Commits on Jan 8, 2024

1. Make all outputs in "streaming" format …

   Doing this to make it easier for Ryan to parse with a singular unified output format, regardless of whether the model parser actually supports streaming or not.
   
   - moved `runPromptStream` -> `runPrompt`, overriding the old definition of `runPrompt` by now passing in an `enableStreaming` flag
   - still relying on `isStreamingSupported()` function to set the `enableStreaming` param to true or false
   - default to `stream` param being `True` on backend server (however this has no effect for non-streaming models like Dall-E)
   
   ## Test Plan
   Test that both streaming and non-streaming settings work as expected, as well as test that a model which does not support streaming (ex: Dall-E) still works. Now when user hasn't explicitly clicked the "stream" setting, it will default to streaming. However if they explictly toggle it turns off. Follow up we should have the "stream" button auto-enabled to reflect this (doesn't have to actually be in the config, we should just have the UI show it as on by default to match user expectation)
   
   Update: Updated this so that the default value for `stream` is now `true` inside of `OpenAIChatModelParserPromptSchema`, `HuggingFaceTextGenerationParserPromptSchema` and `AnyscaleEndpointPromptSchema`. Couldn't see it defined in `PaLMTextParserPromptSchema`
   
   
   https://github.com/lastmile-ai/aiconfig/assets/151060367/34214a66-0cea-4774-a917-9476359f172c

   Rossdan Craig rossdan@lastmileai.dev committed Jan 8, 2024
   Configuration menu

   • View commit details
   • Copy full SHA for 5a811c3
   • Browse repository at this point

   Copy the full SHA

   5a811c3 View commit details

   Browse the repository at this point in the history