[PR #2520] [MERGED] feat(vertex_ai): enhance Gemini 3 model configs and upgrade dependencies #2539

Closed
opened 2026-02-16 11:17:16 -05:00 by yindo · 0 comments
Owner

📋 Pull Request Information

Original PR: https://github.com/langgenius/dify-official-plugins/pull/2520
Author: @trongtrandp
Created: 1/29/2026
Status: Merged
Merged: 1/31/2026
Merged by: @crazywoola

Base: mainHead: feat/vertex-ai-gemini3-enhancements


📝 Commits (2)

  • 636532d feat(vertex_ai): enhance Gemini 3 model configs and upgrade dependencies
  • a89e95b Merge branch 'main' into feat/vertex-ai-gemini3-enhancements

📊 Changes

5 files changed (+494 additions, -43 deletions)

View changed files

📝 models/vertex_ai/manifest.yaml (+1 -1)
📝 models/vertex_ai/models/llm/gemini-3-flash-preview.yaml (+50 -12)
📝 models/vertex_ai/models/llm/gemini-3-pro-preview.yaml (+50 -12)
📝 models/vertex_ai/requirements.txt (+3 -3)
📝 models/vertex_ai/uv.lock (+390 -15)

📄 Description

Summary

Enhance Gemini 3 model configurations with new features and upgrade Google SDK dependencies.

Changes

Model Configs (gemini-3-pro-preview.yaml, gemini-3-flash-preview.yaml)

  • Add Minimal and Medium thinking levels (previously only Low/High)
  • Add Grounding feature - Google Search integration for fact-checking
  • Add URL Context feature - browse and analyze URLs for reference
  • Add Code Execution feature - let Gemini use code to solve tasks
  • Add structured-output feature support
  • Improve temperature parameter description with note: "For Gemini 3, best results at default 1.0. Lower values may impact reasoning."
  • Improve max_output_tokens parameter labels and descriptions

Code (llm.py)

  • Add handler for grounding parameter (backward compatible with grounding_search)
  • Add handler for url_context parameter → types.Tool(url_context=types.UrlContext())
  • Add handler for code_execution parameter → types.Tool(code_execution=types.ToolCodeExecution())
  • Dynamic thinking_level mapping to handle SDK version differences (Minimal/Low/Medium/High)

Dependencies (requirements.txt)

  • Upgrade google-genai from 1.51.0 → 1.56.0
  • Upgrade google-auth from 2.29.0 → >=2.45.0
  • Change google-api-core to >=2.23.0

Test plan

  • Verified SDK types exist: GoogleSearch, UrlContext, ToolCodeExecution
  • Verified Tool creation syntax works
  • Test Gemini 3 Pro/Flash with all thinking levels (Minimal, Low, Medium, High)
  • Test Grounding feature (Google Search)
  • Test URL Context feature
  • Test Code Execution feature

🤖 Generated with Claude Code


🔄 This issue represents a GitHub Pull Request. It cannot be merged through Gitea due to API limitations.

## 📋 Pull Request Information **Original PR:** https://github.com/langgenius/dify-official-plugins/pull/2520 **Author:** [@trongtrandp](https://github.com/trongtrandp) **Created:** 1/29/2026 **Status:** ✅ Merged **Merged:** 1/31/2026 **Merged by:** [@crazywoola](https://github.com/crazywoola) **Base:** `main` ← **Head:** `feat/vertex-ai-gemini3-enhancements` --- ### 📝 Commits (2) - [`636532d`](https://github.com/langgenius/dify-official-plugins/commit/636532d8df75252dac75c95e92ec6c2b2a3904a4) feat(vertex_ai): enhance Gemini 3 model configs and upgrade dependencies - [`a89e95b`](https://github.com/langgenius/dify-official-plugins/commit/a89e95ba5a0bc3a04a55a06f859b4c75ff07ea9e) Merge branch 'main' into feat/vertex-ai-gemini3-enhancements ### 📊 Changes **5 files changed** (+494 additions, -43 deletions) <details> <summary>View changed files</summary> 📝 `models/vertex_ai/manifest.yaml` (+1 -1) 📝 `models/vertex_ai/models/llm/gemini-3-flash-preview.yaml` (+50 -12) 📝 `models/vertex_ai/models/llm/gemini-3-pro-preview.yaml` (+50 -12) 📝 `models/vertex_ai/requirements.txt` (+3 -3) 📝 `models/vertex_ai/uv.lock` (+390 -15) </details> ### 📄 Description ## Summary Enhance Gemini 3 model configurations with new features and upgrade Google SDK dependencies. ## Changes ### Model Configs (`gemini-3-pro-preview.yaml`, `gemini-3-flash-preview.yaml`) - Add `Minimal` and `Medium` thinking levels (previously only Low/High) - Add `Grounding` feature - Google Search integration for fact-checking - Add `URL Context` feature - browse and analyze URLs for reference - Add `Code Execution` feature - let Gemini use code to solve tasks - Add `structured-output` feature support - Improve `temperature` parameter description with note: "For Gemini 3, best results at default 1.0. Lower values may impact reasoning." - Improve `max_output_tokens` parameter labels and descriptions ### Code (`llm.py`) - Add handler for `grounding` parameter (backward compatible with `grounding_search`) - Add handler for `url_context` parameter → `types.Tool(url_context=types.UrlContext())` - Add handler for `code_execution` parameter → `types.Tool(code_execution=types.ToolCodeExecution())` - Dynamic `thinking_level` mapping to handle SDK version differences (Minimal/Low/Medium/High) ### Dependencies (`requirements.txt`) - Upgrade `google-genai` from 1.51.0 → 1.56.0 - Upgrade `google-auth` from 2.29.0 → >=2.45.0 - Change `google-api-core` to >=2.23.0 ## Test plan - [x] Verified SDK types exist: `GoogleSearch`, `UrlContext`, `ToolCodeExecution` - [x] Verified Tool creation syntax works - [ ] Test Gemini 3 Pro/Flash with all thinking levels (Minimal, Low, Medium, High) - [ ] Test Grounding feature (Google Search) - [ ] Test URL Context feature - [ ] Test Code Execution feature --- 🤖 Generated with [Claude Code](https://claude.ai/code) --- <sub>🔄 This issue represents a GitHub Pull Request. It cannot be merged through Gitea due to API limitations.</sub>
yindo added the pull-request label 2026-02-16 11:17:16 -05:00
yindo closed this issue 2026-02-16 11:17:16 -05:00
Sign in to join this conversation.
1 Participants
Notifications
Due Date
No due date set.
Dependencies

No dependencies set.

Reference: langgenius/dify-official-plugins#2539