TS Label
The goal of this project is to provide mutable labels over time series data, where each labels is a span of time series data along its time axis.
Time series datasets can be seeded with labels that are immutable. These type's of labels are for providing useful context to the data.
Entities
Keys:
Ways of indexing into the entire time series dataset to find continuous regions of specific time series data. More than one key(s) represent a single entity in the database, since they index a single piece of time series data.
Creating:
- Creation of a set of keys is typically associated with adding some set of time series values to the database.
Updating:
- updating a set of keys cannot match another set of keys
- changing one or more keys will refer to the same set of data associated with the set of keys
Deleting:
- to delete keys, one must remove all the keys associated with a specific time series identifier. This means that if there are two keys that are responsible for referencing data then both keys must be present for deletion.
Labels:
Represent annotations over a span of time on some subset of the entire time series dataset.
Creating:
- creating a label that was most recently deleted with the same name is akin to reinstating the label
- to create a entity label of a most recently deleted label one must recreate and rename the most recently deleted label, delete it, and then create a new label of the desired name.
Updating:
- changing the name of a label will still refer to the same label entity
- cannot rename to a label entity that already exists
Deleting:
- deleting a label will remove the view of the label from the database