def get(self):
"""
.. py:method:: get(self)
Computes the correlation matrix of the selected columns
of the input dataframe to the object's constructor.
**Parameters**
None
**Returns**
None. However, this method causes side-effects by manipulating
the existing attributes of the object.
"""
if hasattr(self, "method"):
self._isCorMat = True
self._matrixType = "correlation"
if self.method not in ["pearson", "kendall", "spearman"]:
raise Exception(
"The requested correlation type must be one of the following string values,\n"
+ " pearson : standard correlation coefficient\n" +
" kendall : Kendall Tau correlation coefficient\n" +
" spearman : Spearman rank correlation.")
else:
self._isCorMat = False
self._matrixType = "covariance"
# check columns presence
if self.columns is None:
colnames = self._dfref().columns
colindex = range(len(colnames))
elif all(isinstance(element, str) for element in self.columns):
colnames = self.columns
colindex = dfutils.nam2num(self._dfref().columns, self.columns)
elif all(isinstance(element, int) for element in self.columns):
colindex = self.columns
colnames = self._dfref().columns[colindex]
else:
raise Exception(
"The input argument 'columns' must be a list whose elements are all\n"
+
" 1. string-valued, each representing the name of the column from\n"
+
" the input dataframe to the object's constructor, to be included\n"
+ " in the " + self._matrixType +
" matrix construction, or,\n" +
" 2. integer-valued, each representing the index of the column from\n"
+
" the input dataframe to the object's constructor, to be included\n"
+ " in the " + self._matrixType +
" matrix construction.")
# check rows presence
if self.rows is None:
rowindex = range(len(self._dfref().index))
else:
rowindex = self.rows
# construct the matrix dataframe
if self._isCorMat:
self.df = self._dfref().iloc[rowindex,
colindex].corr(method=self.method)
else:
self.df = self._dfref().iloc[rowindex, colindex].cov()
# specify columns/index names
self.df.columns = colnames
self.df.index = colnames
#!DEC$ ifdef PMVIS_ENABLED
# add heatmap plot
heatmap_kws = {}
if self._isCorMat:
annotPrecision = 2
heatmap_kws["cbar_kws"] = {
"label": self.method.capitalize() + "'s Correlation Strength",
"orientation": "vertical",
"ticks": _np.linspace(-1, 1, 9)
}
heatmap_kws["vmin"] = -1
heatmap_kws["vmax"] = +1
heatmap_kws["center"] = 0
else:
annotPrecision = None
heatmap_kws["cbar_kws"] = {
"label": "Covariance Strength",
"orientation": "vertical"
}
from _HeatMapPlot import HeatMapPlot
self.plot = _Struct()
self.plot.heatmap = HeatMapPlot(dataFrame=self.df,
heatmap_kws=heatmap_kws,
annotPrecision=annotPrecision)
def get(self, reself: tp.Optional[bool] = False, **kwargs):
"""
Compute the correlation / covariance matrix of the selected
columns of the input dataframe to the object's constructor.
**Parameters**
reself
A logical variable. If ``True``, an instance of
the object will be returned to the calling routine
upon exit. The default value is ``False``.
**Returns**
The object self if ``reself = True`` otherwise, ``None``.
**NOTE**
This method causes side-effects by manipulating
the existing attributes of the object.
"""
for key in kwargs.keys():
if hasattr(self, key):
setattr(self, key, kwargs[key])
elif key == "dataFrame":
setattr(self, "_dfref", wref.ref(kwargs[key]))
else:
raise Exception("Unrecognized input '" + key +
"' class attribute detected." + newline +
self._getDocString())
if hasattr(self, "method"):
self._isCorMat = True
self._matrixType = "correlation"
if self.method not in ["pearson", "kendall", "spearman"]:
raise Exception(
newline +
"The requested correlation type must be one of the following string values,\n"
+ " pearson : standard correlation coefficient\n" +
" kendall : Kendall Tau correlation coefficient\n" +
" spearman : Spearman rank correlation.")
else:
self._isCorMat = False
self._matrixType = "covariance"
############################################################################################################################
#### check columns presence
############################################################################################################################
if self.columns is None:
colnames = self._dfref().columns
colindex = range(len(colnames))
elif all(isinstance(element, str) for element in self.columns):
colnames = self.columns
colindex = dfutils.nam2num(self._dfref().columns, self.columns)
elif all(isinstance(element, int) for element in self.columns):
colindex = self.columns
colnames = self._dfref().columns[colindex]
else:
raise Exception(
newline +
"The input argument 'columns' must be a list whose elements are all\n"
+
" 1. string-valued, each representing the name of the column from\n"
+
" the input dataframe to the object's constructor, to be included\n"
+ " in the " + self._matrixType +
" matrix construction, or,\n" +
" 2. integer-valued, each representing the index of the column from\n"
+
" the input dataframe to the object's constructor, to be included\n"
+ " in the " + self._matrixType +
" matrix construction.")
############################################################################################################################
#### check rows presence. This must be checked here, because it depends on the integrity of the in input dataFrame.
############################################################################################################################
if self.rows is None: self.rows = range(len(self._dfref().index))
############################################################################################################################
#### construct the matrix dataframe
############################################################################################################################
if self._isCorMat:
self.df = self._dfref().iloc[self.rows,
colindex].corr(method=self.method)
else:
self.df = self._dfref().iloc[self.rows, colindex].cov()
############################################################################################################################
#### specify columns/index names
############################################################################################################################
self.df.columns = colnames
self.df.index = colnames
############################################################################################################################
#### graphics
############################################################################################################################
self._plotTypeList = ["heatmap"]
self._progress.note(msg="adding the " + self._matrixType +
" graphics tools... ",
end=newline,
pre=True)
self.plot = Struct()
self._resetPlot(resetType="hard")
self.plot.reset = self._resetPlot
############################################################################################################################
if reself: return self
def get(self):
"""
.. py:method:: get(self)
Computes the autocorrelations of the selected columns
of the input dataframe to the object's constructor.
**Parameters**
None
**Returns**
None. However, this method causes side-effects by manipulating
the existing attributes of the object.
"""
# check columns presence
if self.columns is None:
colnames = self._dfref().columns
colindex = range(len(colnames))
elif all(isinstance(element, str) for element in self.columns):
colnames = self.columns
colindex = dfutils.nam2num( self._dfref().columns , self.columns )
elif all(isinstance(element,int) for element in self.columns):
colindex = self.columns
colnames = self._dfref().columns[colindex]
else:
raise Exception ( "The input argument 'columns' must be a list whose elements are all\n"
+ " 1. string-valued, each representing the name of the column from\n"
+ " the input dataframe to the object's constructor, to be included\n"
+ " in the computation of autocorrelations, or,\n"
+ " 2. integer-valued, each representing the index of the column from\n"
+ " the input dataframe to the object's constructor, to be included\n"
+ " in the computation of autocorrelations."
)
# check rows presence
if self.rows is None:
rowindex = range(len(self._dfref().index))
else:
rowindex = self.rows
# compute the autocorrelations
nvar = len(colnames)
nlag = len(rowindex)
acf = _np.zeros((nvar,nlag))
from scipy.signal import correlate as _ccor
for icnt, ivar in enumerate(colindex):
xdata = self._dfref().iloc[rowindex,ivar].values.flatten() - _np.mean(self._dfref().iloc[rowindex,ivar].values.flatten())
acf[icnt] = _ccor ( xdata
, xdata
, mode = "full"
)[nlag-1:2*nlag]
acf[icnt] = acf[icnt] / acf[icnt,0]
self.df = _pd.DataFrame(_np.transpose(acf))
# specify columns/index names
colnames = [ "ACF_"+colnames[i] for i in range(len(colnames)) ]
self.df.columns = colnames
# add SampleLogFunc to df for plot coloring, if exists
ccolumns = ()
if "SampleLogFunc" in self._dfref().columns:
ccolumns = "SampleLogFunc"
self.df.insert ( loc = 0
, column = ccolumns
, value = self._dfref()[[ccolumns]].values.flatten()
, allow_duplicates = True
)
# add lags to df
self.df.insert ( loc = 0
, column = "Lag"
, value = [ i for i in self.df.index ]
)
#!DEC$ ifdef PMVIS_ENABLED
#####################
#### add plots
#####################
self.plot = _Struct()
# add LinePlot
from _LinePlot import LinePlot
self.plot.line = LinePlot ( dataFrame = self.df
, xcolumns = "Lag"
, ycolumns = colnames
, ccolumns = None #ccolumns
, lc_kws = {
#"linewidth":0.75,
#"cmap":"viridis",
"cmap":"autumn",
#"alpha":0.5,
}
, colorbar_kws = {
"extend":"neither",
"orientation":"vertical",
#"spacing":"uniform",
}
#, legend_kws = None
)
# add ScatterPlot
from _ScatterPlot import ScatterPlot
self.plot.scatter = ScatterPlot ( dataFrame = self.df
, xcolumns = "Lag"
, ycolumns = colnames
, ccolumns = None #ccolumns
#, scatter_kws = {}
, colorbar_kws = {
"extend":"neither",
"orientation":"vertical",
#"spacing":"uniform",
}
#, legend_kws = None
)
################################################################################################################################
#!DEC$ endif