From a67863563acb2a1aa050b0e13df23379673df900 Mon Sep 17 00:00:00 2001
From: zm2231 <merchantzains@gmail.com>
Date: Fri, 20 Mar 2026 00:31:39 -0400
Subject: [PATCH] fix: dynamic embedding dimensions for HTTP backend

- schema.ts: FLOAT[${EMBEDDING_DIMS}] reads from GITNEXUS_EMBEDDING_DIMS env
- embedding-pipeline.ts: vector search CAST uses actual query vector length
- mcp/core/embedder.ts: HTTP embedding support for MCP query-time search

Without this, using a 1024d model (e.g. bge-large) fails with
'Expected: 384, Actual: 1024' on LadybugDB vector insert.
---
 .../src/core/embeddings/embedding-pipeline.ts |  2 +-
 gitnexus/src/core/lbug/schema.ts              |  4 ++-
 gitnexus/src/mcp/core/embedder.ts             | 26 +++++++++++++++++--
 3 files changed, 28 insertions(+), 4 deletions(-)

diff --git a/gitnexus/src/core/embeddings/embedding-pipeline.ts b/gitnexus/src/core/embeddings/embedding-pipeline.ts
index 5fbd5cd0b..bee01ed5f 100644
--- a/gitnexus/src/core/embeddings/embedding-pipeline.ts
+++ b/gitnexus/src/core/embeddings/embedding-pipeline.ts
@@ -326,7 +326,7 @@ export const semanticSearch = async (
   // Query the vector index on CodeEmbedding to get nodeIds and distances
   const vectorQuery = `
     CALL QUERY_VECTOR_INDEX('CodeEmbedding', 'code_embedding_idx', 
-      CAST(${queryVecStr} AS FLOAT[384]), ${k})
+      CAST(${queryVecStr} AS FLOAT[${queryVec.length}]), ${k})
     YIELD node AS emb, distance
     WITH emb, distance
     WHERE distance < ${maxDistance}
diff --git a/gitnexus/src/core/lbug/schema.ts b/gitnexus/src/core/lbug/schema.ts
index ef1fbad50..689fb4df0 100644
--- a/gitnexus/src/core/lbug/schema.ts
+++ b/gitnexus/src/core/lbug/schema.ts
@@ -398,10 +398,12 @@ CREATE REL TABLE ${REL_TABLE_NAME} (
 // Separate table for vector storage to avoid copy-on-write overhead
 // ============================================================================
 
+export const EMBEDDING_DIMS = parseInt(process.env.GITNEXUS_EMBEDDING_DIMS ?? '384', 10);
+
 export const EMBEDDING_SCHEMA = `
 CREATE NODE TABLE ${EMBEDDING_TABLE_NAME} (
   nodeId STRING,
-  embedding FLOAT[384],
+  embedding FLOAT[${EMBEDDING_DIMS}],
   PRIMARY KEY (nodeId)
 )`;
 
diff --git a/gitnexus/src/mcp/core/embedder.ts b/gitnexus/src/mcp/core/embedder.ts
index ee480a6a9..404e6dd88 100644
--- a/gitnexus/src/mcp/core/embedder.ts
+++ b/gitnexus/src/mcp/core/embedder.ts
@@ -9,7 +9,13 @@ import { pipeline, env, type FeatureExtractionPipeline } from '@huggingface/tran
 
 // Model config
 const MODEL_ID = 'Snowflake/snowflake-arctic-embed-xs';
-const EMBEDDING_DIMS = 384;
+const EMBEDDING_DIMS = parseInt(process.env.GITNEXUS_EMBEDDING_DIMS ?? '384', 10);
+
+// HTTP embedding config
+const HTTP_URL = process.env.GITNEXUS_EMBEDDING_URL ?? '';
+const HTTP_MODEL = process.env.GITNEXUS_EMBEDDING_MODEL ?? '';
+const HTTP_KEY = process.env.GITNEXUS_EMBEDDING_API_KEY ?? 'unused';
+const USE_HTTP = !!(HTTP_URL && HTTP_MODEL);
 
 // Module-level state for singleton pattern
 let embedderInstance: FeatureExtractionPipeline | null = null;
@@ -20,6 +26,8 @@ let initPromise: Promise<FeatureExtractionPipeline> | null = null;
  * Initialize the embedding model (lazy, on first search)
  */
 export const initEmbedder = async (): Promise<FeatureExtractionPipeline> => {
+  if (USE_HTTP) return null as unknown as FeatureExtractionPipeline;
+
   if (embedderInstance) {
     return embedderInstance;
   }
@@ -87,12 +95,26 @@ export const initEmbedder = async (): Promise<FeatureExtractionPipeline> => {
 /**
  * Check if embedder is ready
  */
-export const isEmbedderReady = (): boolean => embedderInstance !== null;
+export const isEmbedderReady = (): boolean => USE_HTTP || embedderInstance !== null;
 
 /**
  * Embed a query text for semantic search
  */
 export const embedQuery = async (query: string): Promise<number[]> => {
+  if (USE_HTTP) {
+    const resp = await fetch(`${HTTP_URL.replace(/\/+$/, '')}/embeddings`, {
+      method: 'POST',
+      headers: {
+        'Content-Type': 'application/json',
+        'Authorization': `Bearer ${HTTP_KEY}`,
+      },
+      body: JSON.stringify({ input: [query], model: HTTP_MODEL }),
+    });
+    if (!resp.ok) throw new Error(`Embedding endpoint ${resp.status}`);
+    const data = (await resp.json()) as { data: Array<{ embedding: number[] }> };
+    return data.data[0].embedding;
+  }
+
   const embedder = await initEmbedder();
   
   const result = await embedder(query, {