Remove outdated documentation and integration test plans

- Deleted the architecture design and implementation plan for CodeGraph.
- Removed the integration test plan document, which outlined testing strategies and execution commands.
- Eliminated the testing context summary document that provided an overview of the testing environment and constraints.
- Cleared the monitoring setup documentation detailing the Docker-based monitoring stack for the application.
- Removed the Rust SDK development plan for the native Rust client SDK for MCP, including development context and library research.

Author: jruokola

.codegraph/index.json

@@ -1,12 +1,12 @@
 {
   "project_path": ".",
-  "indexed_at": "2025-09-11T22:40:30.157223Z",
+  "indexed_at": "2025-09-12T05:16:15.023218Z",
   "stats": {
-    "files": 0,
-    "lines": 161517,
-    "functions": 4197,
+    "files": 315,
+    "lines": 162442,
+    "functions": 4209,
     "classes": 0,
-    "embeddings": 0,
+    "embeddings": 13509,
     "errors": 0
   },
   "config": {

.gitignore

@@ -168,4 +168,4 @@ CLAUDE.md
 AGENTS.md
 CRUSH.md
 OUROBOROS.md
-.codegraph/index.json
+.codegraph/index.json

README.md

@@ -104,16 +104,16 @@ How to use ONNX embeddings
 # CPU-only (default)
 export CODEGRAPH_EMBEDDING_PROVIDER=onnx
 export CODEGRAPH_ONNX_EP=cpu
-export CODEGRAPH_LOCAL_MODEL=sentence-transformers/all-MiniLM-L6-v2
+export CODEGRAPH_LOCAL_MODEL=/path/to/onnx-file
 
 # CoreML (requires CoreML-enabled ORT build)
 export CODEGRAPH_EMBEDDING_PROVIDER=onnx
 export CODEGRAPH_ONNX_EP=coreml
-export CODEGRAPH_LOCAL_MODEL=sentence-transformers/all-MiniLM-L6-v2
+export CODEGRAPH_LOCAL_MODEL=/path/to/onnx-file
 
 
 # Install codegraph
-cargo install --path crates/codegraph-mcp --features "embeddings,codegraph-vector/onnx"
+cargo install --path crates/codegraph-mcp --features "embeddings,codegraph-vector/onnx,faiss"
 ```
 
 Notes
@@ -154,9 +154,9 @@ sudo dnf install cmake clang openssl-devel
 ### Optional Dependencies
 
 - **FAISS** (for vector search acceleration)
-- **Local Embeddings (Hugging Face + Candle + ONNX osx-metal/cuda/cpu)**
+- **Local Embeddings (HuggingFace + Candle + ONNX/ORT(coreML) osx-metal/cuda/cpu)**
   - Enables on-device embedding generation (no external API calls)
-  - Downloads models from Hugging Face Hub on first run and caches them locally
+  - Downloads models from HuggingFace Hub on first run and caches them locally
   - Internet access required for the initial model download (or pre-populate cache)
   - Default runs on CPU; advanced GPU backends (CUDA/Metal) require appropriate hardware and drivers
 - **CUDA** (for GPU-accelerated embeddings)
@@ -260,7 +260,7 @@ codegraph index .
 codegraph index . --languages rust,python,typescript
 
 # Or with more options in Osx
-RUST_LOG=info,codegraph_vector=debug codegraph index . --workers 10 --batch-size 256 --device metal --max-seq-len 512 --force                                                    
+RUST_LOG=info,codegraph_vector=debug codegraph index . --workers 10 --batch-size 256 --max-seq-len 512 --force                                                    
 
 # Index with file watching
 codegraph index . --watch

SESSION-MEMORY.md

@@ -1,12 +1,12 @@
 use codegraph_core::{CodeGraphError, CodeNode, Result};
-#[cfg(any(feature = "local-embeddings", feature = "openai"))]
+#[cfg(any(feature = "local-embeddings", feature = "openai", feature = "onnx"))]
 use std::sync::Arc;
-#[cfg(any(feature = "local-embeddings", feature = "openai"))]
+#[cfg(any(feature = "local-embeddings", feature = "openai", feature = "onnx"))]
 use crate::embeddings::generator::TextEmbeddingEngine;
 
 pub struct EmbeddingGenerator {
     model_config: ModelConfig,
-    #[cfg(any(feature = "local-embeddings", feature = "openai"))]
+    #[cfg(any(feature = "local-embeddings", feature = "openai", feature = "onnx"))]
     pub(crate) advanced: Option<Arc<crate::embeddings::generator::AdvancedEmbeddingGenerator>>,
 }
 
@@ -31,7 +31,7 @@ impl EmbeddingGenerator {
     pub fn new(config: ModelConfig) -> Self {
         Self {
             model_config: config,
-            #[cfg(any(feature = "local-embeddings", feature = "openai"))]
+            #[cfg(any(feature = "local-embeddings", feature = "openai", feature = "onnx"))]
             advanced: None,
         }
     }
@@ -40,7 +40,7 @@ impl EmbeddingGenerator {
         Self::new(ModelConfig::default())
     }
 
-    #[cfg(any(feature = "local-embeddings", feature = "openai"))]
+    #[cfg(any(feature = "local-embeddings", feature = "openai", feature = "onnx"))]
     pub fn set_advanced_engine(
         &mut self,
         engine: Arc<crate::embeddings::generator::AdvancedEmbeddingGenerator>,
@@ -55,15 +55,15 @@ impl EmbeddingGenerator {
     /// Construct an EmbeddingGenerator that optionally wraps the advanced engine based on env.
     /// If CODEGRAPH_EMBEDDING_PROVIDER=local, tries to initialize a local-first engine.
     pub async fn with_auto_from_env() -> Self {
-        #[cfg(any(feature = "local-embeddings", feature = "openai"))]
+        #[cfg(any(feature = "local-embeddings", feature = "openai", feature = "onnx"))]
         let mut base = Self::new(ModelConfig::default());
-        #[cfg(not(any(feature = "local-embeddings", feature = "openai")))]
+        #[cfg(not(any(feature = "local-embeddings", feature = "openai", feature = "onnx")))]
         let base = Self::new(ModelConfig::default());
         let provider = std::env::var("CODEGRAPH_EMBEDDING_PROVIDER")
             .unwrap_or_default()
             .to_lowercase();
         if provider == "local" {
-            #[cfg(any(feature = "local-embeddings", feature = "openai"))]
+            #[cfg(any(feature = "local-embeddings", feature = "openai", feature = "onnx"))]
             {
                 use crate::embeddings::generator::{
                     AdvancedEmbeddingGenerator, EmbeddingEngineConfig, LocalDeviceTypeCompat,
@@ -85,6 +85,22 @@ impl EmbeddingGenerator {
                     base.advanced = Some(Arc::new(engine));
                 }
             }
+        } else if provider == "onnx" {
+            #[cfg(feature = "onnx")]
+            {
+                use crate::embeddings::generator::{AdvancedEmbeddingGenerator, EmbeddingEngineConfig, OnnxConfigCompat};
+                let mut cfg = EmbeddingEngineConfig::default();
+                let model_repo = std::env::var("CODEGRAPH_LOCAL_MODEL").unwrap_or_default();
+                cfg.onnx = Some(OnnxConfigCompat {
+                    model_repo,
+                    model_file: Some("model.onnx".into()),
+                    max_sequence_length: 512,
+                    pooling: "mean".into(),
+                });
+                if let Ok(engine) = AdvancedEmbeddingGenerator::new(cfg).await {
+                    base.advanced = Some(Arc::new(engine));
+                }
+            }
         }
         base
     }
@@ -95,7 +111,7 @@ impl EmbeddingGenerator {
     }
 
     pub async fn generate_embeddings(&self, nodes: &[CodeNode]) -> Result<Vec<Vec<f32>>> {
-        #[cfg(any(feature = "local-embeddings", feature = "openai"))]
+        #[cfg(any(feature = "local-embeddings", feature = "openai", feature = "onnx"))]
         if let Some(engine) = &self.advanced {
             // Use provider's batched path when available
             let texts: Vec<String> = nodes.iter().map(|n| self.prepare_text(n)).collect();