Add graph embeddings query

scripts/ge-query-milvus

@@ -0,0 +1,6 @@
+#!/usr/bin/env python3
+
+from trustgraph.query.graph_embeddings.milvus import run
+
+run()
+

setup.py

@@ -57,6 +57,7 @@ setuptools.setup(
         "scripts/embeddings-ollama",
         "scripts/embeddings-vectorize",
         "scripts/ge-dump-parquet",
+        "scripts/ge-query-milvus",
         "scripts/ge-write-milvus",
         "scripts/graph-rag",
         "scripts/graph-show",

trustgraph/query/graph_embeddings/milvus/__init__.py

@@ -0,0 +1,3 @@
+
+from . service import *
+

trustgraph/query/graph_embeddings/milvus/__main__.py

@@ -0,0 +1,7 @@
+#!/usr/bin/env python3
+
+from . hf import run
+
+if __name__ == '__main__':
+    run()
+

trustgraph/query/graph_embeddings/milvus/service.py

@@ -0,0 +1,83 @@
+
+"""
+Graph embeddings query service.  Input is vector, output is list of
+e
+"""
+
+from ... direct.milvus import TripleVectors
+from ... schema import GraphEmbeddingsRequest, GraphEmbeddingsResponse
+from ... schema import graph_embeddings_request_queue
+from ... schema import graph_embeddings_response_queue
+from ... base import ConsumerProducer
+
+module = ".".join(__name__.split(".")[1:-1])
+
+default_input_queue = graph_embeddings_request_queue
+default_output_queue = graph_embeddings_response_queue
+default_subscriber = module
+default_store_uri = 'http://localhost:19530'
+
+class Processor(ConsumerProducer):
+
+    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)
+
+        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,
+            }
+        )
+
+        self.vecstore = TripleVectors(store_uri)
+
+    def handle(self, msg):
+
+        v = msg.value()
+
+        # Sender-produced ID
+        id = msg.properties()["id"]
+
+        print(f"Handling input {id}...", flush=True)
+
+        entities = set()
+
+        for vec in v.vectors:
+
+            resp = self.vecstore.search(vec, limit=v.limit)
+
+            for r in resp:
+                entities.add(r)
+
+        print("Send response...", flush=True)
+        r = GraphEmbeddingsResponse(entities=entities)
+        self.producer.send(r, properties={"id": id})
+
+        print("Done.", flush=True)
+
+    @staticmethod
+    def add_args(parser):
+
+        ConsumerProducer.add_args(
+            parser, default_input_queue, default_subscriber,
+            default_output_queue,
+        )
+
+        parser.add_argument(
+            '-t', '--store-uri',
+            default=default_store_uri,
+            help=f'Milvus store URI (default: {default_store_uri})'
+        )
+
+def run():
+
+    Processor.start(module, __doc__)
+

trustgraph/schema.py

@@ -75,9 +75,10 @@ graph_embeddings_store_queue = topic('graph-embeddings-store')
 
 class GraphEmbeddingsQueryRequest(Record):
     vectors = Array(Array(Double()))
+    limit = Integer()
 
 class GraphEmbeddingsQueryResponse(Record):
-    vectors = Array(Value())
+    entities = Array(Value())
 
 graph_embeddings_request_queue = topic(
     'graph-embeddings', kind='non-persistent', namespace='request'

trustgraph/storage/graph_embeddings/milvus/write.py

@@ -51,8 +51,8 @@ class Processor(Consumer):
 
         parser.add_argument(
             '-t', '--store-uri',
-            default="http://milvus:19530",
-            help=f'Milvus store URI (default: http://milvus:19530)'
+            default=default_store_uri,
+            help=f'Milvus store URI (default: {default_store_uri})'
         )
 
 def run():