You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
{{ message }}
This repository was archived by the owner on Dec 24, 2019. It is now read-only.
We need to refine the existing documentation at https://scava-docs.readthedocs.io/en/latest/users/Scava-Metrics/ . A good starting point would be to review each metric's description with a least add the type (string) unit (hours, urls..) and possibly the range and increment steps it is expected to get as values.
I'm quoting what mentioned in a previous correspondence about that:
I already told Yannis about that : as you now, Boris and I
initiated the "Scava Metrics" document [1]. That's a good start but
it still misses some important data for some of the metrics that could
ease up building our quality models:
what is the type of the value returned : Int ? String ?
list ? URL ?
what's the unit ? seconds ? hours ? scalar ?
what is the expected range values ?
what is the increment value within the interval ?
For example, if we take the metric "Topics" within the Newsgroup
metrics, it is described : "Using the clustering algorithm, we
determine which are the most frequent topics discussed in the
newsgroup.". The questions are then: what's the type of the returned
value ? How many references can we expect ? Do we expect URL(s) or topic
name(s) ? etc.
Of course, if the documentation provided from the REST API at /metrics endpoint is complete and structured enough we could choose to mention that the best way to get started with it is to use the REST API (and explain how).