class monad.targets.DataSourceEvents
monad.targets.DataSourceEventsContains main entity events data source defined in the pretrain phase.
Properties
DataSourceEvents.timestamps
DataSourceEvents.timestampsTimestamps of the events.
| Returns |
|---|
numpy.ndarray
Methods
DataSourceEvents.__getitem__(self, column_name)
DataSourceEvents.__getitem__(self, column_name)def target_fn(history: Events, future: Events, attributes: Attributes, ctx: Dict) -> np.ndarray:
...
# here we get price feature from the datasource transactions
# Note, it is ModalityEvents object, not the value of the feature
city = attributes["transactions"]["price"]
...| Parameters |
|---|
column_name: str | list[str]
Name of the feature from the data source.
| Returns |
|---|
DataSourceEvents.__len__(self)
DataSourceEvents.__len__(self)def target_fn(history: Events, future: Events, attributes: Attributes, ctx: Dict) -> np.ndarray:
...
no_events = len(future["transactions"])
...| Returns |
|---|
int
DataSourceEvents.count(self)
DataSourceEvents.count(self)Calculates the number of events.
def target_fn(history: Events, future: Events, attributes: Attributes, ctx: Dict) -> np.ndarray:
...
churn = 0 if future['product_buy'].count() > 0 else 1
...| Returns |
|---|
int
DataSourceEvents.sum(self, column)
DataSourceEvents.sum(self, column)Calculates the sum of feature values.
def target_fn(history: Events, future: Events, attributes: Attributes, ctx: Dict) -> np.ndarray:
...
future['transactions'].sum(column='price')
...| Parameters |
|---|
column: str
Name of the feature from the data source.
| Returns |
|---|
float
DataSourceEvents.mean(self, column)
DataSourceEvents.mean(self, column)Calculates the mean of feature values.
def target_fn(history: Events, future: Events, attributes: Attributes, ctx: Dict) -> np.ndarray:
...
future['transactions'].mean(column='price')
...| Parameters |
|---|
column: str
Name of the feature from the data source.
| Returns |
|---|
float
DataSourceEvents.min(self, column)
DataSourceEvents.min(self, column)Calculates the minimum of feature values.
def target_fn(history: Events, future: Events, attributes: Attributes, ctx: Dict) -> np.ndarray:
...
future['transactions'].min(column='price')
...| Parameters |
|---|
column: str
Name of the feature from the data source.
| Returns |
|---|
float
DataSourceEvents.max(self, column)
DataSourceEvents.max(self, column)Calculates the maximum of feature values.
def target_fn(history: Events, future: Events, attributes: Attributes, ctx: Dict) -> np.ndarray:
...
future['transactions'].max(column='price')
...| Parameters |
|---|
column: str
Name of the feature from the data source.
| Returns |
|---|
float
DataSourceEvents.apply(self, func, target)
DataSourceEvents.apply(self, func, target)Applies function to a target column.
def target_fn(history: Events, future: Events, attributes: Attributes, ctx: Dict) -> np.ndarray:
...
trans_modified = future['transactions'].apply(lambda x: x.lower()), target='brand')
...| Parameters |
|---|
func: Callable[[Any], Any]
Function to apply.
target: str
Name of the target column.
| Returns |
|---|
DataSourceEvents with column target transformed.
DataSourceEvents.filter(self, by, condition)
DataSourceEvents.filter(self, by, condition)Filters events based on the condition checked against column by.
def target_fn(history: Events, future: Events, attributes: Attributes, ctx: Dict) -> np.ndarray:
...
PRODUCTS_IN_CAMPAIGN = ["prod_1", "prod_2", "prod_3", "prod_4", "prod_5", "prod_6", "prod_7"]
trans_filtered = future['transactions'].filter(by="product_id", condition=lambda x: x in PRODUCTS_IN_CAMPAIGN)
...| Parameters |
|---|
by: str
Column to check condition against.
condition: Callable[[Any], bool]
Filtering condition.
| Returns |
|---|
DataSourceEvents with filtered events.
DataSourceEvents.groupBy(self, by)
DataSourceEvents.groupBy(self, by)Groups the events by values in a column (or list of columns) provided after by.
def target_fn(history: Events, future: Events, attributes: Attributes, ctx: Dict) -> np.ndarray:
...
PRODUCTS_IN_CAMPAIGN = ["prod_1", "prod_2", "prod_3", "prod_4", "prod_5", "prod_6", "prod_7"]
future['transactions'].groupBy('brand').exists(groups=['Nike', 'Adidas'])
...| Parameters |
|---|
by: str | list[str]
Columns to group by.
| Returns |
|---|
A proxy object EventsGroupBy .