[Amazing-Stardom]

Summary

• Updated the langchaingo Go library to the latest version (v0.1.15).
• Changed the default AI provider and model shown in the UI to OpenRouter with DeepSeek Pro v4 (deepseek/deepseek-pro-v4).
• Updated the helper text and backend model catalog so DeepSeek Pro v4 is automatically selected as default.

Validation

• Ran go mod tidy to make sure all Go dependencies are clean and matched.
• Ran npx tsc --noEmit in the UI directory to ensure there are no frontend type errors.

Checklist

• This PR fulfills an agreed issue.
• I kept the change narrow and scoped.
• I ran the most specific relevant validation and described it above.
• If I changed behavior, I called that out clearly in this PR.
• If I touched UI, I attached a GIF or video walkthrough. This is required.
• If this change touches security, disclosure flow, credentials, storage, network behavior, or licensing/entitlement enforcement, I reviewed SECURITY.md and updated related documentation if needed.
• I have read and accept the Contributor License Agreement.

Notes For Reviewers

• Creating a new connector for OpenRouter will now automatically fill in deepseek/deepseek-pro-v4 instead of leaving it empty or pointing to the older route.

[LiveReview-Bot]

Migrate Codebase to Gemini 3.8 Flash and Remove Sampling Parameters

Overview

The codebase migrates to Gemini 3.8 Flash as the default AI model. All integrations, scripts, and configuration templates now omit unsupported sampling parameters like temperature, top_p, and top_k.

Technical Highlights

• internal/ai/gemini/gemini.go: Updates the default model constant to Gemini 3.8 Flash and strips unsupported generation configuration payloads.
• internal/aiconnectors/connector.go: Adds conditional checks to omit temperature, top_p, and top_k options for Gemini models.
• internal/ai/langchain/provider.go: Introduces buildCallOptions to conditionally exclude temperature settings for Gemini and Vertex AI.

Impact

• Functionality: Standardizes model usage on Gemini 3.8 Flash and prevents API validation errors from unsupported settings.
• Risk: Custom configurations or external scripts using removed sampling parameters will ignore those values.

[LiveReview-Bot]

Update Default AI Models and Remove Temperature Parameters

Overview

The system migrates to Gemini 3.8 Flash and DeepSeek Pro v4. It also removes deprecated temperature and sampling parameters across all AI providers.

Technical Highlights

• internal/ai/gemini/gemini.go: Sets default model to Gemini 3.8 Flash and removes legacy temperature fields.
• internal/aiconnectors/connector.go: Adds helper logic to bypass temperature and sampling parameters for Gemini providers.
• config/livereview.toml.example: Updates default sample configuration to use the new Gemini and DeepSeek models.

Impact

• Functionality: Modernizes default AI models and prevents invalid parameter errors on Gemini APIs.
• Risk: Existing custom configurations relying on removed temperature parameters may require manual cleanup.

[LiveReview-Bot]

Adapt to Gemini 3.8 API Parameter Deprecations

Overview

This change adapts the system to Google's Gemini 3.x API updates. It removes deprecated sampling parameters like temperature. This ensures continued compatibility with Gemini 3.8 Flash models.

Technical Highlights

• config/livereview.toml.example: Updates the default Gemini model; removes the temperature parameter.
• docs/gemini_3.8_migration.md: Documents Gemini 3.x parameter deprecations; lists required code changes.
• go.mod: Upgrades the langchaingo dependency to v0.1.15.
• internal/ai/gemini/gemini.go: Sets gemini-3.8-flash as the default model; removes all deprecated parameters.
• internal/ai/langchain/provider.go: Adds buildCallOptions to conditionally apply llms.WithTemperature.
• internal/aiconnectors/connector.go: Excludes deprecated parameters from llms.CallOptions for Gemini.
• internal/config/config.go: Removes temperature from generated Gemini configurations.
• scripts/prelivi/interpretation.py: Removes hardcoded temperature from Gemini REST calls.
• ui/src/pages/AIProviders/AIProviders.tsx: Sets gemini-3.8-flash as the default in UI metadata.

Impact

• Functionality: Enables continued, error-free use of Gemini 3.x models.
• Risk: Existing configurations or scripts not updated could cause API 400 INVALID_ARGUMENT errors.