feat(prompts)!: prompt template management #5724
Draft
+46,645
−5,822
Conversation
This file contains bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
dosubot
bot
added
the
size:M
mikeldking
added
the
feature branch
RogerHYang
requested changes
Dec 12, 2024
mikeldking
force-pushed
the
prompts
branch
4 times, most recently
from
6da6bdc to
3c3c5e0
Compare
mikeldking
force-pushed
the
prompts
branch
2 times, most recently
from
b34c106 to
4973813
Compare
|
Check out this pull request on See visual diffs & provide feedback on Jupyter Notebooks. Powered by ReviewNB |
* wip * WIP * WIP * add the prompt type * add initial listing * feat: temporary prompt * final changes * add aria role
…rompt specific playground (#5748) * feat(prompts): Add prompt selection combobox to playground instances * Implement prompt/:id/playground route and add link from prompt page * Rename prompt combobox components for clarity
* feat(prompts): Implement SavePromptForm for creating new prompts * Rename button to 'save'
* Add prompt table migrations * Fix typo * Update migration test * Address migration feedback * Use default names for constraints * Add source prompt id
* cleanup * feat(prompts): rudimentary prompt details page #5741 * Update app/src/pages/prompt/PromptTabContent.tsx Co-authored-by: Anthony Powell <[email protected]> --------- Co-authored-by: Anthony Powell <[email protected]>
* Add PromptVersion query * Update src/phoenix/server/api/types/PromptVersion.py * reorganize * cleanup * Update the schema --------- Co-authored-by: Mikyo King <[email protected]>
* refactor(prompts): URL based routing and loader sharing for prompts - /prompts/:prompId has been converted to a layout component - sub-routes of promptId share loader data with promptLoader - prompt detail page tabs perform navigations * feat(prompts): Add query fragment to prompt versions sub-route * feat(prompts): Implement routing, loaders, and fragments for prompt version list * Add PromptVersion to node query * fix imports * Ensure all unset fields are set * fix comma * Distinguish prompt version details from one another * Move truncate component * Consolidate prompt tab nav logic --------- Co-authored-by: Dustin Ngo <[email protected]>
…5780) * feat(prompts): Render model invocation params in propt details view * Use fragment in PromptInvocationParameters * temporarily use component styles for lists until we can migrate * fix scrolling --------- Co-authored-by: Mikyo King <[email protected]>
* feat(prompts): Render prompt messages on prompt detail view * Remove unused import * Implement copy and paste button on prompt chat messages * Render chat messages on prompt version details page * Redirect to latest version when clicking versions tab Additionally highlight active version in versions list * Update example data to match template language * Fix prompt details page width, prompt versions borders / scroll * Distinguish readOnly TemplateEditor styling from normal TemplateEditor * Refactor active prompt id into prop * Replace custom button with styled anchor * Rename node in promptVersion query
* Flesh out PromptVersionTemplate type * Return GraphQL objects with the correct type * Use new types in node query * Decouple pydantic models and gql types * Rebuild gql schema * Rework model names * Update gql schema * Propagate name into schema * Incorporate feedback * Update schema * adjust UI to new schema * cleanup * Remove `hub` naming and clean up type annotations --------- Co-authored-by: Mikyo King <[email protected]>
* chore: fix lint errors * cleanup * cleanup
* feat(prompts): preview of last 5 versions * feat(prompts): preview of last 5 versions * rename
* Include prompt name in playground experiment name * Remove old type ignore
* feat(prompts): Denormalize tool choice before saving/using We were previously storing and sending tool choice in openai format, regardless of the destination provider. This caused silent failures with anthropic, and invalid validation on the backend (now that we are validating) * Clean up ToolChoiceSelector
* refactor(prompts): Reorganize llm schemas and converters * feat(prompts): Enhance LLM SDK provider converters and constants * refactor(prompts): Enforce zod schemas match openapi types
* feat(prompts): edit description * rudamentry styles * feat(prompts): update description * cleanup
…nes (#6325) * refactor(components): switch to leadingVisual * WIP * refactor
* feat(playground): hotkeys * key bindings * wip * feat(prompts): command enter to run the playground * cleanup * cleanup * add windows support * add missing
* fix(prompts): Make playground page blocking less intrusive * Only block dirty navigations on saved prompts
…#6330) If you saved an azure openai provider prompt, and then tried to load it back into playground, it would load as "openai" and you would lose your endpoint, and base url settings. Now we properly carry over azure_openai, and we will restore any saved azure (or openai) specific settings when loading prompt into playground
…roviders (#6339) * fix(prompts): Do not wipe playground invocation parameters when switching between providers * Ensure only supported invocation parameters are stored per-provider * Fix response format autocomplete
* feat(components): migrate text field * WIP * migrate more * final push * change message according to mode * Update app/.eslintrc.js Co-authored-by: Anthony Powell <[email protected]> --------- Co-authored-by: Anthony Powell <[email protected]>
* fix(prompts): Fix tool definition copy paste button It should copy current content of the editor, even if it is invalid.
* fix: restore playwrite tests * add ci
* docs(prompts): Add deno notebook example for TS phoenix-client * Add cross-sdk notebook example * Update text * Update import hash * Add Vercel AI SDK notebook
mikeldking
changed the title
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
Add this suggestion to a batch that can be applied as a single commit.
This suggestion is invalid because no changes were made to the code.
Suggestions cannot be applied while the pull request is closed.
Suggestions cannot be applied while viewing a subset of changes.
Only one suggestion per line can be applied in a batch.
Add this suggestion to a batch that can be applied as a single commit.
Applying suggestions on deleted lines is not supported.
You must change the existing code in this line in order to create a valid suggestion.
Outdated suggestions cannot be applied.
This suggestion has been applied or marked resolved.
Suggestions cannot be applied from pending reviews.
Suggestions cannot be applied on multi-line comments.
Suggestions cannot be applied while the pull request is queued to merge.
Suggestion cannot be applied right now. Please check back later.
resolves #5582