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:
...
# below we get price feature from the datasource transactions
# note, it is ModalityEvents object, not the value of the feature
price = attributes["transactions"]["price"]
...| Parameters |
|---|
column_name: str | list[str]
Name of the feature from the data source.
| Returns |
|---|
DataSourceEvents.__len__(self)
DataSourceEvents.__len__(self)Returns the length of the DataSourceEvents object, ie. the count of events.
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:
...
# sum up total transaction spend from future events
total_spend = future["transactions"].sum(column="value")
...
| 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:
...
# calculate average transaction price from future events
avg_price = 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:
...
# find minimum transaction price from future events
min_price = 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:
...
# find maximum transaction price from future events
max_price = 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:
...
# apply lowercase transformation to brand values in future transactions
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:
...
# filter transactions to include only products from the campaign
PRODUCTS_IN_CAMPAIGN = ["prd_1", "prd_2", "prd_3", "prd_4", "prd_5", "prd_6", "prd_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:
...
# check if purchases exist for nike or adidas brands in future transactions
target = (
future["transactions"]
.groupBy("brand")
.exists(groups=["Nike", "Adidas"])
)
...
| Parameters |
|---|
by: str | list[str]
Columns to group by.
| Returns |
|---|
A proxy object EventsGroupBy see this article.