Skip to content

fix(nodes): remove llm_vertex, fix llm_gemini key handling, migrate vision nodes to LLMBase - #902

Merged
Rod-Christensen merged 4 commits into
developfrom
fix/gdg-fixes
May 16, 2026
Merged

Rod-Christensen merged 4 commits into
developfrom
fix/gdg-fixes

Conversation

@ryan-t-christensen

@ryan-t-christensen ryan-t-christensen commented May 16, 2026 •

Copy link
Copy Markdown
Collaborator

Summary

  • Remove llm_vertex node — Vertex AI is now supported directly through llm_gemini (via genai.Client(vertexai=True, ...)); the separate node is deprecated and deleted along with its icon, mock, and UI references
  • Fix llm_gemini API key handling — drop the strict startsWith('AI') format check (rejects valid keys), add sub-key fallback for old pipe files where the stored sub-key is profile.split('_', 1)[1] instead of the full profile name; also early-return in validateConfig when apikey is empty (secure fields aren't decrypted at validate time)
  • Migrate vision nodes to LLMBase — llm_vision_gemini and llm_vision_openai IInstance now inherit from ai.common.llm_base.LLMBase instead of the deprecated nodes.llm_base.IInstanceGenericLLM; also remove the redundant startsWith('AI') key check from gemini_vision.py
  • UI/infra cleanup — hide modelSource field (was readonly), remove vertex icon asset and map entry, bump client-typescript SDK to 1.1.0, update test mocks

Test plan

  • Open a pipeline with llm_gemini in the chat chain and confirm it initializes and responds
  • Open a pipeline with llm_vision_gemini and confirm image description still works
  • Open a pipeline with llm_vision_openai and confirm image description still works
  • Confirm no references to llm_vertex remain in the UI node picker
  • Run existing node tests (gemini-2_5-flash mock test case)

🤖 Generated with Claude Code

Summary by CodeRabbit

  • Removed Features

    • VertexAI (Enterprise & Personal) provider nodes, related docs and test mocks removed
  • Improvements

    • Gemini API key handling: profile-based fallback and relaxed validation
    • Task token extraction made more flexible for backward compatibility
    • Web UIs: URL auth parameter now takes priority for auth tokens
  • Updates

    • Gemini model profile IDs renamed to underscore format (e.g., gemini-3_1-*)
    • SDK version bumped to 1.1.0
    • A cloud model-source field hidden from the UI
    • Bundled "vertex" icon removed (falls back to unknown)

Review Change Stack

…ision nodes to LLMBase

- Remove deprecated llm_vertex node (Vertex AI now supported via llm_gemini)
- Fix llm_gemini API key validation: drop strict startsWith('AI') check, add
  sub-key fallback for old pipe files where sub-key is profile.split('_',1)[1]
- Fix llm_gemini validateConfig to early-return when apikey is empty (secure
  fields are not decrypted at validate time)
- Migrate llm_vision_gemini and llm_vision_openai IInstance from deprecated
  IInstanceGenericLLM to ai.common.llm_base.LLMBase
- Remove startsWith('AI') key format check from gemini_vision.py
- Remove vertex icon from UI icon map and asset bundle
- Hide modelSource field in services.common.llm.json (was readonly)
- Bump client-typescript SDK version to 1.1.0
- Update test mocks to remove langchain_google_vertexai mock

Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
@github-actions github-actions Bot added docs Documentation module:nodes Python pipeline nodes module:ui Chat UI and Dropper UI module:client-typescript TypeScript client SDK labels May 16, 2026
@github-actions

Copy link
Copy Markdown
Contributor

No description provided.

@coderabbitai

coderabbitai Bot commented May 16, 2026 •

Copy link
Copy Markdown
Contributor

No actionable comments were generated in the recent review. 🎉

ℹ️ Recent review info
⚙️ Run configuration

Configuration used: Path: .coderabbit.yaml

Review profile: ASSERTIVE

Plan: Pro

Run ID: 16f7e22f-8bd5-4170-a59d-835afc3250af

📥 Commits

Reviewing files that changed from the base of the PR and between 0357200 and 8414545.

📒 Files selected for processing (1)
  • apps/dropper-ui/src/App.tsx

📝 Walkthrough

Walkthrough

Removes Vertex AI artifacts; renames Gemini profiles to underscore IDs; Chat derives API keys from profile fallback and IGlobal minimal probe short-circuits; vision IInstance classes rebase to shared LLMBase; SDK/UI/version and auth/token extraction tweaks.

Changes

Vertex AI Removal and Gemini Modernization

Layer / File(s) Summary
Vertex AI module removal
nodes/src/nodes/llm_vertex/IGlobal.py, nodes/src/nodes/llm_vertex/IInstance.py, nodes/src/nodes/llm_vertex/vertex.py, nodes/src/nodes/llm_vertex/services.enterprise.json, nodes/src/nodes/llm_vertex/services.personal.json, nodes/src/nodes/llm_vertex/*/README.md, nodes/test/mocks/langchain_google_vertexai/*
Complete deletion of Vertex AI implementation and related mock/docs: IGlobal validation, lifecycle hooks, secret decoding, IInstance, service JSON configs, and mock module were removed.
Gemini profile identifier normalization
nodes/src/nodes/llm_gemini/services.json, nodes/src/nodes/llm_gemini/IGlobal.py, nodes/src/nodes/llm_gemini/gemini.py
Renamed Gemini profile IDs from dot to underscore across preconfig and fields; Chat initialization now derives missing API keys via connConfig.profile sub-profile lookup; IGlobal minimal probe now returns early when apikey is absent.
Vision model base class migration
nodes/src/nodes/llm_vision_gemini/IInstance.py, nodes/src/nodes/llm_vision_openai/IInstance.py, nodes/src/nodes/llm_vision_gemini/gemini_vision.py
IInstance classes now inherit from ai.common.llm_base.LLMBase; Gemini Vision API-key validation updated to reject OpenAI-style sk- keys.
Version and UI metadata cleanup
packages/client-typescript/src/client/constants.ts, packages/shared-ui/src/components/canvas/util/get-icon-path.ts, nodes/src/nodes/core/services.common.llm.json, nodes/test/mocks/__init__.py
SDK_VERSION bumped to 1.1.0; removed vertex icon import and iconMap entry; llm.cloud.modelSource marked hidden; mock docstring updated to remove Vertex references.
Task token backward-compatible extraction
packages/ai/src/ai/modules/task/task_conn.py
get_task_token now falls back from request.arguments['token'] to top-level request['token'] for compatibility.
Auth token URL-first precedence
apps/chat-ui/src/App.tsx, apps/dropper-ui/src/App.tsx
Startup auth resolution now prioritizes the auth URL query parameter, clears it from the URL when used, and falls back to sessionStorage or dev API key as before.

Sequence Diagram(s)

sequenceDiagram
  participant IGlobal as IGlobal.beginGlobal()
  participant Chat as Chat.__init__
  participant ConnConfig as connConfig
  participant SubProfile as subprofile lookup
  IGlobal->>Chat: Chat(connConfig)
  Chat->>ConnConfig: get('profile') (e.g. 'gemini-2_5-pro')
  ConnConfig-->>Chat: profile string
  Chat->>SubProfile: extract substring after first '_' (e.g. '2_5-pro' -> '5-pro')
  SubProfile-->>Chat: subprofile key
  Chat->>ConnConfig: get(subprofile).get('apikey')
  ConnConfig-->>Chat: api_key or None
  alt api_key present
    Chat->>Chat: genai.Client(api_key=api_key)
  else
    Chat->>Chat: raise ValueError("Please enter your Gemini API key.")
  end
Loading

Estimated code review effort

🎯 4 (Complex) | ⏱️ ~45 minutes

Possibly related PRs

Suggested reviewers

  • jmaionchi
  • stepmikhaylov
  • asclearuc
  • Rod-Christensen

Poem

A rabbit nibbles code at dawn,
Vertex gates quietly withdrawn,
Gemini swaps dots for underscore cheer,
Vision shares a base — hops forward, not rear,
SDK hops up a version — carrots near! 🐇

🚥 Pre-merge checks | ✅ 5
✅ Passed checks (5 passed)
Check name Status Explanation
Description Check ✅ Passed Check skipped - CodeRabbit’s high-level summary is enabled.
Title check ✅ Passed The title accurately summarizes the three main changes: removing llm_vertex, fixing llm_gemini key handling, and migrating vision nodes to LLMBase.
Docstring Coverage ✅ Passed Docstring coverage is 100.00% which is sufficient. The required threshold is 80.00%.
Linked Issues check ✅ Passed Check skipped because no linked issues were found for this pull request.
Out of Scope Changes check ✅ Passed Check skipped because no linked issues were found for this pull request.

✏️ Tip: You can configure your own custom pre-merge checks in the settings.

✨ Finishing Touches
📝 Generate docstrings
  • Create stacked PR
  • Commit on current branch
🧪 Generate unit tests (beta)
  • Create PR with unit tests
  • Commit unit tests in branch fix/gdg-fixes

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.

❤️ Share

Comment @coderabbitai help to get the list of available commands and usage tips.

get_task_token() now checks arguments.token first, then falls back to
request.token so older clients and debug command token injection still work.

Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>
@github-actions github-actions Bot added the module:ai AI/ML modules label May 16, 2026
The URL param carries the freshly-minted key for the current task and
must not be shadowed by a stale sessionStorage value from a previous
task on the same origin.

Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>

@coderabbitai coderabbitai Bot left a comment

Copy link
Copy Markdown
Contributor

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Actionable comments posted: 1

🤖 Prompt for all review comments with AI agents
Verify each finding against current code. Fix only still-valid issues, skip the
rest with a brief reason, keep changes minimal, and validate.

Inline comments:
In `@apps/dropper-ui/src/App.tsx`:
- Around line 71-74: The conditional is redundant: when token is set from
urlParams.get('auth') the search already contains 'auth=', so remove the inner
check and unconditionally call window.history.replaceState({}, '',
window.location.pathname) inside the if (token) block (referencing token,
urlParams.get('auth') and window.history.replaceState in App.tsx) to match the
chat-ui behavior.
🪄 Autofix (Beta)

Fix all unresolved CodeRabbit comments on this PR:

  • Push a commit to this branch (recommended)
  • Create a new PR with the fixes

ℹ️ Review info
⚙️ Run configuration

Configuration used: Path: .coderabbit.yaml

Review profile: ASSERTIVE

Plan: Pro

Run ID: 5a8b167d-b09f-43b4-9e49-8d53d5913bba

📥 Commits

Reviewing files that changed from the base of the PR and between 9e17976 and 0357200.

📒 Files selected for processing (2)
  • apps/chat-ui/src/App.tsx
  • apps/dropper-ui/src/App.tsx

Comment thread apps/dropper-ui/src/App.tsx Resolved
Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>
@Rod-Christensen
Rod-Christensen merged commit 283291c into develop May 16, 2026
24 of 26 checks passed
@Rod-Christensen
Rod-Christensen deleted the fix/gdg-fixes branch May 16, 2026 09:49
@coderabbitai coderabbitai Bot mentioned this pull request May 16, 2026
7 tasks
nihalnihalani added a commit to nihalnihalani/rocketride-server that referenced this pull request May 19, 2026
Resolves conflicts in llm_vision_{gemini,openai}/IInstance.py from upstream
rocketride-org#902 (migrate vision nodes to LLMBase). Both sides add the same
`from ai.common.llm_base import LLMBase` import; kept the existing
HEAD position adjacent to the other ai.common imports and dropped the
duplicate inserted by upstream. Circuit-breaker behavior (CircuitBreakerOpenError
handling + _mark_current_object_retry calls) preserved.
AviNormie pushed a commit to AviNormie/rocketride-server that referenced this pull request Nov 9, 2026
…ision nodes to LLMBase (rocketride-org#902)

* fix(nodes): remove llm_vertex, fix llm_gemini key handling, migrate vision nodes to LLMBase

- Remove deprecated llm_vertex node (Vertex AI now supported via llm_gemini)
- Fix llm_gemini API key validation: drop strict startsWith('AI') check, add
  sub-key fallback for old pipe files where sub-key is profile.split('_',1)[1]
- Fix llm_gemini validateConfig to early-return when apikey is empty (secure
  fields are not decrypted at validate time)
- Migrate llm_vision_gemini and llm_vision_openai IInstance from deprecated
  IInstanceGenericLLM to ai.common.llm_base.LLMBase
- Remove startsWith('AI') key format check from gemini_vision.py
- Remove vertex icon from UI icon map and asset bundle
- Hide modelSource field in services.common.llm.json (was readonly)
- Bump client-typescript SDK version to 1.1.0
- Update test mocks to remove langchain_google_vertexai mock

Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>

* fix(task): add backward-compat fallback for token in DAP request root

get_task_token() now checks arguments.token first, then falls back to
request.token so older clients and debug command token injection still work.

Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>

* fix(ui): prioritize URL auth param over sessionStorage token

The URL param carries the freshly-minted key for the current task and
must not be shadowed by a stale sessionStorage value from a previous
task on the same origin.

Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>

* fix(ui): remove redundant auth check in dropper-ui to match chat-ui

Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>

---------

Co-authored-by: Claude Sonnet 4.6 <noreply@anthropic.com>
Co-authored-by: Rod.Christensen <rod.christensen@rocketride.ai>
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Labels

docs Documentation module:ai AI/ML modules module:client-typescript TypeScript client SDK module:nodes Python pipeline nodes module:ui Chat UI and Dropper UI

Projects

None yet

Development

Successfully merging this pull request may close these issues.

2 participants