feat: Enhance Qwen provider support and documentation#1083
Closed
LoneWolf36 wants to merge 12 commits intocoleam00:mainfrom
Closed
feat: Enhance Qwen provider support and documentation#1083LoneWolf36 wants to merge 12 commits intocoleam00:mainfrom
LoneWolf36 wants to merge 12 commits intocoleam00:mainfrom
Conversation
…m00#1023) * feat: make artifact file paths clickable in chat messages (coleam00#1016) Artifact paths in AI responses (e.g. `artifacts/runs/{uuid}/report.md`) were rendered as plain inline code. Now they render as clickable buttons (.md opens ArtifactViewerModal) or links (other types open in new tab). Changes: - Add artifact path detection via regex in MessageBubble code renderer - Convert MARKDOWN_COMPONENTS to factory function for modal state access - Add ArtifactViewerModal integration in chat view - Add 'workflow_artifact' to WORKFLOW_EVENT_TYPES in store.ts Fixes coleam00#1016 Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com> * refactor(web): improve MessageBubble type clarity and code comments - Replace opaque `ComponentPropsWithoutRef<never>` return type on `makeMarkdownComponents` with the idiomatic `Components` type from react-markdown, which is the actual contract the prop expects - Add explanatory comment on `ARTIFACT_PATH_RE` describing group semantics and cross-platform path handling intent - Add comment on `useMemo` empty dep array explaining that `setArtifactViewer` is a stable React state setter Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com> * simplify: merge duplicate lucide-react imports in MessageBubble Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com> * fix(web): harden extractArtifactInfo UUID regex and path traversal check Make ARTIFACT_PATH_RE case-insensitive for hex digits ([a-fA-F0-9-]+) so uppercase UUID characters are matched correctly. Also reject filenames containing '..' path segments to prevent path traversal. Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com> * fix(web): add artifact path detection to WorkflowResultCard markdown The artifact-aware code component was only in MessageBubble.tsx but WorkflowResultCard in MessageList.tsx used its own static markdown components without artifact detection. Artifact paths in workflow results were rendering as plain code spans instead of clickable links. Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com> * fix(server,web): artifact route uses codebase name for owner/repo + bright link styling The artifact route derived owner/repo from the working_path, which fails for worktree-based runs (worktrees use local filesystem username, not GitHub username). Now looks up the codebase record by codebase_id to get the correct owner/repo from the codebase name. Also fixes artifact link styling in WorkflowResultCard — uses text-accent-bright instead of text-[inherit] which was inheriting the dim text-text-secondary color from the parent container. Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com> * fix: resolve merge conflict markers in api.ts Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com> * style(web): brighter artifact link color in workflow result cards Bumped artifact link color from accent-bright (0.72) to a brighter oklch(0.78 0.18 250) with font-medium for better readability against the dark text-text-secondary card background. Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com> * style(web): use inline styles for artifact links to guarantee visibility Tailwind classes were being overridden by parent text-text-secondary. Inline styles ensure the bright blue color and pointer cursor always apply regardless of CSS cascade. Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com> * fix(web): use <a> tags for all artifact links with !text-accent-bright Replaced <button> with <a> for .md artifacts (pointer cursor by default), unified styling using !text-accent-bright to override parent text-text-secondary, hover brightens to oklch(0.85). All artifact links now consistently bright and clickable. Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com> * fix(web,server): restore missing import, fix artifact link colors and cursor - Restore getArchonWorkspacesPath import removed by prior commit (still used in 3 other places in api.ts, breaking the build) - Use text-accent-bright instead of dark text-accent for artifact links in MessageBubble (was nearly invisible on dark background) - Add cursor-pointer to .md artifact buttons in MessageBubble - Replace arbitrary oklch hover value with hover:!text-primary in MessageList artifact links Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com> --------- Co-authored-by: Claude Opus 4.6 (1M context) <noreply@anthropic.com>
fix(server): use BUNDLED_VERSION for app version in binary mode
Integrate Qwen as a first-class assistant across config, workflows, CLI setup, API schemas, and the web UI. Add a Qwen smoke test and tighten the docs so the remaining provider-specific limits are explicit. This keeps Archon aligned with the new Qwen path while documenting where behavior still differs from Claude and Codex. Co-Authored-By: Codex GPT-5 <noreply@openai.com>
Preserve existing Qwen settings when assistant config is updated so the web settings flow no longer drops qwen entries from global config. Stop forcing OpenAI auth in the Qwen client and allow the SDK to follow its native auth resolution unless Archon is configured explicitly. Also pass through OpenAI-compatible env vars and wire the Qwen smoke test into the package test script so the reviewed behavior stays covered. Co-Authored-By: Codex GPT-5 <noreply@openai.com>
Prevent duplicate tool-call emission when partial Qwen streaming emits a tool start before the final assistant message, and pass native Qwen auth environment variables through the subprocess allowlist. Co-Authored-By: Codex GPT-5 <noreply@openai.com>
- Config persistence: unconditionally preserve qwen in updateGlobalConfig() - CLI setup: filter default assistant options by actually configured assistants - Add mode: read hasQwen and defaultAssistant from existing config instead of hardcoding - Missing env vars: inject config.envVars into Qwen node options in dag-executor - Provider misrouting: return undefined instead of 'codex' for unknown models - Streaming dedup: track sawPartialText separately from tool_use partials - Security: scope Qwen credentials to QwenClient only, remove from shared allowlist - UI robustness: add optional chaining for config.assistants.qwen?.model in SettingsPage Co-Authored-By: Codex GPT-5 <noreply@openai.com> Co-authored-by: Qwen-Coder <qwen-coder@alibabacloud.com>
…ator warnings - Schema: make qwen optional+nullable in safeConfigSchema for unset state - Docstring: fix authType docstring to reflect SDK-resolved default (not 'openai') - Validator: add Qwen warnings for MCP, skills, hooks, and tool restrictions (previously only warned for Codex; now warns for both non-Claude providers) Co-Authored-By: Codex GPT-5 <noreply@openai.com> Co-authored-by: Qwen-Coder <qwen-coder@alibabacloud.com>
- Improve model recognition for Qwen variants (qwen-coder, qwen-max, qwen-turbo, qwen-plus, qwq-*, qvq-*) - Add comprehensive test coverage for Qwen model validation and provider inference - Create detailed Qwen setup guide in docs/qwen-setup.md - Update README.md to highlight Qwen as first-class AI assistant - Add configuration examples and migration guide for binary v0.3.5 users - Ensure all Qwen model patterns are properly recognized in isQwenModel() - Add tests for provider inference with various Qwen model names - Document Qwen-specific workflow configuration patterns Closes: Qwen support gap between source code and binary v0.3.5 Co-authored-by: Qwen-Coder <qwen-coder@alibabacloud.com>
|
Important Review skippedAuto reviews are disabled on base/target branches other than the default branch. Please check the settings in the CodeRabbit UI or the ⚙️ Run configurationConfiguration used: defaults Review profile: CHILL Plan: Pro Run ID: You can disable this status message by setting the Use the checkbox below for a quick retry:
✨ Finishing Touches🧪 Generate unit tests (beta)
Thanks for using CodeRabbit! It's free for OSS, and your support helps us grow. If you like it, consider giving us a shout-out. Comment |
Collaborator
|
This PR looks similar to #1085, which was closed on April 11, 2026. You may want to read the discussion there before pushing further. |
Collaborator
|
Closing this for the same reason as the Qwen implementation please follow the community registry pattern and resolve conflicts |
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
3 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.
Summary
Problem: Archon binary v0.3.5 had incomplete Qwen support despite source code being fully ready for Qwen integration.
Why it matters: Qwen users couldn't use Archon binary releases effectively - they had to build from source to get proper Qwen support.
What changed:
qwen-coder,qwen-max,qwen-turbo,qwen-plus,qwq-*,qvq-*,coder-model)model-validation.ts)docs/qwen-setup.md) with authentication, configuration, and troubleshootingWhat did not change: Core
QwenClientimplementation (already complete and production-ready in source)Changes
1. Enhanced Model Recognition (
packages/workflows/src/model-validation.ts)Extended
isQwenModel()to recognize all common Qwen model naming patterns:qwen-*prefix (qwen-coder, qwen-max, qwen-turbo, qwen-plus)qwq-*prefix (Qwen reasoning models)qvq-*prefix (Qwen vision models)*-coder-modelsuffixcoder-modelexact matchThis ensures provider inference works correctly when workflows specify Qwen models.
2. Comprehensive Test Coverage (
packages/workflows/src/model-validation.test.ts)Added 15 new test cases covering:
Results: 25 tests passing, 100% code coverage on
model-validation.ts3. Documentation (
docs/qwen-setup.md)Created comprehensive 300+ line setup guide including:
4. README Updates (
README.md)Added "Supported AI Assistants" section:
Testing
All tests pass:
Manual testing with Qwen provider:
Checklist
Related Issues
Addresses user reports of Qwen provider not working in binary releases despite source code supporting it.
Note: The core Qwen implementation was already complete in the source code. This PR focuses on enhancing model recognition, adding comprehensive tests, and creating documentation to make Qwen a truly first-class citizen in Archon.