# !/usr/bin/env python3
# -*- coding:utf-8 -*-
from abc import ABC, abstractmethod
from typing import Any, Callable, List, Optional, Sequence, Tuple, Union, Dict
import numpy as np
import pandas as pd
import matplotlib.pyplot
from paddlets import TimeSeries, TSDataset
from paddlets.logger import Logger, raise_if_not, raise_if, raise_log
logger = Logger(__name__)
[docs]class Analyzer(ABC):
"""
``Analyzer`` is the base class for all analyzers.
Analyzer module is designed to perform specific mathematical analysis on time series data.
All analyzer need to override analyze method.
Args:
kwargs: Argument positions left for sub-classes.
"""
def __init__(self, **kwargs):
self._res = None
self._kwargs = kwargs
def _build_analysis_data(
self,
tsdataset: "TSDataset",
columns: Optional[Union[str, List[str]]] = None,
**kwargs
) -> Union[pd.Series, pd.DataFrame]:
"""
Build a pd.Series or pd.DataFrame From a TSDataset or specific columns in a TSDataset.
Args:
tsdataset(TSDataset): TSDataset to be analyzed.
columns(str|list[str]): Specific columns used to build pd.Series or pd.DataFrame.
kwargs: argument positions Left for overide methods.
Returns:
pd.Series|pd.DataFrame: pd.Series or pd.DataFrame that convert from the TSDataset.
Raise:
ValueError
"""
if columns is None:
if tsdataset.get_target() is None and tsdataset.get_all_cov() is None:
raise_log(
ValueError('tsdataset is empty!')
)
elif tsdataset.get_all_cov() is None:
return tsdataset.get_target().data
elif tsdataset.get_target() is None:
return tsdataset.get_all_cov().data
else:
return pd.concat([tsdataset.get_target().data, tsdataset.get_all_cov().data], axis=1)
else:
return tsdataset[columns]
[docs] @abstractmethod
def analyze(
self,
X: Union[pd.Series, pd.DataFrame],
**kwargs
) -> Any:
"""
Analyze data, need to be implemented by sub-classes.
Args:
X(pd.Series|pd.DataFrame): Pd.Series or pd.DataFrame to be analyzed.
kwargs: Argument positions left for override methods.
Returns:
Any: Analysis results.
Raise:
ValueError
"""
pass
[docs] @abstractmethod
def get_properties(self) -> Dict:
"""
Get the properties of the analyzer.
All sub-classses should implements.
Args:
None
Returns:
Dict
Raise:
None
"""
pass
[docs] def plot(self) -> "pyplot":
"""
The plot method of the Analyzer to show figures of Analysis results, optionally override this method.
If analyzers need to displays figures in the analysis report, this method needs to be overrided.
Args:
None
Returns:
matplotlib.pyplot object
Raise:
None
Examples:
.. code-block:: python
# Analysis results is DataFrame
def plot(self):
fig = self.res.plot().get_figure()
return fig
# Other
#Implement the plot method by yourself, return need to be matplotlib.pyplot object.
"""
return None
def __call__(
self,
tsdataset: "TSDataset",
columns: Optional[Union[str, List[str]]] = None
)-> Any:
"""
Compute analysis's value from TSDataset
Args:
tsdataset(TSDataset): TSDataset to be analyzed.
columns(str|List[str]): Specific columns to be analyzed. Analyze complete TSDataset by default.
Returns:
Any: Analysis results.
Raise:
ValueError
"""
array = self._build_analysis_data(tsdataset, columns, **self._kwargs)
self._res = self.analyze(array, **self._kwargs)
return self._res