# independent_components.py
# Contact: Jacob Schreiber <jmschreiber91@gmail.com>

import torch

from .._utils import _cast_as_tensor
from .._utils import _cast_as_parameter
from .._utils import _update_parameter
from .._utils import _check_parameter
from .._utils import _reshape_weights

from ._distribution import Distribution


class IndependentComponents(Distribution):
	"""An independent components distribution object.

	A distribution made up of independent, univariate, distributions that each
	model a single feature in the data. This means that instead of using a
	single type of distribution to model all of the features in your data, you
	use one distribution per feature. Note that this will likely be slower
	than using a single distribution because the amount of batching possible
	will go down significantly.

	There are two ways to initialize this object. The first is to pass in a
	set of distributions that are all initialized with parameters, at which
	point this distribution can be immediately used for inference. The second
	is to pass in a set of distributions that are not initialized with
	parameters, and then call either `fit` or `summary` + `from_summaries` to
	learn the parameters of all the distributions.


	Parameters
	----------
	distributions: list, tuple, numpy.ndarray, torch.Tensor, shape=(d,)
		An ordered iterable containing all of the distributions, one per
		feature, that will be used.

	check_data: bool, optional
		Whether to check properties of the data and potentially recast it to
		torch.tensors. This does not prevent checking of parameters but can
		slightly speed up computation when you know that your inputs are valid.
		Setting this to False is also necessary for compiling.
	"""

	def __init__(self, distributions, check_data=False):
		super().__init__(inertia=0.0, frozen=False, check_data=check_data)
		self.name = "IndependentComponents"

		if len(distributions) <= 1:
			raise ValueError("Must pass in at least 2 distributions.")
		for distribution in distributions:
			if not isinstance(distribution, Distribution):
				raise ValueError("All passed in distributions must " +
					"inherit from the Distribution object.")

		self.distributions = distributions
		self._initialized = all(d._initialized for d in distributions)
		self.d = len(distributions)
		self._reset_cache()


	def _initialize(self, d):
		"""Initialize the probability distribution.

		This method is meant to only be called internally. It initializes the
		parameters of the distribution and stores its dimensionality. For more
		complex methods, this function will do more.


		Parameters
		----------
		d: int
			The dimensionality the distribution is being initialized to.
		"""

		for distribution in self.distributions:
			distribution._initialize(d)

		self._initialized = True


	def _reset_cache(self):
		"""Reset the internally stored statistics.

		This method is meant to only be called internally. It resets the
		stored statistics used to update the model parameters as well as
		recalculates the cached values meant to speed up log probability
		calculations.
		"""

		if self._initialized == False:
			return

		for distribution in self.distributions:
			distribution._reset_cache()


	def sample(self, n):
		"""Sample from the probability distribution.

		This method will return `n` samples generated from the underlying
		probability distribution.


		Parameters
		----------
		n: int
			The number of samples to generate.
		

		Returns
		-------
		X: torch.tensor, shape=(n, self.d)
			Randomly generated samples.
		"""

		return torch.hstack([d.sample(n) for d in self.distributions])


	def log_probability(self, X):
		"""Calculate the log probability of each example.

		This method calculates the log probability of each example given the
		parameters of the distribution. The examples must be given in a 2D
		format.

		Note: This differs from some other log probability calculation
		functions, like those in torch.distributions, because it is not
		returning the log probability of each feature independently, but rather
		the total log probability of the entire example.


		Parameters
		----------
		X: list, tuple, numpy.ndarray, torch.Tensor, shape=(-1, self.d)
			A set of examples to evaluate.


		Returns
		-------
		logp: torch.Tensor, shape=(-1,)
			The log probability of each example.
		"""

		X = _check_parameter(_cast_as_tensor(X), "X", ndim=2, 
			shape=(-1, self.d))

		logp = torch.zeros(X.shape[0])
		for i, d in enumerate(self.distributions):
			if isinstance(X, torch.masked.MaskedTensor):
				logp.add_(d.log_probability(X[:, i:i+1])._masked_data)
			else:
				logp.add_(d.log_probability(X[:, i:i+1]))

		return logp


	def summarize(self, X, sample_weight=None):
		"""Extract the sufficient statistics from a batch of data.

		This method calculates the sufficient statistics from optionally
		weighted data and adds them to the stored cache. The examples must be
		given in a 2D format. Sample weights can either be provided as one
		value per example or as a 2D matrix of weights for each feature in
		each example.


		Parameters
		----------
		X: list, tuple, numpy.ndarray, torch.Tensor, shape=(-1, self.d)
			A set of examples to summarize.

		sample_weight: list, tuple, numpy.ndarray, torch.Tensor, optional
			A set of weights for the examples. This can be either of shape
			(-1, self.d) or a vector of shape (-1,). Default is ones.
		"""

		if self.frozen == True:
			return
			
		X = _check_parameter(_cast_as_tensor(X), "X", ndim=2, 
			shape=(-1, self.d))

		sample_weight = _reshape_weights(X, _cast_as_tensor(sample_weight, 
			dtype=torch.float32), device=self.device)

		for i, d in enumerate(self.distributions):
			d.summarize(X[:, i:i+1], sample_weight=sample_weight[:, i:i+1])


	def from_summaries(self):
		"""Update the model parameters given the extracted statistics.

		This method uses calculated statistics from calls to the `summarize`
		method to update the distribution parameters. Hyperparameters for the
		update are passed in at initialization time.

		Note: Internally, a call to `fit` is just a successive call to the
		`summarize` method followed by the `from_summaries` method.
		"""
		
		if self.frozen == True:
			return

		for distribution in self.distributions:
			distribution.from_summaries()