Context defines all properties, dataframes, models and other data passed through pipeline steps.
- properties are accessible on properties property.
- output dataframe(s) use return statement.
- input dataframe(s) get them by parameters or use dataframe property directly.
- for model use the set_model statement : support only one model per script.
- for plot use the set_model statement : support only one plot per script.
- dataframes, models and additional plots used within the pipelines can also be accessed, read-only.
Attributes
Name | Type | Description |
---|---|---|
api | ApiInterface | api instance with limited functionality accessible during runner execution. |
config | dict | all Wizata custom.ini configuration keys. |
dataframe | pandas.Dataframe | single instance of pandas Dataframe inputted as parameter. |
dataframes | dict | dict of all dataframes parameters and generated by previously executed steps. |
datapoints | dict | wizata_dsapi.Datapoint for each name (key) used in queries. |
execution_id | uuid.UUID | Execution ID currently processed. |
grafana_api | grafana_client.GrafanaApi | instance of grafana api. |
models | dict | dict of all models generated by previously executed steps. |
now | datetime | Now timestamp as used by the pipeline. |
pipeline_id | uuid.UUID | Pipeline ID associated with the ongoing execution. |
plots | dict | dict of all plots generated by previously executed steps. |
properties | dict | properties dictionary containing useful information such as variables. |
step | PipelineStep | Currently processed step definition. |
template | Template | Template if set on the pipeline. |
registration | TwinRegistration | Registration associated with the pipeline and the twin currently processed. |
warnings | list | all warnings and error messages. |
Methods
append()
append an object (pandas.Dataframe or any properties)
Name | Type | Default | Description |
---|---|---|---|
key | str | dictionary identifier - name inside your pipeline. | |
obj | ML Model, Dataframe or any properties (must be JSON serializable type). | ||
overwrite | bool | True | by default - allow modifying an existing object. can be set to false. |
current_dataframes()
-
current dataframes a dictionary with all current named dataframes specific for this script.
-
dataframes contains all accessible dataframes for the pipeline mapped.
-
single dataframe context is accessible with context.dataframe and is not named
get()
get key from either dataframes, models, plots or properties.
Name | Type | Default | Description |
---|---|---|---|
key | str |
return: None if not found.
get_model()
get model to be added to the context.
get_model_config()
extract model configuration from the context.
get_plot()
get plot set to be added to the context.
get_script_config()
extract script configuration from the context.
notify_execution()
notify the listeners and watchers on current execution status.
Name | Type | Default | Description |
---|---|---|---|
execution | Execution |
notify_step()
notify the listeners and watchers on current step status.
Name | Type | Default | Description |
---|---|---|---|
step_log | ExecutionStepLog |
reset()
reset context between step execution - remove all step info, but keep all data.
set_model()
set model to be added to the context.
Name | Type | Default | Description |
---|---|---|---|
trained_model | Trained Model to be stored as a pickled object. | ||
input_columns | |||
output_columns | None | ||
has_anomalies | False | ||
scaler | None | Scaler to be stored if necessary. |
return: ML Model object prepared.
set_plot()
set plot to be added to the context.
Name | Type | Default | Description |
---|---|---|---|
figure | Plotly figure. | ||
name | Unkwown | Name of the plot. |
return: Plot object prepared.
show()
return plot figure to be shown in a notebook.
return: plot figure
write_log()
write log in console (only in experiment mode) - if running locally it will print in the console.
Name | Type | Default | Description |
---|---|---|---|
message | str | message to write. | |
level | int | 7 | from 7=DEBUG, 6=INFO, 3=ERROR to 1=CRITICAL |