feat(knowledgebase): replace litellm-pgvector connector with memory-retrieval

Per docs/research/langchain-pgvector-vs-litellm-pgvector.md (issue #25):
the vendored litellm-pgvector connector (793 lines, Prisma migrations, a
fragile git-context build) is replaced by a ~90-line FastAPI service
(services/memory-retrieval/) wrapping langchain_postgres.PGVector directly
against pgvector-db. Same gateway boundary — it still calls litellm for
embeddings, nothing talks to Postgres or the model directly except this
service.

- New services/memory-retrieval/ (main.py, Dockerfile, requirements.txt):
  POST /ingest, POST /query, GET /health.
- docker-compose.yml: litellm-pgvector service replaced by memory-retrieval;
  pgvector-db and embedding-server untouched.
- litellm-config.yaml: vector_store_registry block removed (no
  langchain_postgres provider exists to register against; callers query
  memory-retrieval directly instead of an in-band file_search tool call —
  that mechanism was never confirmed working per issue #24 anyway).
- scripts/ingest-memory.sh rewritten for the new /ingest endpoint (same
  per-line chunking, no dedup).
- .env vars renamed: LITELLM_PGVECTOR_API_KEY/LITELLM_PGVECTOR_EMBEDDING_KEY
  -> MEMORY_RETRIEVAL_API_KEY/MEMORY_RETRIEVAL_EMBEDDING_KEY.
- vendor/litellm-pgvector/ removed entirely.
- docs/memory-knowledgebase.md updated for the new setup/query flow.

Co-Authored-By: Claude Sonnet 5 <noreply@anthropic.com>

services/memory-retrieval/main.py

@@ -0,0 +1,79 @@
+"""Thin FastAPI wrapper around langchain_postgres.PGVector, replacing the
+vendored litellm-pgvector connector. See
+docs/research/langchain-pgvector-vs-litellm-pgvector.md for the rationale
+and docs/memory-knowledgebase.md for usage.
+
+ponytail: one fixed collection (COLLECTION_NAME) — this stack only needs one
+knowledgebase ("memory-and-notes"), not a multi-tenant store registry.
+"""
+import os
+
+from fastapi import Depends, FastAPI, HTTPException
+from fastapi.security import HTTPAuthorizationCredentials, HTTPBearer
+from langchain_openai import OpenAIEmbeddings
+from langchain_postgres import PGVector
+from pydantic import BaseModel
+
+DATABASE_URL = os.environ["DATABASE_URL"]
+LITELLM_BASE_URL = os.environ["LITELLM_BASE_URL"]
+LITELLM_API_KEY = os.environ["LITELLM_API_KEY"]
+EMBEDDING_MODEL = os.environ.get("EMBEDDING_MODEL", "local-embedding")
+SERVER_API_KEY = os.environ["SERVER_API_KEY"]
+COLLECTION_NAME = os.environ.get("COLLECTION_NAME", "memory-and-notes")
+
+app = FastAPI(title="memory-retrieval", version="1.0.0")
+
+embeddings = OpenAIEmbeddings(
+    model=EMBEDDING_MODEL, base_url=LITELLM_BASE_URL, api_key=LITELLM_API_KEY
+)
+vector_store = PGVector(
+    embeddings=embeddings,
+    collection_name=COLLECTION_NAME,
+    connection=DATABASE_URL,
+    use_jsonb=True,
+)
+
+security = HTTPBearer()
+
+
+def check_api_key(credentials: HTTPAuthorizationCredentials = Depends(security)):
+    if credentials.credentials != SERVER_API_KEY:
+        raise HTTPException(status_code=401, detail="Invalid API key")
+
+
+class Chunk(BaseModel):
+    content: str
+    metadata: dict = {}
+
+
+class IngestRequest(BaseModel):
+    chunks: list[Chunk]
+
+
+class QueryRequest(BaseModel):
+    query: str
+    k: int = 5
+
+
+@app.get("/health")
+def health():
+    return {"status": "ok"}
+
+
+@app.post("/ingest", dependencies=[Depends(check_api_key)])
+def ingest(req: IngestRequest):
+    texts = [c.content for c in req.chunks]
+    metadatas = [c.metadata for c in req.chunks]
+    ids = vector_store.add_texts(texts=texts, metadatas=metadatas)
+    return {"ingested": len(ids)}
+
+
+@app.post("/query", dependencies=[Depends(check_api_key)])
+def query(req: QueryRequest):
+    results = vector_store.similarity_search_with_relevance_scores(req.query, k=req.k)
+    return {
+        "results": [
+            {"content": doc.page_content, "metadata": doc.metadata, "score": score}
+            for doc, score in results
+        ]
+    }