def heatmap_cross_correlation(df_dev, figsize=(10, 8)):
title = 'Devices cross-correlation'
cmap = get_diverging_color()
cbarlabel = 'similarity'
ct = cross_correlation(df_dev)
vals = ct.values.T.astype(np.float64)
devs = list(ct.index)
fig, ax = plt.subplots(figsize=figsize)
im, cbar = heatmap_square(vals,
devs,
devs,
ax=ax,
cmap=cmap,
cbarlabel=cbarlabel,
vmin=-1,
vmax=1)
annotate_heatmap(im,
textcolors=("black", "white"),
valfmt="{x:.2f}",
threshold=0.6)
ax.set_title(title)
fig.tight_layout()
plt.show()
def corr_devices_01(rep, figsize=(19,14)):
""" correlation between the on and off states of every device.
Parameters
----------
rep: pd.DataFrame
A dataframe where columns are devices and rows are a binary representation
e.g raw, changepoint, lastfired representation
"""
df = rep.iloc[:,:-1] # get raw without activities
df = df.reset_index(drop=True)
df = df.astype(int)
for device in df.columns:
mask1 = df[[device]] == 1
col_off = device + ' Off'
df[[col_off]] = df[[device]].astype(bool)
df[[col_off]] = ~df[[col_off]]
df[[col_off]] = df[[col_off]].astype(int)
df = df.rename(columns={device: device + ' On'})
dev_count = int(len(df.columns)/2)
ct = df.corr()
ct_on_on = ct.iloc[:dev_count, :dev_count]
ct_on_on_values = ct_on_on.values.T
ct_on_on_devs = list(ct_on_on.index)
ct_on_off = ct.iloc[dev_count:, :dev_count]
ct_on_off_values = ct_on_off.values.T
ct_on_off_rows = list(ct_on_off.index)
ct_on_off_cols = list(ct_on_off.columns)
fig, (ax1, ax2) = plt.subplots(1,2, figsize=figsize)
# plot on-on corr
im, cbar = heatmap(ct_on_on_values, ct_on_on_devs, ct_on_on_devs,
ax=ax1, cmap='PuOr', cbarlabel='counts',
vmin=-1, vmax=1)
texts = annotate_heatmap(im, textcolors=("black", "white"), valfmt="{x:.2f}")
# plot on-off corr
im, cbar = heatmap(ct_on_off_values, ct_on_off_rows, ct_on_off_cols,
ax=ax2, cmap='PuOr', cbarlabel='counts',
vmin=-1, vmax=1)
texts = annotate_heatmap(im, textcolors=("black", "white"), valfmt="{x:.2f}")
ax1.set_title("Correlation of on-on signals")
ax2.set_title("Correlation of on-off signals")
fig.tight_layout()
plt.show()
def heatmap_transitions(df_act, z_scale=None, figsize=(8, 6), idle=False):
""" """
assert z_scale in [None,
'log'], 'z-scale has to be either of type None or log'
title = 'Activity transitions'
z_label = 'count'
df_act = add_idle(df_act) if idle else df_act
# get the list of cross tabulations per t_window
df = activities_transitions(df_act)
act_lst = list(df.columns)
x_labels = act_lst
y_labels = act_lst
values = df.values
log = True if z_scale == 'log' else False
valfmt = '{x:.0f}'
# begin plotting
fig, ax = plt.subplots(figsize=figsize)
im, cbar = heatmap_square(values,
y_labels,
x_labels,
log=log,
ax=ax,
cbarlabel=z_label)
texts = annotate_heatmap(im,
textcolors=("white", "black"),
log=log,
valfmt=valfmt)
ax.set_title(title)
return fig
def heatmap_contingency_triggers(df_dev_rep3,
df_act,
figsize=(12, 10),
idle=False,
z_scale=None):
"""
plot a heatmap TODO write docu
"""
title = 'Triggers'
cbarlabel = 'counts'
tmp = contingency_table_triggers(df_dev_rep3, df_act, idle=idle)
vals = tmp.values.T
acts = tmp.columns
devs = tmp.index
fig, ax = plt.subplots(figsize=figsize)
if z_scale == 'log':
log = True
else:
log = False
im, cbar = heatmap(vals, acts, devs, ax=ax, log=log, cbarlabel=cbarlabel)
texts = annotate_heatmap(im,
textcolors=("white", "black"),
log=log,
valfmt="{x}")
ax.set_title(title)
fig.tight_layout()
plt.show()
def heatmap_cross_correlation(df_dev, figsize=(10, 8)):
""" plots the cross correlation between the device signals
Parameters
----------
df_dev: pd.DataFrame
devices in representation 1
"""
cmap = 'BrBG'
cbarlabel = 'counts'
title = 'Cross-correlation of signals'
ct = duration_correlation(df_dev)
vals = ct.values.T
devs = list(ct.index)
fig, ax = plt.subplots(figsize=figsize)
im, cbar = heatmap_square(vals,
devs,
devs,
ax=ax,
cmap=cmap,
cbarlabel=cbarlabel,
vmin=-1,
vmax=1)
texts = annotate_heatmap(im,
textcolors=("black", "white"),
threshold=0.5,
valfmt="{x:.2f}")
ax.set_title(title)
fig.tight_layout()
plt.show()
def heatmap_trigger_time(df_dev, t_window='5s', figsize=(8, 8), z_scale=None):
color = 'trigger count'
cbarlabel = 'counts'
# get the list of cross tabulations per t_window
df = device_tcorr(df_dev, t_window)[0]
vals = df.astype(int).values.T
devs = list(df.index)
fig, ax = plt.subplots(figsize=figsize)
log = True if z_scale == 'log' else False
valfmt = "{x:.0f}"
im, cbar = heatmap_square(
vals,
devs,
devs,
ax=ax, #cmap='viridis',
cbarlabel=cbarlabel,
log=log) #, cbar_kw=cbar_kw)
texts = annotate_heatmap(im,
textcolors=("white", "black"),
log=log,
valfmt=valfmt)
ax.set_title("Triggercount with sliding window of " + t_window)
fig.tight_layout()
plt.show()
def heatmap_cross_correlation(df_dev, figsize=(10,8)):
ct = cross_correlation(df_dev)
vals = ct.values.T.astype(np.float64)
devs = list(ct.index)
fig, ax = plt.subplots(figsize=figsize)
im, cbar = heatmap_square(vals, devs, devs, ax=ax, cmap='BrBG', cbarlabel='counts',
vmin=-1, vmax=1)
texts = annotate_heatmap(im, textcolors=("black", "white"), valfmt="{x:.2f}",
threshold=0.6)
ax.set_title("Cross-correlation of signals")
fig.tight_layout()
plt.show()
def heatmap_contingency_01(X, y, rep='', z_scale=None, figsize=(16,12)):
""" plots the contingency between features and labels of data
Parameters
----------
X: pd.DataFrame
one of the representation raw, lastfired, changepoint
y: pd.DataFrame
a series of labels
rep: string
the name of the representation to add to the title
"""
cbarlabel = 'counts'
title = "On/Off contingency for representation " + rep
df_con = contingency_table_01(X, y)
vals = df_con.values.T
acts = df_con.columns
devs = list(df_con.index)
# format x labels
for i in range(0,len(devs)):
if i % 2 == 0:
tmp = devs[i][:-3]
devs[i] = tmp + 'Off'
else:
devs[i] = 'On'
if z_scale == 'log':
log = True
else:
log = False
fig, ax = plt.subplots(figsize=figsize)
im, cbar = heatmap(vals, acts, devs, ax=ax, log=log, cbarlabel=cbarlabel)
texts = annotate_heatmap(im, log=log, textcolors=("white", "black"), valfmt="{x:.0f}")
# create grid for heatmap into every pair
tcks = np.arange((vals.shape[1])/2)*2 + 1.5
ax.set_xticks(tcks, minor=True)
ax.grid(which="minor", color="w", linestyle='-', linewidth=2)
ax.tick_params(which="minor", bottom=False, left=False)
ax.set_title(title)
fig.tight_layout()
return fig
def heatmap_contingency_triggers_01(df_dev_rep3,
df_act,
figsize=(12, 10),
idle=True,
z_scale=None):
"""
"""
df_con = contingency_table_triggers_01(df_dev_rep3, df_act, idle=idle)
tmp = df_con.reset_index()
tmp['index'] = tmp['device'] + ' ' + tmp['val'].astype(str)
tmp = tmp.set_index('index')
tmp = tmp.drop(['device', 'val'], axis=1)
vals = tmp.values.T
acts = tmp.columns
devs = list(tmp.index)
# format x labels
for i in range(0, len(devs)):
if i % 2 == 0:
tmp = devs[i][:-5]
devs[i] = tmp + 'Off'
else:
devs[i] = 'On'
if z_scale == 'log':
log = True
else:
log = False
fig, ax = plt.subplots(figsize=figsize)
im, cbar = heatmap(vals, acts, devs, ax=ax, log=log, cbarlabel='counts')
texts = annotate_heatmap(im,
textcolors=("white", "black"),
log=log,
valfmt="{x}")
# create grid for heatmap into every pair
tcks = np.arange((vals.shape[1]) / 2) * 2 + 1.5
ax.set_xticks(tcks, minor=True)
ax.grid(which="minor", color="w", linestyle='-', linewidth=1)
ax.tick_params(which="minor", bottom=False, left=False)
ax.set_title("0-1 triggers")
fig.tight_layout()
return fig
def heatmap_cross_correlation(df_devs=None,
lst_devs=None,
df_dur_corr=None,
figsize=None,
numbers=None,
file_path=None):
"""
Plots the cross correlation between the device signals
Parameters
----------
df_devs : pd.DataFrame, optional
recorded devices from a dataset. Fore more information refer to the
:ref:`user guide<device_dataframe>`.
lst_devs : lst of str, optional
A list of devices that are included in the statistic. The list can be a
subset of the recorded activities or contain activities that are not recorded.
df_tcorr : pd.DataFrame
A precomputed correlation table. If the *df_tcorr* parameter is given, parameters
*df_devs* and *lst_devs* are ignored. The transition table can be computed
in :ref:`stats <stats_devs_tcorr>`.
figsize : (float, float), default: None
width, height in inches. If not provided, the figsize is inferred by automatically.
numbers : bool, default: True
Whether to display numbers inside the heatmaps fields or not.
file_path : str, optional
If set, saves the plot under the given file path and return *None* instead
of returning the figure.
Examples
--------
>>> from pyadlml.plots import plot_dev_hm_similarity
>>> plot_dev_hm_similarity(data.df_devices)
.. image:: ../_static/images/plots/dev_hm_dur_cor.png
:height: 400px
:width: 500 px
:scale: 90 %
:alt: alternate text
:align: center
Returns
-------
res : fig or None
Either a figure if file_path is not specified or nothing.
"""
assert not (df_devs is None and df_dur_corr is None)
title = 'Devices cross-correlation'
cmap = 'RdBu'
cbarlabel = 'similarity'
if df_dur_corr is None:
ct = duration_correlation(df_devs, lst_devs=lst_devs)
else:
ct = df_dur_corr
ct = ct.replace(pd.NA, np.inf)
vals = ct.values.T
devs = list(ct.index)
num_dev = len(devs)
figsize = (_num_items_2_heatmap_square_figsize_ver2(num_dev)
if figsize is None else figsize)
fig, ax = plt.subplots(figsize=figsize)
im, cbar = heatmap_square(vals,
devs,
devs,
ax=ax,
cmap=cmap,
cbarlabel=cbarlabel,
vmin=-1,
vmax=1)
if numbers is None:
if num_dev < 15:
valfmt = "{x:.2f}"
texts = annotate_heatmap(im,
textcolors=("black", "white"),
threshold=0.5,
valfmt=valfmt)
elif num_dev < 30:
valfmt = "{x:.1f}"
texts = annotate_heatmap(im,
textcolors=("black", "white"),
threshold=0.5,
valfmt=valfmt)
if numbers:
texts = annotate_heatmap(im,
textcolors=("black", "white"),
threshold=0.5,
valfmt="{x:.2f}")
ax.set_title(title)
fig.tight_layout()
if file_path is not None:
savefig(fig, file_path)
return
else:
return fig
def heatmap_trigger_time(df_devs=None,
lst_devs=None,
df_tcorr=None,
t_window='5s',
figsize=None,
z_scale="linear",
cmap=None,
numbers=None,
file_path=None):
"""
Plot todo
Parameters
----------
df_devs : pd.DataFrame, optional
recorded devices from a dataset. Fore more information refer to the
:ref:`user guide<device_dataframe>`.
lst_devs : lst of str, optional
A list of devices that are included in the statistic. The list can be a
subset of the recorded activities or contain activities that are not recorded.
df_tcorr : pd.DataFrame
A precomputed correlation table. If the *df_tcorr* parameter is given, parameters
*df_devs* and *lst_devs* are ignored. The transition table can be computed
in :ref:`stats <stats_devs_tcorr>`.
t_window : str of {'[1-12]h', default='1h'
the resolution, time_bins in hours. The number are
figsize : (float, float), default: None
width, height in inches. If not provided, the figsize is inferred by automatically.
z_scale : {"log", "linear"}, default: None
The axis scale type to apply.
numbers : bool, default: True
Whether to display numbers inside the heatmaps fields or not.
cmap : str or Colormap, optional
The Colormap instance or registered colormap name used to map scalar
data to colors. This parameter is ignored for RGB(A) data.
Defaults 'viridis'.
file_path : str, optional
If set, saves the plot under the given file path and return *None* instead
of returning the figure.
Examples
--------
>>> from pyadlml.plots import plot_device_hm_time_trigger
>>> plot_device_hm_time_trigger(data.df_devices, t_res='1h')
.. image:: ../_static/images/plots/dev_hm_trigger_one_day.png
:height: 300px
:width: 500 px
:scale: 100 %
:alt: alternate text
:align: center
Returns
-------
res : fig or None
Either a figure if file_path is not specified or nothing.
"""
assert not (df_devs is None and df_tcorr is None)
title = "Triggercount with sliding window of " + t_window
color = 'trigger count'
cbarlabel = 'counts'
if df_tcorr is None:
df = device_tcorr(df_devs, lst_devs=lst_devs, t_window=t_window)
else:
df = df_tcorr
# get the list of cross tabulations per t_window
vals = df.astype(int).values.T
devs = list(df.index)
num_dev = len(devs)
figsize = (_num_items_2_heatmap_square_figsize_ver2(num_dev)
if figsize is None else figsize)
cmap = (get_sequential_color() if cmap is None else cmap)
fig, ax = plt.subplots(figsize=figsize)
log = True if z_scale == 'log' else False
valfmt = "{x:.0f}"
im, cbar = heatmap_square(vals,
devs,
devs,
ax=ax,
cmap=cmap,
cbarlabel=cbarlabel,
log=log) #, cbar_kw=cbar_kw)
# show numbers for small sizes
if numbers is None:
if num_dev < 20:
texts = annotate_heatmap(im,
textcolors=("white", "black"),
log=log,
valfmt=valfmt)
elif numbers:
texts = annotate_heatmap(im,
textcolors=("white", "black"),
log=log,
valfmt=valfmt)
ax.set_title(title)
fig.tight_layout()
if file_path is not None:
savefig(fig, file_path)
return
else:
return fig
def heatmap_transitions(df_acts=None,
lst_acts=None,
df_trans=None,
z_scale="linear",
figsize=None,
idle=False,
numbers=True,
grid=True,
cmap=None,
file_path=None):
"""
Parameters
----------
df_acts : pd.DataFrame, optional
recorded activities from a dataset. Fore more information refer to the
:ref:`user guide<activity_dataframe>`.
lst_acts : lst of str, optional
A list of activities that are included in the statistic. The list can be a
subset of the recorded activities or contain activities that are not recorded.
df_trans : pd.DataFrame
A precomputed transition table. If the *df_trans* parameter is given, parameters
*df_acts* and *lst_acts* are ignored. The transition table can be computed
in :ref:`stats <stats_acts_trans>`.
figsize : (float, float), default: None
width, height in inches. If not provided, the figsize is inferred by automatically.
z_scale : {"log", "linear"}, default: None
The axis scale type to apply.
numbers : bool, default: True
Whether to display numbers inside the heatmaps fields or not.
idle : bool, default: False
Determines whether gaps between activities should be assigned
the activity *idle* or be ignored.
cmap : str or Colormap, optional
The Colormap instance or registered colormap name used to map scalar
data to colors. This parameter is ignored for RGB(A) data.
Defaults 'viridis'.
grid : bool, default: True
determines whether to display a white grid, seperating the fields or not.
file_path : str, optional
If set, saves the plot under the given file path and return *None* instead
of returning the figure.
Examples
--------
>>> from pyadlml.plots import plot_activity_hm_transitions
>>> plot_activity_hm_transitions(data.df_activities)
.. image:: ../_static/images/plots/act_hm_trans.png
:height: 300px
:width: 500 px
:scale: 90 %
:alt: alternate text
:align: center
Returns
-------
res : fig or None
Either a figure if file_path is not specified or nothing.
"""
assert z_scale in [None,
'log'], 'z-scale has to be either of type None or log'
assert not (df_acts is None and df_trans is None)
title = 'Activity transitions'
z_label = 'count'
if df_trans is None:
df_acts = add_idle(df_acts) if idle else df_acts
df = activities_transitions(df_acts, lst_acts=lst_acts)
else:
df = df_trans
# get the list of cross tabulations per t_window
act_lst = list(df.columns)
num_act = len(act_lst)
figsize = (_num_items_2_heatmap_square_figsize(num_act)
if figsize is None else figsize)
cmap = (get_sequential_color() if cmap is None else cmap)
x_labels = act_lst
y_labels = act_lst
values = df.values
log = True if z_scale == 'log' else False
valfmt = '{x:.0f}'
# begin plotting
fig, ax = plt.subplots(figsize=figsize)
im, cbar = heatmap_square(values,
y_labels,
x_labels,
log=log,
cmap=cmap,
ax=ax,
cbarlabel=z_label,
grid=grid)
if numbers:
texts = annotate_heatmap(im,
textcolors=("white", "black"),
log=log,
valfmt=valfmt)
ax.set_title(title)
if file_path is not None:
savefig(fig, file_path)
return
else:
return fig