"""Module for validating data and raising informative errors."""
from typing import Sequence, Tuple, Union
import numpy as np
import pandas as pd
def check_1d(seq: Sequence) -> Sequence:
"""Check given seq is one-dimensional. Raise error if can't be easily transformed."""
e = ValueError("Too many dimensions.")
for ndx, el in enumerate(seq):
if isinstance(el, (list, tuple)):
if len(el) > 1:
raise e
seq[ndx] = el[0]
elif isinstance(el, (np.ndarray, pd.Series)):
if el.size > 1:
raise e
if isinstance(el, np.ndarray):
seq[ndx] = el.reshape(1, -1)[0][0]
else:
seq[ndx] = el[0]
return seq
def check_2d(seq: Union[list, np.ndarray]) -> list:
"""Check given seq is two-dimensional. Raise error if can't be easily transformed."""
for ndx, el in enumerate(seq):
if not isinstance(el, (list, tuple, np.ndarray, pd.Series)):
raise ValueError("Too few dimensions.")
seq[ndx] = check_1d(el)
return seq
def check_column_types(seq: Sequence, certainty_proportion: float = 0.2) -> Sequence[type]:
"""Check all elements of each column are of the same type."""
output_types = None
for ndx in range(min(int(len(seq) * certainty_proportion) + 1, len(seq) - 1)):
el = seq[ndx]
el_types = [type(val) for val in el]
if output_types is None:
output_types = el_types
else:
if len(output_types) != len(el_types):
raise ValueError("Found inconsistent number of columns.")
for i, output_type in enumerate(output_types):
if output_type != el_types[i]:
raise ValueError("Found inconsistent types in column.")
return output_types
def check_num_columns(seq: Sequence, certainty_proportion: float = 0.2) -> int:
"""Return the number of columns of in the dataset."""
num_columns = None
for ndx in range(min(int(len(seq) * certainty_proportion) + 1, len(seq) - 1)):
el_len = len(list(seq[ndx]))
if num_columns is None:
num_columns = el_len
else:
if num_columns != el_len:
raise ValueError("Found inconsistent number of columns.")
return num_columns
def check_X_y(X, y) -> Tuple[Sequence, Sequence, int, Sequence]:
"""Check the given X and y datasets are prepared to be passed to a PushEstimator.
Inspired by the scikit-learn function with the same name.
https://scikit-learn.org/stable/modules/generated/sklearn.utils.check_X_y.html#sklearn.utils.check_X_y
"""
X = check_2d(X)
y = check_2d(y)
X_column_count = check_num_columns(X)
y_dim_types = check_column_types(y)
return X, y, X_column_count, y_dim_types
def check_is_fitted(estimator, attribute):
"""Check the given estimator has already been fit.
Inspired by the scikit-learn function with the same name.
https://scikit-learn.org/stable/modules/generated/sklearn.utils.validation.check_is_fitted.html#sklearn.utils.validation.check_is_fitted
"""
if not hasattr(estimator, 'fit'):
raise TypeError("{e} is not an estimator.".format(e=estimator))
if not hasattr(estimator, attribute):
raise ValueError(
"Estimator has not been fit yet. Call 'fit' with appropriate arguments before using this method."
)
class PushError(Exception):
"""Error raised during Push program execution."""
@classmethod
def no_type(cls, thing):
"""Raise PushError when no PushType can be found for something."""
return cls("Unkown PushType for {th}.".format(th=thing))
@classmethod
def failed_coerce(cls, thing, push_type):
"""Raise PushError when something fails to coerce to a PushType."""
return cls("Could not convert {typ1} {th} to {typ2}.".format(
th=thing,
typ1=type(thing),
typ2=push_type
))
@classmethod
def empty_character(cls):
"""Raise PushError when Character is made from empty string."""
return cls("Char object cannot be created from empty string.")
@classmethod
def long_character(cls):
"""Raise PushError when Character is made from string length more than 1."""
return cls("Char object cannot be created from string of length > 1.")
