# Copyright 2018 Amazon.com, Inc. or its affiliates. All Rights Reserved.
#
# Licensed under the Apache License, Version 2.0 (the "License").
# You may not use this file except in compliance with the License.
# A copy of the License is located at
#
# http://www.apache.org/licenses/LICENSE-2.0
#
# or in the "license" file accompanying this file. This file is distributed
# on an "AS IS" BASIS, WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either
# express or implied. See the License for the specific language governing
# permissions and limitations under the License.
import logging
from functools import singledispatch
from typing import Callable, Iterator, List, Optional
import numpy as np
from gluonts.core.component import validated
from gluonts.dataset.common import DataEntry
from gluonts.dataset.field_names import FieldName
from gluonts.dataset.loader import DataLoader
from gluonts.model.forecast import Forecast, QuantileForecast, SampleForecast
logger = logging.getLogger(__name__)
OutputTransform = Callable[[DataEntry, np.ndarray], np.ndarray]
LOG_CACHE = set()
OUTPUT_TRANSFORM_NOT_SUPPORTED_MSG = (
"The `output_transform` argument is not supported and will be ignored."
)
NOT_SAMPLE_BASED_MSG = (
"Forecast is not sample based. Ignoring parameter `num_samples` from"
" predict method."
)
[docs]def log_once(msg):
global LOG_CACHE
if msg not in LOG_CACHE:
logger.info(msg)
LOG_CACHE.add(msg)
# different deep learning frameworks generate predictions and the tensor to
# numpy conversion differently, use a dispatching function to prevent needing
# a ForecastGenerators for each framework
[docs]@singledispatch
def predict_to_numpy(prediction_net, args) -> np.ndarray:
raise NotImplementedError
def _unpack(batched) -> Iterator:
"""
Unpack batches.
This assumes that arrays are wrapped in a nested structure of lists and
tuples, and each array has the same shape::
>>> a = np.arange(5)
>>> batched = [a, (a, [a, a, a])]
>>> list(_unpack(batched))
[[0, (0, [0, 0, 0])],
[1, (1, [1, 1, 1])],
[2, (2, [2, 2, 2])],
[3, (3, [3, 3, 3])],
[4, (4, [4, 4, 4])]]
"""
if isinstance(batched, (list, tuple)):
T = type(batched)
return map(T, zip(*map(_unpack, batched)))
return batched
[docs]@singledispatch
def make_distribution_forecast(distr, *args, **kwargs) -> Forecast:
raise NotImplementedError
[docs]class ForecastGenerator:
"""
Classes used to bring the output of a network into a class.
"""
def __call__(
self,
inference_data_loader: DataLoader,
prediction_net,
input_names: List[str],
output_transform: Optional[OutputTransform],
num_samples: Optional[int],
**kwargs
) -> Iterator[Forecast]:
raise NotImplementedError()
[docs]class QuantileForecastGenerator(ForecastGenerator):
@validated()
def __init__(self, quantiles: List[str]) -> None:
self.quantiles = quantiles
def __call__(
self,
inference_data_loader: DataLoader,
prediction_net,
input_names: List[str],
output_transform: Optional[OutputTransform],
num_samples: Optional[int],
**kwargs
) -> Iterator[Forecast]:
for batch in inference_data_loader:
inputs = [batch[k] for k in input_names]
outputs = predict_to_numpy(prediction_net, inputs)
if output_transform is not None:
outputs = output_transform(batch, outputs)
if num_samples:
log_once(NOT_SAMPLE_BASED_MSG)
i = -1
for i, output in enumerate(outputs):
yield QuantileForecast(
output,
start_date=batch[FieldName.FORECAST_START][i],
item_id=batch[FieldName.ITEM_ID][i]
if FieldName.ITEM_ID in batch
else None,
info=batch["info"][i] if "info" in batch else None,
forecast_keys=self.quantiles,
)
assert i + 1 == len(batch[FieldName.FORECAST_START])
[docs]class SampleForecastGenerator(ForecastGenerator):
@validated()
def __init__(self):
pass
def __call__(
self,
inference_data_loader: DataLoader,
prediction_net,
input_names: List[str],
output_transform: Optional[OutputTransform],
num_samples: Optional[int],
**kwargs
) -> Iterator[Forecast]:
for batch in inference_data_loader:
inputs = [batch[k] for k in input_names]
outputs = predict_to_numpy(prediction_net, inputs)
if output_transform is not None:
outputs = output_transform(batch, outputs)
if num_samples:
num_collected_samples = outputs[0].shape[0]
collected_samples = [outputs]
while num_collected_samples < num_samples:
outputs = predict_to_numpy(prediction_net, inputs)
if output_transform is not None:
outputs = output_transform(batch, outputs)
collected_samples.append(outputs)
num_collected_samples += outputs[0].shape[0]
outputs = [
np.concatenate(s)[:num_samples]
for s in zip(*collected_samples)
]
assert len(outputs[0]) == num_samples
i = -1
for i, output in enumerate(outputs):
yield SampleForecast(
output,
start_date=batch[FieldName.FORECAST_START][i],
item_id=batch[FieldName.ITEM_ID][i]
if FieldName.ITEM_ID in batch
else None,
info=batch["info"][i] if "info" in batch else None,
)
assert i + 1 == len(batch[FieldName.FORECAST_START])
[docs]class DistributionForecastGenerator(ForecastGenerator):
@validated()
def __init__(self, distr_output) -> None:
self.distr_output = distr_output
def __call__(
self,
inference_data_loader: DataLoader,
prediction_net,
input_names: List[str],
output_transform: Optional[OutputTransform],
num_samples: Optional[int],
**kwargs
) -> Iterator[Forecast]:
for batch in inference_data_loader:
inputs = [batch[k] for k in input_names]
outputs = prediction_net(*inputs)
if output_transform:
log_once(OUTPUT_TRANSFORM_NOT_SUPPORTED_MSG)
if num_samples:
log_once(NOT_SAMPLE_BASED_MSG)
distributions = [
self.distr_output.distribution(*u) for u in _unpack(outputs)
]
i = -1
for i, distr in enumerate(distributions):
yield make_distribution_forecast(
distr,
start_date=batch[FieldName.FORECAST_START][i],
item_id=batch[FieldName.ITEM_ID][i]
if FieldName.ITEM_ID in batch
else None,
info=batch["info"][i] if "info" in batch else None,
)
assert i + 1 == len(batch[FieldName.FORECAST_START])