Integrate semantic retrieval into the AI chat layer

[dhruvi-16-me]

Summary

This PR integrates the existing semantic retrieval pipeline into Ell-ena's AI chat flow.

The changes replace broad workspace context injection with targeted semantic retrieval and introduce Riverpod-based state management for retrieval results and AI context.

Changes

Semantic Retrieval

• Integrated the existing Week 13 rag_search pipeline into the AI chat flow.
• Reused the existing retrieval infrastructure without introducing a duplicate search implementation.
• Added typed retrieval results with explicit handling for successful, empty, and failed retrievals.
• Added safe error handling to prevent internal database errors or sensitive information from being exposed.

Targeted AI Context

• Removed the previous loading of all assigned tasks and tickets during chat initialization.

• Removed unused userTasks and userTickets parameters from the AI flow.

• Gemini now receives only the workspace records relevant to the user's query.

• Excluded embeddings, similarity scores, and unnecessary database fields from the AI context.

• Preserved essential non-RAG context, including:

  • Current date
  • Team members
  • Conversation history
  • System guidelines
  • Tool schemas

Riverpod State Management

• Introduced Riverpod-based state management for the chat and retrieval flow.

• chatControllerProvider manages:

  • Conversation messages
  • Processing state
  • Retrieval status
  • Retrieval results
  • Retrieval errors
  • AI context

• Made ragRetrieverProvider overridable in tests, allowing retrieval logic to be tested without requiring a live database.

• Prevented duplicate retrieval calls during a single chat request.

Tool Context Optimization

• Reduced query_tasks and query_tickets responses to only the fields required by the AI.
• Preserved record IDs where required by modification tools.
• Removed unnecessary database data from Gemini prompts.

Testing

Added and updated tests covering:

• RAG result mapping
• AI context construction
• Irrelevant result exclusion
• Empty retrieval handling
• Retrieval error handling
• Prompt construction without RAG results
• ChatController success, empty, error, and conversation-history states

Validation

• ✅ flutter test — 17 tests passed
• ✅ flutter analyze — no new errors
• ✅ Dart formatting applied

✅ Checklist

• I have read the contributing guidelines.
• I have added tests that prove my fix is effective or that my feature works.
• I have added necessary documentation (if applicable).
• Any dependent changes have been merged and published in downstream modules.

[coderabbitai]

Important

Review skipped

Auto reviews are disabled on base/target branches other than the default branch.

🗂️ Base branches to auto review (2)

• main
• develop

Please check the settings in the CodeRabbit UI or the .coderabbit.yaml file in this repository. To trigger a single review, invoke the @coderabbitai review command.

⚙️ Run configuration

Configuration used: Path: .coderabbit.yaml

Review profile: ASSERTIVE

Plan: Pro Plus

Run ID: 1eaf5d38-6e95-405a-9efe-13b55fa50a01

You can disable this status message by setting the reviews.review_status to false in the CodeRabbit configuration file.

Use the checkbox below for a quick retry:

• 🔍 Trigger review

---

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

• X
• Mastodon
• Reddit
• LinkedIn

_{Comment @coderabbitai help to get the list of available commands.}