feat(web): add image generation model config to routing settings#82
Open
DivXPro wants to merge 33 commits into
Open
feat(web): add image generation model config to routing settings#82DivXPro wants to merge 33 commits into
DivXPro wants to merge 33 commits into
Conversation
Add ImageGenModelSettingControl to web routing settings page, using platform settings (image_generative.model) for persistence. Also increase image model test timeout from 15s to 60s to accommodate slower image generation APIs. Co-Authored-By: Claude Opus 4.7 <noreply@anthropic.com>
There was a problem hiding this comment.
💡 Codex Review
Here are some automated review suggestions for this pull request.
Reviewed commit: 5d05b3593e
ℹ️ About Codex in GitHub
Codex has been enabled to automatically review pull requests in this repo. Reviews are triggered when you
- Open a pull request for review
- Mark a draft as ready
- Comment "@codex review".
If Codex has suggestions, it will comment; otherwise it will react with 👍.
When you sign up for Codex through ChatGPT, Codex can also answer questions or update the PR, like "@codex address that feedback".
| setTestErrorMenuStyle(null) | ||
| try { | ||
| if (value === '') { | ||
| await updatePlatformSetting(accessToken, IMAGE_GENERATIVE_MODEL_KEY, '') |
There was a problem hiding this comment.
Use DELETE when clearing image model setting
Selecting the empty option (Not Set) currently sends PUT /v1/admin/platform-settings/image_generative.model with an empty value, but the API rejects empty values (value must not be empty in platformapi/v1_platform_settings.go). In this scenario, users cannot clear the image generation model and the async handler rejects, so the UI cannot persist the unset action. This branch should call deletePlatformSetting(...) instead of updatePlatformSetting(..., '').
Useful? React with 👍 / 👎.
- Add image_generate to AUXILIARY_RENDERED_TOOL_NAMES - Create StreamingArtifactEntry on image_generate tool-result - Render image artifacts via ArtifactImage instead of ArtifactIframe - Filter inline-rendered image artifacts from MarkdownRenderer to avoid duplication Co-Authored-By: Claude Opus 4.7 <noreply@anthropic.com>
- Add messageAttachmentStore to image_generate executor with WithMessageAttachmentStore builder
- loadInputImages fallback to attachment store when artifact store misses
- artifactKeyMatchesAccount supports attachments/{accountID}/ prefix
- Inject uploaded image key annotation into LLM context so model can reference attachments
Co-Authored-By: Claude Opus 4.7 <noreply@anthropic.com>
… in messages - Scan all assistant messages for artifact:<key> references - Only show image_generate streaming artifact at top-level when the key is NOT already referenced in an assistant message bubble - Remove currentRunArtifacts filtering so MarkdownRenderer can render images during streaming inside the message bubble Co-Authored-By: Claude Opus 4.7 <noreply@anthropic.com>
…liminate flash - Never show image_generate in visibleStreamingArtifacts during streaming - Compute fallbackImageGenerateArtifacts only after run completes and the artifact key is not referenced in messages - Preserve image_generate streaming artifacts on run completion for fallback - Update persona prompts and toolmeta: image_generate does not need visualize_read_me Co-Authored-By: Claude Opus 4.7 <noreply@anthropic.com>
…story for fallback rendering Do not clear streamingArtifactsRef when releasing completed handoff to history. image_generate artifacts need to persist for fallback rendering after the run ends. They are cleared by resetLiveState() when the next run starts. Co-Authored-By: Claude Opus 4.7 <noreply@anthropic.com>
- Persist pending form messages as thread messages on run.input_requested - Finalize form messages (submitted/dismissed) on input submission - Expire pending forms on terminal run end (completed/cancelled/failed) - Add AskUserFormMessageCard component for editable and read-only views - Route form-mode input-request events through message stream instead of overlay - Widen web message types to support AskUserFormContent tagged union Co-Authored-By: Claude Opus 4.7 <noreply@anthropic.com>
…assthrough - normalizeInputRequest now preserves display_mode field from SSE events - Add display_mode to AgentInputRequestData type - Create optimistic form message in message store on form-mode input-request - Remove setAwaitingInput for form mode (form card handles all input) - Rewrite AskUserFormMessageCard from wizard to unified scrollable form - Update display_mode description in tool spec Co-Authored-By: Claude Opus 4.7 <noreply@anthropic.com>
Co-Authored-By: Claude Opus 4.7 <noreply@anthropic.com>
Co-Authored-By: Claude Opus 4.7 <noreply@anthropic.com>
Co-Authored-By: Claude Opus 4.7 <noreply@anthropic.com>
Co-Authored-By: Claude Opus 4.7 <noreply@anthropic.com>
Co-Authored-By: Claude Opus 4.7 <noreply@anthropic.com>
Co-Authored-By: Claude Opus 4.7 <noreply@anthropic.com>
Co-Authored-By: Claude Opus 4.7 <noreply@anthropic.com>
Co-Authored-By: Claude Opus 4.7 <noreply@anthropic.com>
Co-Authored-By: Claude Opus 4.7 <noreply@anthropic.com>
…rlay, remove filename bar, remove count label Co-Authored-By: Claude Opus 4.7 <noreply@anthropic.com>
…mage support Replace single artifactRef with generatedImages data model to handle multiple images per image_generate call without overwriting. Images render via GeneratedImageGroup in the assistant turn path, with markdown suppression via suppressedArtifactKeys to prevent duplicates. Fix: when hasAssistantTurn=true, skip passing generatedImages to MessageBubble to avoid double rendering in both segment and bubble paths. Co-Authored-By: Claude Opus 4.7 <noreply@anthropic.com>
… skill Active trigger is better controlled via ecommerce system SKILL or MCP. Co-Authored-By: Claude Opus 4.7 <noreply@anthropic.com>
Split Step 1 into two explicit ask_user calls (product info then image config) to ensure module selection and resolution are always collected. Add guardrail to Step 4: image_generate must not be called before user confirms prompts. Co-Authored-By: Claude Opus 4.7 <noreply@anthropic.com>
ArrayEnumFieldSchema items.enum does not support enumNames in the ask_user form renderer. Switch to items.anyOf (const + title) so users see Chinese module names instead of raw enum keys. Co-Authored-By: Claude Opus 4.7 <noreply@anthropic.com>
…rwrites Multiple image_generate calls within the same run produced identical artifact keys (accountID/runID/generated-image.png), causing later calls to overwrite earlier images. Add toolCallID to the key path so each call produces a unique artifact key. Co-Authored-By: Claude Opus 4.7 <noreply@anthropic.com>
- Move form-mode ask_user from message flow to input area during filling; submitted/dismissed forms appear in chat flow only via message sync - Remove "Submitted"/"Dismissed" status badge from submitted form view - Fix SelectField/OneOfSelectField dropdown clipping by using createPortal + position:fixed with auto up/down direction, outside click dismiss, and scroll-close to prevent misalignment - Extract shared PopoverSelect component to eliminate duplication between SelectField and OneOfSelectField Co-Authored-By: Claude Opus 4.7 <noreply@anthropic.com>
Co-Authored-By: Claude Opus 4.7 <noreply@anthropic.com>
The dedup logic in extractFilesFromTransfer was keyed by MIME type prefix
('image'), so only one image file survived each drop. Switch to identity-based
dedup (name|size|lastModified) so that truly distinct files all pass through
while still suppressing genuine duplicates from Electron clipboard bugs.
Co-Authored-By: Claude Opus 4.7 <noreply@anthropic.com>
SkyAerope
requested changes
May 26, 2026
This file contains hidden or 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
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
2 participants
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.
Add ImageGenModelSettingControl to web routing settings page, using platform settings (image_generative.model) for persistence. Also increase image model test timeout from 15s to 60s to accommodate slower image generation APIs.