Feature/configure flows (#345)

- Keeps processing in different flows separate so that data can go to different stores / collections etc.
- Potentially supports different processing flows
- Tidies the processing API with common base-classes for e.g. LLMs, and automatic configuration of 'clients' to use the right queue names in a flow

trustgraph-flow/trustgraph/query/graph_embeddings/milvus/service.py

@@ -11,7 +11,7 @@ from .... schema import graph_embeddings_request_queue
 from .... schema import graph_embeddings_response_queue
 from .... base import ConsumerProducer
 
-module = ".".join(__name__.split(".")[1:-1])
+module = "ge-query"
 
 default_input_queue = graph_embeddings_request_queue
 default_output_queue = graph_embeddings_response_queue

trustgraph-flow/trustgraph/query/graph_embeddings/pinecone/service.py

@@ -16,7 +16,7 @@ from .... schema import graph_embeddings_request_queue
 from .... schema import graph_embeddings_response_queue
 from .... base import ConsumerProducer
 
-module = ".".join(__name__.split(".")[1:-1])
+module = "ge-query"
 
 default_input_queue = graph_embeddings_request_queue
 default_output_queue = graph_embeddings_response_queue

trustgraph-flow/trustgraph/query/graph_embeddings/qdrant/service.py

@@ -7,44 +7,32 @@ entities
 from qdrant_client import QdrantClient
 from qdrant_client.models import PointStruct
 from qdrant_client.models import Distance, VectorParams
-import uuid
 
-from .... schema import GraphEmbeddingsRequest, GraphEmbeddingsResponse
+from .... schema import GraphEmbeddingsResponse
 from .... schema import Error, Value
-from .... schema import graph_embeddings_request_queue
-from .... schema import graph_embeddings_response_queue
-from .... base import ConsumerProducer
+from .... base import GraphEmbeddingsQueryService
 
-module = ".".join(__name__.split(".")[1:-1])
+default_ident = "ge-query"
 
-default_input_queue = graph_embeddings_request_queue
-default_output_queue = graph_embeddings_response_queue
-default_subscriber = module
 default_store_uri = 'http://localhost:6333'
 
-class Processor(ConsumerProducer):
+class Processor(GraphEmbeddingsQueryService):
 
     def __init__(self, **params):
 
-        input_queue = params.get("input_queue", default_input_queue)
-        output_queue = params.get("output_queue", default_output_queue)
-        subscriber = params.get("subscriber", default_subscriber)
         store_uri = params.get("store_uri", default_store_uri)
+
+        #optional api key
         api_key = params.get("api_key", None)
 
         super(Processor, self).__init__(
             **params | {
-                "input_queue": input_queue,
-                "output_queue": output_queue,
-                "subscriber": subscriber,
-                "input_schema": GraphEmbeddingsRequest,
-                "output_schema": GraphEmbeddingsResponse,
                 "store_uri": store_uri,
                 "api_key": api_key,
             }
         )
 
-        self.client = QdrantClient(url=store_uri, api_key=api_key)
+        self.qdrant = QdrantClient(url=store_uri, api_key=api_key)
 
     def create_value(self, ent):
         if ent.startswith("http://") or ent.startswith("https://"):
@@ -52,34 +40,27 @@ class Processor(ConsumerProducer):
         else:
             return Value(value=ent, is_uri=False)
         
-    async def handle(self, msg):
+    async def query_graph_embeddings(self, msg):
 
         try:
 
-            v = msg.value()
-
-            # Sender-produced ID
-            id = msg.properties()["id"]
-
-            print(f"Handling input {id}...", flush=True)
-
             entity_set = set()
             entities = []
 
-            for vec in v.vectors:
+            for vec in msg.vectors:
 
                 dim = len(vec)
                 collection = (
-                    "t_" + v.user + "_" + v.collection + "_" +
+                    "t_" + msg.user + "_" + msg.collection + "_" +
                     str(dim)
                 )
 
                 # Heuristic hack, get (2*limit), so that we have more chance
                 # of getting (limit) entities
-                search_result = self.client.query_points(
+                search_result = self.qdrant.query_points(
                     collection_name=collection,
                     query=vec,
-                    limit=v.limit * 2,
+                    limit=msg.limit * 2,
                     with_payload=True,
                 ).points
 
@@ -92,10 +73,10 @@ class Processor(ConsumerProducer):
                         entities.append(ent)
 
                     # Keep adding entities until limit
-                    if len(entity_set) >= v.limit: break
+                    if len(entity_set) >= msg.limit: break
 
                 # Keep adding entities until limit
-                if len(entity_set) >= v.limit: break
+                if len(entity_set) >= msg.limit: break
 
             ents2 = []
 
@@ -105,36 +86,19 @@ class Processor(ConsumerProducer):
             entities = ents2
 
             print("Send response...", flush=True)
-            r = GraphEmbeddingsResponse(entities=entities, error=None)
-            await self.send(r, properties={"id": id})
+            return entities
 
             print("Done.", flush=True)
 
         except Exception as e:
 
             print(f"Exception: {e}")
-
-            print("Send error response...", flush=True)
-
-            r = GraphEmbeddingsResponse(
-                error=Error(
-                    type = "llm-error",
-                    message = str(e),
-                ),
-                entities=None,
-            )
-
-            await self.send(r, properties={"id": id})
-
-            self.consumer.acknowledge(msg)
+            raise e
 
     @staticmethod
     def add_args(parser):
 
-        ConsumerProducer.add_args(
-            parser, default_input_queue, default_subscriber,
-            default_output_queue,
-        )
+        GraphEmbeddingsQueryService.add_args(parser)
 
         parser.add_argument(
             '-t', '--store-uri',
@@ -150,5 +114,5 @@ class Processor(ConsumerProducer):
 
 def run():
 
-    Processor.launch(module, __doc__)
+    Processor.launch(default_ident, __doc__)