

    @property
    def pool_size(self):
        """Pool-size of the experiment"""
        return self.experiment.pool_size

    def backoff(self):
        """Wait some time and update algorithm."""
        waiting_time = min(0, random.gauss(1, 0.2))
        log.info('Waiting %d seconds', waiting_time)
        time.sleep(waiting_time)
        log.info('Updating algorithm.')
        self.update()

    def produce(self):
        """Create and register new trials."""
        sampled_points = 0

        start = time.time()
        while sampled_points < self.pool_size and not self.algorithm.is_done:
            if time.time() - start > self.max_idle_time:
                raise RuntimeError(
                    "Algorithm could not sample new points in less than {} seconds"
                    .format(self.max_idle_time))

            log.debug("### Algorithm suggests new points.")

            new_points = self.naive_algorithm.suggest(self.pool_size)
            # Sync state of original algo so that state continues evolving.
            self.algorithm.set_state(self.naive_algorithm.state_dict)
            if new_points is None:
                log.info("### Algo opted out.")
                self.backoff()
                continue

            for new_point in new_points:
                log.debug("#### Convert point to `Trial` object.")
                new_trial = format_trials.tuple_to_trial(new_point, self.space)
                try:
                    new_trial.parents = self.naive_trials_history.children
                    log.debug("#### Register new trial to database: %s",
                              new_trial)
                    self.experiment.register_trial(new_trial)
                    sampled_points += 1
                except DuplicateKeyError:
                    log.debug("#### Duplicate sample.")
                    self.backoff()
                    break

    def update(self):
        """Pull all trials to update model with completed ones and naive model with non completed
        ones.
        """
        trials = self.experiment.fetch_trials()

        self._update_algorithm(
            [trial for trial in trials if trial.status == 'completed'])
        self._update_naive_algorithm(
            [trial for trial in trials if trial.status != 'completed'])

    def _update_algorithm(self, completed_trials):
        """Pull newest completed trials to update local model."""
        log.debug("### Fetch completed trials to observe:")

        new_completed_trials = []
        for trial in completed_trials:
            if trial not in self.trials_history:
                new_completed_trials.append(trial)

        log.debug("### %s", new_completed_trials)

        if new_completed_trials:
            log.debug("### Convert them to list of points and their results.")
            points = list(
                map(
                    lambda trial: format_trials.trial_to_tuple(
                        trial, self.space), new_completed_trials))
            results = list(
                map(format_trials.get_trial_results, new_completed_trials))

            log.debug("### Observe them.")
            self.trials_history.update(new_completed_trials)
            self.algorithm.observe(points, results)
            self.strategy.observe(points, results)

    def _produce_lies(self, incomplete_trials):
        """Add fake objective results to incomplete trials

        Then register the trials in the db
        """
        log.debug("### Fetch active trials to observe:")
        lying_trials = []
        log.debug("### %s", incomplete_trials)

        for trial in incomplete_trials:
            log.debug(
                "### Use defined ParallelStrategy to assign them fake results."
            )
            lying_result = self.strategy.lie(trial)
            if lying_result is not None:
                lying_trial = copy.deepcopy(trial)
                lying_trial.results.append(lying_result)
                lying_trials.append(lying_trial)
                log.debug("### Register lie to database: %s", lying_trial)
                lying_trial.parents = self.trials_history.children
                try:
                    self.experiment.register_lie(lying_trial)
                except DuplicateKeyError:
                    log.debug(
                        "#### Duplicate lie. No need to register a duplicate in DB."
                    )

        return lying_trials

    def _update_naive_algorithm(self, incomplete_trials):
        """Pull all non completed trials to update naive model."""
        self.naive_algorithm = copy.deepcopy(self.algorithm)
        self.naive_trials_history = copy.deepcopy(self.trials_history)
        log.debug("### Create fake trials to observe:")
        lying_trials = self._produce_lies(incomplete_trials)
        log.debug("### %s", lying_trials)
        if lying_trials:
            log.debug("### Convert them to list of points and their results.")
            points = list(
                map(
                    lambda trial: format_trials.trial_to_tuple(
                        trial, self.space), lying_trials))
            results = list(map(format_trials.get_trial_results, lying_trials))

            log.debug("### Observe them.")
            self.naive_trials_history.update(lying_trials)
            self.naive_algorithm.observe(points, results)