Fix/document embeddings (#247)

* Update schema for doc embeddings

* Rename embeddings-vectorize to graph-embeddings

* Added document-embeddings processor (broken, needs fixing)

* Added scripts

* Fixed DE queue schema

* Add missing DE process

* Fix doc RAG processing, put graph-rag and doc-rag in appropriate component files.

trustgraph-flow/trustgraph/embeddings/document_embeddings/__init__.py

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

trustgraph-flow/trustgraph/embeddings/document_embeddings/__main__.py

@@ -1,5 +1,5 @@
 
-from . vectorize import run
+from . embeddings import run
 
 if __name__ == '__main__':
     run()

trustgraph-flow/trustgraph/embeddings/document_embeddings/embeddings.py

@@ -0,0 +1,109 @@
+
+"""
+Document embeddings, calls the embeddings service to get embeddings for a
+chunk of text.  Input is chunk of text plus metadata.
+Output is chunk plus embedding.
+"""
+
+from ... schema import Chunk, ChunkEmbeddings, DocumentEmbeddings
+from ... schema import chunk_ingest_queue
+from ... schema import document_embeddings_store_queue
+from ... schema import embeddings_request_queue, embeddings_response_queue
+from ... clients.embeddings_client import EmbeddingsClient
+from ... log_level import LogLevel
+from ... base import ConsumerProducer
+
+module = ".".join(__name__.split(".")[1:-1])
+
+default_input_queue = chunk_ingest_queue
+default_output_queue = document_embeddings_store_queue
+default_subscriber = module
+
+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)
+        emb_request_queue = params.get(
+            "embeddings_request_queue", embeddings_request_queue
+        )
+        emb_response_queue = params.get(
+            "embeddings_response_queue", embeddings_response_queue
+        )
+
+        super(Processor, self).__init__(
+            **params | {
+                "input_queue": input_queue,
+                "output_queue": output_queue,
+                "embeddings_request_queue": emb_request_queue,
+                "embeddings_response_queue": emb_response_queue,
+                "subscriber": subscriber,
+                "input_schema": Chunk,
+                "output_schema": DocumentEmbeddings,
+            }
+        )
+
+        self.embeddings = EmbeddingsClient(
+            pulsar_host=self.pulsar_host,
+            input_queue=emb_request_queue,
+            output_queue=emb_response_queue,
+            subscriber=module + "-emb",
+        )
+
+    def handle(self, msg):
+
+        v = msg.value()
+        print(f"Indexing {v.metadata.id}...", flush=True)
+
+        try:
+
+            vectors = self.embeddings.request(v.chunk)
+
+            embeds = [
+                ChunkEmbeddings(
+                    chunk=v.chunk,
+                    vectors=vectors,
+                )
+            ]
+
+            r = DocumentEmbeddings(
+                metadata=v.metadata,
+                chunks=embeds,
+            )
+
+            self.producer.send(r)
+
+        except Exception as e:
+            print("Exception:", e, flush=True)
+
+            # Retry
+            raise e
+
+        print("Done.", flush=True)
+
+    @staticmethod
+    def add_args(parser):
+
+        ConsumerProducer.add_args(
+            parser, default_input_queue, default_subscriber,
+            default_output_queue,
+        )
+
+        parser.add_argument(
+            '--embeddings-request-queue',
+            default=embeddings_request_queue,
+            help=f'Embeddings request queue (default: {embeddings_request_queue})',
+        )
+
+        parser.add_argument(
+            '--embeddings-response-queue',
+            default=embeddings_response_queue,
+            help=f'Embeddings request queue (default: {embeddings_response_queue})',
+        )
+
+def run():
+
+    Processor.start(module, __doc__)
+

trustgraph-flow/trustgraph/embeddings/graph_embeddings/__init__.py

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

trustgraph-flow/trustgraph/embeddings/graph_embeddings/__main__.py

@@ -0,0 +1,6 @@
+
+from . embeddings import run
+
+if __name__ == '__main__':
+    run()
+

trustgraph-flow/trustgraph/embeddings/graph_embeddings/embeddings.py

@@ -1,7 +1,8 @@
 
 """
-Vectorizer, calls the embeddings service to get embeddings for a chunk.
-Input is text chunk, output is chunk and vectors.
+Graph embeddings, calls the embeddings service to get embeddings for a
+set of entity contexts.  Input is entity plus textual context.
+Output is entity plus embedding.
 """
 
 from ... schema import EntityContexts, EntityEmbeddings, GraphEmbeddings
@@ -51,11 +52,6 @@ class Processor(ConsumerProducer):
             subscriber=module + "-emb",
         )
 
-    def emit(self, rec, vectors):
-
-        r = GraphEmbeddings(metadata=metadata, chunk=chunk, vectors=vectors)
-        self.producer.send(r)
-
     def handle(self, msg):
 
         v = msg.value()

trustgraph-flow/trustgraph/embeddings/vectorize/__init__.py

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