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
Allow for alternate sources, which could include a single or multiple documents, a folder of files, a GitHub repository, etc., that would extend the capability of the Predictor for use with local data without having the requirement of creating a separate Predictor which requires knowledge of creating code.
@mikenelson-io Can you please be more specific on the behavior and user experience you'd like to see? It would be helpful if you can provide some imagined scenarios as examples.
Example:
I am not a developer and have no real experience with Visual Studio. i am a system administrator that writes scripts to automate my work. I have created a simple PowerShell module that contains a few functions that I created to help me do my work faster and better. They are my own cmdlets with my own parameters, switches, etc. I will distribute this module to others within my organization and outside as open source.
I need a way to utilize the Predictor to assist people when using my module in the command line completion for the cmdlets I wrote. Again, I do not want to take on the burden of creating a Predictor in VS. I simply wish to provide a text or CVS file that contains the cmdlets and allowable parameters and switches. I want the PSReadline PredictionSource to allow me to specify the file so it can search that as part of discovery and provide the cmdlet assistance if it sees a match.
I see this as one example. Allowing the Prediction source to be something other than History or Plugin will allow a more diverse and inclusive discovery for any cmdlet.
So a generalized predictor that can be fed some local input as the "training data" that will be used in predictors. Me and @mikenelson-io talked about this and this does not have to be a different Predictor Source (i.e History, Plugin, none...) necessarily but could be a predictor module that is able to be configured to some location on file to use as the source.
I do really like this idea because yeah you are right that creating a predictor may be hard for folks unfamiliar with C# and binary module development but creating a configurable prediction source is easier way to enable more people. and enables proprietary scripts to be used as predictors 😉
Prerequisites
Description of the new feature/enhancement
Allow for alternate sources, which could include a single or multiple documents, a folder of files, a GitHub repository, etc., that would extend the capability of the Predictor for use with local data without having the requirement of creating a separate Predictor which requires knowledge of creating code.
Proposed technical implementation details (optional)
No response
The text was updated successfully, but these errors were encountered: