Log

The model log is useful for debugging a model and for optimizing model performance.

HubitLog dataclass

Hubit log. For each query, various run data such as the number of workers spawned and the executions time is added to the log as a LogItem as the first element.

Often your simply want to print the log for a HubitModel instance e.g. print(hmodel.log()).

get_all(self, attr)

Get all log item values corresponding to attribute name attr.

Examples:

To get the elapsed time for all queries on the HubitModel instance hmodel execute hmodel.log().get_all("elapsed_time"). If two queries has been executed on the model, the return value is a list of times e.g. [0.5028373999812175, 0.6225477000162937] where the first element represent the elapsed time for for latest query.

Parameters:

Name	Type	Description	Default
attr	str	Valid values are attributes names of the LogItem class.	required

Returns:

Type	Description
List	Log item values for attr

LogItem dataclass

Hubit log item. Keys in all attribute dicts (e.g. worker_counts and cache_counts) are the same.

Parameters:

Name	Type	Description	Default
elapsed_time	float	Query execution time	required
worker_counts	Dict[str, int]	For each component function name the value is the count of spawned workers.	required
cache_counts	Dict[str, int]	For each component function name the value is the number of workers that used the cache.	required