- Always stream the visible scenarios, if the modelProvider supports it
- Never stream the invisible scenarios
Also actually runs our query tasks in a background worker, which we weren't quite doing before.
* Move DeleteButton into a separate file
* Rename plural relations
* Add ability to fork
* Fork automatically after auth upon return
* Add experiment card skeleton
* Create HeaderButtons component
* return no header buttons while experiment loading
* Fix prettier
* Remove unused variable
* Remove newline
* Default json values to undefined
* Change header styles
* Fix prettier
* Give AddScenario icon less width
* Move useEffect
* Skip invalidating experiments list after forking
* Require user to be able to view experiment to fork it
* Move experiment creation into same transaction
* Only return the forked experiment id
* Put delete button in experiment settings drawer
* Move useEffect hook
I got somewhat confused by the extra fields, sorry.
Also makes some frontend changes to track that state more directly although in retrospect not sure the frontend changes make things any better.
* Add descriptions of fields in llama 2 input schema
* Let GPT-4 know when the provider stays the same
* Allow refetching in the event of any errors
* Define refinement actions in model providers
* Fix prettier
* Make CompareFunctions more configurable
* Change RefinePromptModal styles
* Accept newModel in getModifiedPromptFn
* Show prompt comparison in SelectModelModal
* Pass variant to SelectModelModal
* Update instructions
* Properly use isDisabled
Adds a `modelProvider` field to `promptVariants`, currently just set to "openai/ChatCompletion" for all variants for now.
Adds a `modelProviders/` directory where we can define and store pluggable model providers. Currently just OpenAI. Not everything is pluggable yet -- notably the code to actually generate completions hasn't been migrated to this setup yet.
Does a lot of work to get the types working. Prompts are now defined with a function `definePrompt(modelProvider, config)` instead of `prompt = config`. Added a script to migrate old prompt definitions.
This is still partial work, but the diff is large enough that I want to get it in. I don't think anything is broken but I haven't tested thoroughly.
