Fix Embeddings

[yilinxia]

No description provided.

[Copilot]

Pull Request Overview

This PR renames the GraphRAG system to DeepGitAI throughout the codebase, representing a rebranding from "GraphRAG" to "DeepGitAI" and adds support for semantic search through repository embeddings.

• Comprehensive renaming from GraphRAG to DeepGitAI across all files
• Added embedding functionality for semantic search on README content
• Enhanced dependency management with new machine learning packages

Reviewed Changes

Copilot reviewed 12 out of 12 changed files in this pull request and generated 4 comments.

Show a summary per file

File	Description
start.sh	Updated command line output to show DeepGitAI instead of GraphRAG
src/views/GraphRAGPanel.tsx	Renamed component to DeepGitAIPanel and updated all references from GraphRAG to DeepGitAI
src/views/ContextPanel.tsx	Updated imports and panel references to use DeepGitAI naming
src/styles/_layout.scss	Updated CSS class names and styling from graphrag to deepgit-ai
src/styles/_graph.scss	Updated CSS selectors and class names to reflect DeepGitAI branding
src/lib/config.ts	Renamed API endpoints from GraphRAG to DeepGitAI format
environment.yml	Added ML dependencies for embeddings: sentence-transformers and torch
backend/requirements.txt	Added sentence-transformers and torch for embedding functionality
backend/app/services/deepgit_ai_service.py	Major service rename with added embedding support and semantic search capabilities
backend/app/main.py	Updated all Flask endpoints and variable names from GraphRAG to DeepGitAI
backend/app/config_manager.py	Updated configuration keys from graphrag to deepgit_ai
backend/app/config.example.json	Updated example configuration with DeepGitAI settings

Comments suppressed due to low confidence (1)

backend/app/services/deepgit_ai_service.py:1399

• The SQL query is missing the embedding fields that are now part of the data structure. The query should include SET r.readme_embedding = $embedding, r.embedding_provider = $embedding_provider to match the data being processed in the loop.

                        MERGE (r:Repository {id: $repo_id})
                        SET r.readme_content = $content, r.readme_length = $length

---

_{Tip: Customize your code reviews with copilot-instructions.md. Create the file or learn how to get started.}

[Copilot]

The inline comment with field descriptions should be updated to match the actual parameter order and names. The fields appear to be readme_content, readme_length, readme_embedding, embedding_provider but the comment shows different spacing and structure.

Suggested change

	contributors, stargazers, '', 0, None, None # readme_content, readme_length, readme_embedding, embedding_provider
	contributors, stargazers, # up to stargazers
	'', # readme_content
	0, # readme_length
	None, # readme_embedding
	None # embedding_provider

[Copilot]

These ALTER TABLE statements are executed inside the exception handler but the retry query at line 1412 is still missing the embedding fields. The retry should include both readme_embedding and embedding_provider in the SET clause.

[Copilot]

The CSS class name 'deepgit-ai-chat' should be consistent with the hyphenated naming convention used elsewhere in the codebase. Consider using 'deepgit-ai-chat' consistently or update the corresponding CSS to match.

Suggested change

	<div className="d-flex flex-column h-100 deepgit-ai-chat" style={{ minHeight: 0 }}>
	<div className="d-flex flex-column h-100 deepgitai-chat" style={{ minHeight: 0 }}>

[Copilot]

[nitpick] The database schema includes embedding fields but there's no validation to ensure the embedding array has a consistent dimension. Consider adding a constraint or comment documenting the expected embedding dimension (e.g., 384 for sentence-transformers/all-MiniLM-L6-v2).

Suggested change

	readme_length INT64,
	readme_length INT64,
	-- readme_embedding is expected to be a 384-dimensional vector (e.g., from sentence-transformers/all-MiniLM-L6-v2)