Pre-pgvector migration checkpoint — upsert, allow_replace_deleted, maintenance timer

scripts/api.py

@@ -62,7 +62,7 @@ embedder = SentenceTransformer("all-MiniLM-L6-v2")
 chroma_client = chromadb.PersistentClient(path=DB_PATH)
 collection = chroma_client.get_or_create_collection(
     name="aaronai",
-    metadata={"hnsw:space": "cosine"}
+    metadata={"hnsw:space": "cosine", "hnsw:allow_replace_deleted": True}
 )
 anthropic_client = anthropic.Anthropic(api_key=os.getenv("ANTHROPIC_API_KEY"))
 

scripts/dream.py

@@ -122,7 +122,7 @@ def retrieve(mode, task=None, project=None, n_results=8):
     client     = chromadb.PersistentClient(path=DB_PATH)
     collection = client.get_or_create_collection(
         name="aaronai",
-        metadata={"hnsw:space": "cosine"}
+        metadata={"hnsw:space": "cosine", "hnsw:allow_replace_deleted": True}
     )
 
     low, high = MODE_RANGES[mode]

scripts/ingest.py

@@ -18,7 +18,7 @@ db_path = str(Path.home() / "aaronai" / "db")
 client = chromadb.PersistentClient(path=db_path)
 collection = client.get_or_create_collection(
     name="aaronai",
-    metadata={"hnsw:space": "cosine"}
+    metadata={"hnsw:space": "cosine", "hnsw:allow_replace_deleted": True}
 )
 
 def extract_text_from_docx(path):

scripts/ingest_chatgpt.py

@@ -14,7 +14,7 @@ embedder = SentenceTransformer("all-MiniLM-L6-v2")
 client = chromadb.PersistentClient(path=db_path)
 collection = client.get_or_create_collection(
     name="aaronai",
-    metadata={"hnsw:space": "cosine"}
+    metadata={"hnsw:space": "cosine", "hnsw:allow_replace_deleted": True}
 )
 
 def extract_messages(convo):

scripts/ingest_claude.py

@@ -13,7 +13,7 @@ embedder = SentenceTransformer("all-MiniLM-L6-v2")
 client = chromadb.PersistentClient(path=db_path)
 collection = client.get_or_create_collection(
     name="aaronai",
-    metadata={"hnsw:space": "cosine"}
+    metadata={"hnsw:space": "cosine", "hnsw:allow_replace_deleted": True}
 )
 
 def extract_messages(convo):
@@ -102,7 +102,7 @@ def ingest_conversations(path):
             continue
 
         embeddings = embedder.encode([n[1] for n in new]).tolist()
-        collection.add(
+        collection.upsert(
             ids=[n[0] for n in new],
             documents=[n[1] for n in new],
             metadatas=[n[2] for n in new],
@@ -138,7 +138,7 @@ def ingest_memories(path):
         collection.delete(ids=[chunk_id])
 
     embedding = embedder.encode([chunk_text]).tolist()
-    collection.add(
+    collection.upsert(
         ids=[chunk_id],
         documents=[chunk_text],
         metadatas=[{