Feat: Implement dual-mode search architecture (Phase 1)

Implements automatic mode detection and routing for CodeGraph search:

**Local Mode** (FAISS + local/ollama embeddings):
- Existing FAISS implementation extracted into faiss_search_impl()
- Used when CODEGRAPH_EMBEDDING_PROVIDER=local/ollama/lmstudio

**Cloud Mode** (SurrealDB HNSW + Jina embeddings):
- New cloud_search_impl() with basic SurrealDB integration
- Used when CODEGRAPH_EMBEDDING_PROVIDER=jina
- MVP: Returns filtered results (HNSW search + reranking TODO Phase 2)

**Mode Detection**:
- SearchMode enum (Local/Cloud)
- detect_search_mode() reads CODEGRAPH_EMBEDDING_PROVIDER
- bin_search_with_scores_shared() routes to correct implementation

**Documentation**:
- Updated .env.example with comprehensive dual-mode configuration
- Documents Local vs Cloud trade-offs
- SurrealDB connection settings for cloud mode

**Status**: Phase 1 complete, compiles successfully
**Next**: Phase 2 - SurrealDB HNSW search + Jina reranking integration

Author: jruokola

.env.example

@@ -13,7 +13,7 @@ CODEGRAPH_EMBEDDING_PROVIDER=auto
 
 # Embedding Provider Configuration
 # ----------------------------------
-# Provider options: "auto", "onnx", "ollama", "openai", or "lmstudio"
+# Provider options: "auto", "onnx", "ollama", "openai", "jina", or "lmstudio"
 # CODEGRAPH_EMBEDDING_PROVIDER=auto
 
 # ONNX: Specify model path (or leave empty for auto-detection from HuggingFace cache)
@@ -33,12 +33,54 @@ CODEGRAPH_EMBEDDING_PROVIDER=auto
 # OpenAI: Model name (API key configured below in Security section)
 # CODEGRAPH_EMBEDDING_MODEL=text-embedding-3-small
 
+# Jina AI: Cloud embeddings with reranking (requires JINA_API_KEY)
+# CODEGRAPH_EMBEDDING_PROVIDER=jina
+# JINA_API_KEY=your-jina-api-key-here
+
+# ============================================================================
+# Dual-Mode Search Configuration
+# ============================================================================
+# CodeGraph supports two search modes based on CODEGRAPH_EMBEDDING_PROVIDER:
+#
+# Local Mode (FAISS + local/ollama embeddings)
+# ---------------------------------------------
+# - Uses FAISS for in-memory vector search
+# - Embeddings: ONNX, Ollama, or LM Studio
+# - Best for: Desktop development, privacy-focused setups
+# - Requires: Build with --features faiss
+# Example:
+#   CODEGRAPH_EMBEDDING_PROVIDER=local  # or ollama or lmstudio
+#
+# Cloud Mode (SurrealDB HNSW + Jina embeddings + reranking)
+# ----------------------------------------------------------
+# - Uses SurrealDB HNSW indexes for scalable vector search
+# - Embeddings: Jina AI (2048 dimensions)
+# - Reranking: Jina reranker-v3 for improved relevance
+# - Best for: Cloud deployments, multi-user systems, scalability
+# - Requires: SurrealDB instance, Jina API key
+# Example:
+#   CODEGRAPH_EMBEDDING_PROVIDER=jina
+#   JINA_API_KEY=your-jina-api-key-here
+#
+# SurrealDB Connection (required for cloud mode)
+# SURREALDB_URL=ws://localhost:3004
+# SURREALDB_NAMESPACE=codegraph
+# SURREALDB_DATABASE=main
+# SURREALDB_USERNAME=root
+# SURREALDB_PASSWORD=root
+#
+# Important: HNSW index dimension must match embedding provider
+# - Jina v4: 2048 dimensions
+# - Local ONNX: typically 384 or 768 dimensions
+# - Update schema/codegraph.surql if changing providers
+
 # LLM Configuration (for local insights generation)
 # --------------------------------------------------
 # Leave empty to use context-only mode (fastest, recommended for agents like Claude/GPT-4)
 # Set to enable local LLM insights generation
 
 # LM Studio with DeepSeek Coder v2 Lite Instruct (recommended)
+# Supported LLM provider options: "auto", "onnx", "lmstudio", "openai", "claude" or "ollama"
 # Superior MLX support and Flash Attention 2 on macOS
 # CODEGRAPH_LLM_PROVIDER=lmstudio
 # CODEGRAPH_MODEL=lmstudio-community/DeepSeek-Coder-V2-Lite-Instruct-GGUF/DeepSeek-Coder-V2-Lite-Instruct-Q4_K_M.gguf

.gitignore

@@ -169,4 +169,5 @@ AGENTS.md
 CRUSH.md
 OUROBOROS.md
 .codegraph/
-SESSION-MEMORY.md
+SESSION-MEMORY.md
+.serena/

crates/codegraph-mcp/Cargo.toml

@@ -79,6 +79,7 @@ embeddings = ["dep:codegraph-vector"]
 embeddings-local = ["embeddings", "codegraph-vector/local-embeddings"]
 embeddings-openai = ["embeddings", "codegraph-vector/openai"]
 embeddings-ollama = ["embeddings", "codegraph-vector/ollama"]
+embeddings-jina = ["embeddings", "codegraph-vector/jina"]
 server-http = ["dep:axum", "dep:hyper"]
 qwen-integration = []
 ai-enhanced = ["dep:codegraph-ai", "faiss", "embeddings"]