Adds delta logic (#144)

* Adds delta logic

* Sbumodule update

Author: JWittmeyer

controller.py

@@ -44,6 +44,7 @@ def generate_batches(
     attribute_values_raw: List[str],
     embedder: Transformer,
     attribute_name: str,
+    for_delta: bool = False,
 ) -> Iterator[Dict[List[str], List[Any]]]:
     length = len(record_ids)
     record_batches = []
@@ -61,7 +62,10 @@ def generate_batches(
         record_batches.append(record_ids_batch)
         document_batches.extend(documents)
 
-    embedding_batches = embedder.fit_transform(document_batches, as_generator=True)
+    if for_delta:
+        embedding_batches = embedder.transform(document_batches, as_generator=True)
+    else:
+        embedding_batches = embedder.fit_transform(document_batches, as_generator=True)
     for record_batch in record_batches:
         yield {"record_ids": record_batch, "embeddings": next(embedding_batches)}
 
@@ -193,6 +197,13 @@ def run_encoding(
         initial_count = record.count_attribute_list_entries(project_id, attribute_name)
     else:
         initial_count = record.count(project_id)
+
+    is_delta = False
+    # refinery gateway handles delta logic beforehand so if count is 0 we can be sure it's not a delta
+    if tensor_count := embedding.get_tensor_count(embedding_id) != 0:
+        is_delta = True
+        initial_count -= tensor_count
+
     seed_str = embedding_name
     torch.manual_seed(zlib.adler32(bytes(seed_str, "utf-8")))
     notification.create(
@@ -214,15 +225,18 @@ def run_encoding(
         else:
             config_string = model
 
-        embedder = get_embedder(
-            project_id,
-            embedding_type,
-            iso2_code,
-            platform,
-            model,
-            api_token,
-            additional_data,
-        )
+        if is_delta:
+            embedder = __setup_tmp_embedder(project_id, embedding_id)
+        else:
+            embedder = get_embedder(
+                project_id,
+                embedding_type,
+                iso2_code,
+                platform,
+                model,
+                api_token,
+                additional_data,
+            )
 
         if not embedder:
             raise Exception(
@@ -253,7 +267,7 @@ def run_encoding(
 
     try:
         record_ids, attribute_values_raw = record.get_attribute_data(
-            project_id, attribute_name
+            project_id, attribute_name, is_delta, embedding_id
         )
         embedding.update_embedding_state_encoding(
             project_id,
@@ -279,7 +293,8 @@ def run_encoding(
             True,
         )
         send_project_update(project_id, f"notification_created:{user_id}", True)
-        embedding.delete_tensors(embedding_id, with_commit=True)
+        if not is_delta:
+            embedding.delete_tensors(embedding_id, with_commit=True)
         chunk = 0
         embedding_canceled = False
         for pair in generate_batches(
@@ -289,6 +304,7 @@ def run_encoding(
             attribute_values_raw,
             embedder,
             attribute_name,
+            for_delta=is_delta,
         ):
             if chunk % 10 == 0:
                 session_token = general.remove_and_refresh_session(session_token, True)