def transform(self, df_activities):
df = label_data(self.df_index, df_activities, self.idle)
df = df.iloc[self.window_size:, :]
encoded_labels = self._lbl_enc.transform(df['activity'].values)
return pd.DataFrame(index=df.index,
data=encoded_labels,
columns=['activity'])
def transform(self, X):
"""
Parameters
----------
X :
"""
#if isinstance(X, pd.DataFrame) and set(X.columns) == {START_TIME, END_TIME, ACTIVITY}:
# df = label_data(X, self.df_acts, self.idle)
# encoded_labels = self._lbl_enc.fit_transform(df[ACTIVITY].values)
# return pd.DataFrame(index=df[TIME], data=encoded_labels, columns=[ACTIVITY])
if isinstance(X, pd.DataFrame) and TIME in X.columns:
df = label_data(X, self.df_acts_, self.idle)
encoded_labels = self.lbl_enc.transform(df[ACTIVITY].values)
return pd.DataFrame(data={
TIME: df[TIME].values,
ACTIVITY: encoded_labels
})
# return only the labels for a nd array
elif isinstance(X, np.ndarray):
return self.lbl_enc.transform(X)
else:
raise ValueError
def fit(self, df_activities, y=None):
""" labels data and creates the numeric representations
Parameters
----------
df_activities : pd.DataFrame
Columns are end_time, start_time, activity.
"""
df = label_data(self.df_devices, df_activities, self.idle)
self._lbl_enc.fit(df[ACTIVITY].values)
def transform(self, x):
"""
"""
# if the input is a dataframe of activities, than fit and
# transform the data accordingly
col_names = {'start_time', 'end_time', ACTIVITY}
if isinstance(x, pd.DataFrame) and set(x.columns) == col_names:
df = label_data(self.df_devices, x, self.idle)
encoded_labels = self._lbl_enc.transform(df[ACTIVITY].values)
return pd.DataFrame(index=df.index, data=encoded_labels, columns=[ACTIVITY])
# return only the labels for a nd array
elif isinstance(x , np.ndarray):
return self._lbl_enc.transform(x)
else:
raise ValueError
def contingency_table_triggers(df_devs, df_acts, idle=False):
"""
output:
read like this: dev 1 was 123 times triggered while act 1 was present
example
---------------------------
| act 1 | .... | act n|
-------------------------------
dev 1 | 123 | | 123 |
...
dev n | 123 | | 123 |
"""
df = label_data(df_devs, df_acts, idle=idle)
df['val'] = 1
return pd.pivot_table(df,
columns='activity',
index='device',
values='val',
aggfunc=len,
fill_value=0)
def fit_transform(self, df_acts, X):
"""
Fit label encoder and return encoded labels.
Parameters
----------
df_acts : pd.DataFrame, optional
recorded activities from a dataset. Fore more information refer to the
:ref:`user guide<activity_dataframe>`.
Returns
------
df : pd.DataFrame
"""
self.fit(df_acts)
df = label_data(X, self.df_acts_, self.idle)
encoded_labels = self.lbl_enc.fit_transform(df[ACTIVITY].values)
return pd.DataFrame(data={
TIME: df[TIME].values,
ACTIVITY: encoded_labels
})
def contingency_table_triggers(df_devs, df_acts, idle=False):
"""
Compute the amount of device triggers occuring during the different activities.
Parameters
----------
df_devs : pd.DataFrame
All recorded devices from a dataset. For more information refer to
:ref:`user guide<device_dataframe>`.
df_acts : pd.DataFrame
All recorded activities from a dataset. Fore more information refer to the
:ref:`user guide<activity_dataframe>`.
idle : bool
Determines whether gaps between activities should be assigned
the activity *idle* or be ignored.
Examples
--------
>>> from pyadlml.stats import contingency_triggers
>>> contingency_triggers(data.df_devices, data.df_activities)
activity get drink go to bed ... use toilet
device
Cups cupboard 36 0 ... 0
Dishwasher 2 0 ... 0
... ... ... ... ...
Washingmachine 0 0 ... 0
[7 rows x 7 columns]
Results
-------
df : pd.DataFrame
"""
df = label_data(df_devs, df_acts, idle=idle)
df[VAL] = 1
return pd.pivot_table(df,
columns=ACTIVITY,
index=DEVICE,
values=VAL,
aggfunc=len,
fill_value=0)
def contingency_table_triggers_01(df_devs, df_acts, idle=False):
"""
output:
read like this: dev 1 turned 123 times from 1 to 0 while act 1 was present
example
---------------------------
| act 1 | .... | act n|
-------------------------------
dev 1 0 | 123 | | 123 |
dev 1 1 | 122 | | 141 |
...
dev n 0 | 123 | | 123 |
dev n 1 | 122 | | 141 |
"""
df = label_data(df_devs, df_acts, idle=idle)
df['val2'] = df['val'].astype(int)
return pd.pivot_table(df,
columns='activity',
index=['device', 'val'],
values='val2',
aggfunc=len,
fill_value=0)
def fit(self, df_activities):
df = label_data(self.df_index, df_activities, self.idle)
# start where the labeling begins
df = df.iloc[self.window_size:, :]
self._lbl_enc.fit(df['activity'].values)
def transform(self, df_activities):
df = label_data(self.df_devices, df_activities, self.idle)
encoded_labels = self._lbl_enc.transform(df['activity'].values)
return pd.DataFrame(index=df.index,
data=encoded_labels,
columns=['activity'])
def fit(self, df_activities):
df = label_data(self.df_devices, df_activities, self.idle)
self._lbl_enc.fit(df['activity'].values)
def contingency_table_triggers_01(df_devs, df_acts, idle=False):
"""
Compute the amount a device turns "on" or "off" respectively
during the different activities.
Parameters
----------
df_devs : pd.DataFrame
All recorded devices from a dataset. For more information refer to
:ref:`user guide<device_dataframe>`.
df_acts : pd.DataFrame
All recorded activities from a dataset. Fore more information refer to the
:ref:`user guide<activity_dataframe>`.
idle : bool, optional, default: False
Determines whether gaps between activities should be assigned
the activity *idle* or be ignored.
Examples
--------
>>> from pyadlml.stats import contingency_triggers_01
>>> contingency_triggers_01(data.df_devices, data.df_activities)
activity get drink go to bed ... use toilet
devices ...
Cups cupboard Off 18 0 ... 0
Cups cupboard On 18 0 ... 0
Dishwasher Off 1 0 ... 0
Dishwasher On 1 0 ... 0
... ... ... ... ...
Washingmachine Off 0 0 ... 0
Washingmachine On 0 0 ... 0
[14 rows x 7 columns]
Results
-------
df : pd.DataFrame
"""
dev_index = 'devices'
ON = 'On'
OFF = 'Off'
df = label_data(df_devs, df_acts, idle=idle)
df['val2'] = df[VAL].astype(int)
df = pd.pivot_table(df,
columns=ACTIVITY,
index=[DEVICE, VAL],
values='val2',
aggfunc=len,
fill_value=0)
# format text strings
def func(x):
if "False" in x:
return x[:-len("False")] + " Off"
else:
return x[:-len("True")] + " On"
df = df.reset_index()
df[dev_index] = df[DEVICE] + df[VAL].astype(str)
df[dev_index] = df[dev_index].apply(func)
df = df.set_index(dev_index)
df = df.drop([DEVICE, VAL], axis=1)
return df
