The recommender system needs to connect to the Playground frontend to enable real-time recommendations to a user.
Requirements:
The workflow is written in Python and can be encapsulated into a container for easy use
Once a user makes a selection (e.g., ArchiveType), the system needs to receive that selection and returns a list of possible and probable choices for the next property in the hierarchy (e.g., Inferred/Measured variables). We need to enable an API to do so.
Redesign of the interface to support a mix of dropdown/text boxes. i.e. ability to select something already existing or write a new entry of the a user chooses "other"
Paths for time and depth VT's are working correctly
path for measured VT is almost done. I have data going in and out of the interpretation block (the hard part), but it needs to call the predict API one more time from within the interpretation block to get the prediction list for interpretationVariableDetail
new method for showing/hiding fields is working (for current tested paths)
encoding the API query url so for / and other characters.
resetting the fields if the VT is changed.
Still to go:
path for inferred VT
function: re-purposing function for getting all measured columns in a table for the inferredFrom field
function: copy fields from measured column to inferred column (when possible)
function: upon download, move variableNames to proxyObservationType and inferredVariableType fields where necessary.
The recommender system needs to connect to the Playground frontend to enable real-time recommendations to a user.
Requirements: