trustgraph-flow 0.16.11__tar.gz → 0.17.0__tar.gz

{trustgraph-flow-0.16.11 → trustgraph-flow-0.17.0}/PKG-INFO

@@ -1,12 +1,12 @@
 Metadata-Version: 2.1
 Name: trustgraph-flow
-Version: 0.16.11
+Version: 0.17.0
 Summary: TrustGraph provides a means to run a pipeline of flexible AI processing components in a flexible means to achieve a processing pipeline.
 Home-page: https://github.com/trustgraph-ai/trustgraph
 Author: trustgraph.ai
 Author-email: security@trustgraph.ai
 License: UNKNOWN
-Download-URL: https://github.com/trustgraph-ai/trustgraph/archive/refs/tags/v0.16.11.tar.gz
+Download-URL: https://github.com/trustgraph-ai/trustgraph/archive/refs/tags/v0.17.0.tar.gz
 Platform: UNKNOWN
 Classifier: Programming Language :: Python :: 3
 Classifier: License :: OSI Approved :: GNU General Public License v3 or later (GPLv3+)

trustgraph-flow-0.17.0/scripts/wikipedia-lookup

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

{trustgraph-flow-0.16.11 → trustgraph-flow-0.17.0}/setup.py

@@ -34,7 +34,7 @@ setuptools.setup(
     python_requires='>=3.8',
     download_url = "https://github.com/trustgraph-ai/trustgraph/archive/refs/tags/v" + version + ".tar.gz",
     install_requires=[
-        "trustgraph-base>=0.16,<0.17",
+        "trustgraph-base>=0.17,<0.18",
         "urllib3",
         "rdflib",
         "pymilvus",
@@ -106,5 +106,6 @@ setuptools.setup(
         "scripts/triples-query-neo4j",
         "scripts/triples-write-cassandra",
         "scripts/triples-write-neo4j",
+        "scripts/wikipedia-lookup",
     ]
 )

trustgraph-flow-0.17.0/trustgraph/api/gateway/agent.py

@@ -0,0 +1,30 @@
+
+from ... schema import AgentRequest, AgentResponse
+from ... schema import agent_request_queue
+from ... schema import agent_response_queue
+
+from . endpoint import MultiResponseServiceEndpoint
+
+class AgentEndpoint(MultiResponseServiceEndpoint):
+    def __init__(self, pulsar_host, timeout):
+
+        super(AgentEndpoint, self).__init__(
+            pulsar_host=pulsar_host,
+            request_queue=agent_request_queue,
+            response_queue=agent_response_queue,
+            request_schema=AgentRequest,
+            response_schema=AgentResponse,
+            endpoint_path="/api/v1/agent",
+            timeout=timeout,
+        )
+
+    def to_request(self, body):
+        return AgentRequest(
+            question=body["question"]
+        )
+
+    def from_response(self, message):
+        if message.answer:
+            return { "answer": message.answer }, True
+        else:
+            return {}, False

trustgraph-flow-0.17.0/trustgraph/api/gateway/dbpedia.py

@@ -0,0 +1,29 @@
+
+from ... schema import LookupRequest, LookupResponse
+from ... schema import dbpedia_lookup_request_queue
+from ... schema import dbpedia_lookup_response_queue
+
+from . endpoint import ServiceEndpoint
+
+class DbpediaEndpoint(ServiceEndpoint):
+    def __init__(self, pulsar_host, timeout):
+
+        super(DbpediaEndpoint, self).__init__(
+            pulsar_host=pulsar_host,
+            request_queue=dbpedia_lookup_request_queue,
+            response_queue=dbpedia_lookup_response_queue,
+            request_schema=LookupRequest,
+            response_schema=LookupResponse,
+            endpoint_path="/api/v1/dbpedia",
+            timeout=timeout,
+        )
+
+    def to_request(self, body):
+        return LookupRequest(
+            term=body["term"],
+            kind=body.get("kind", None),
+        )
+
+    def from_response(self, message):
+        return { "text": message.text }
+

trustgraph-flow-0.17.0/trustgraph/api/gateway/embeddings.py

@@ -0,0 +1,27 @@
+
+from ... schema import EmbeddingsRequest, EmbeddingsResponse
+from ... schema import embeddings_request_queue
+from ... schema import embeddings_response_queue
+
+from . endpoint import ServiceEndpoint
+
+class EmbeddingsEndpoint(ServiceEndpoint):
+    def __init__(self, pulsar_host, timeout):
+
+        super(EmbeddingsEndpoint, self).__init__(
+            pulsar_host=pulsar_host,
+            request_queue=embeddings_request_queue,
+            response_queue=embeddings_response_queue,
+            request_schema=EmbeddingsRequest,
+            response_schema=EmbeddingsResponse,
+            endpoint_path="/api/v1/embeddings",
+            timeout=timeout,
+        )
+
+    def to_request(self, body):
+        return EmbeddingsRequest(
+            text=body["text"]
+        )
+
+    def from_response(self, message):
+        return { "vectors": message.vectors }

trustgraph-flow-0.17.0/trustgraph/api/gateway/encyclopedia.py

@@ -0,0 +1,29 @@
+
+from ... schema import LookupRequest, LookupResponse
+from ... schema import encyclopedia_lookup_request_queue
+from ... schema import encyclopedia_lookup_response_queue
+
+from . endpoint import ServiceEndpoint
+
+class EncyclopediaEndpoint(ServiceEndpoint):
+    def __init__(self, pulsar_host, timeout):
+
+        super(EncyclopediaEndpoint, self).__init__(
+            pulsar_host=pulsar_host,
+            request_queue=encyclopedia_lookup_request_queue,
+            response_queue=encyclopedia_lookup_response_queue,
+            request_schema=LookupRequest,
+            response_schema=LookupResponse,
+            endpoint_path="/api/v1/encyclopedia",
+            timeout=timeout,
+        )
+
+    def to_request(self, body):
+        return LookupRequest(
+            term=body["term"],
+            kind=body.get("kind", None),
+        )
+
+    def from_response(self, message):
+        return { "text": message.text }
+

trustgraph-flow-0.17.0/trustgraph/api/gateway/endpoint.py

@@ -0,0 +1,153 @@
+
+import asyncio
+from pulsar.schema import JsonSchema
+from aiohttp import web
+import uuid
+import logging
+
+from . publisher import Publisher
+from . subscriber import Subscriber
+
+logger = logging.getLogger("endpoint")
+logger.setLevel(logging.INFO)
+
+class ServiceEndpoint:
+
+    def __init__(
+            self,
+            pulsar_host,
+            request_queue, request_schema,
+            response_queue, response_schema,
+            endpoint_path,
+            subscription="api-gateway", consumer_name="api-gateway",
+            timeout=600,
+    ):
+
+        self.pub = Publisher(
+            pulsar_host, request_queue,
+            schema=JsonSchema(request_schema)
+        )
+
+        self.sub = Subscriber(
+            pulsar_host, response_queue,
+            subscription, consumer_name,
+            JsonSchema(response_schema)
+        )
+
+        self.path = endpoint_path
+        self.timeout = timeout
+
+    async def start(self):
+
+        self.pub_task = asyncio.create_task(self.pub.run())
+        self.sub_task = asyncio.create_task(self.sub.run())
+
+    def add_routes(self, app):
+
+        app.add_routes([
+            web.post(self.path, self.handle),
+        ])
+
+    def to_request(self, request):
+        raise RuntimeError("Not defined")
+
+    def from_response(self, response):
+        raise RuntimeError("Not defined")
+
+    async def handle(self, request):
+
+        id = str(uuid.uuid4())
+
+        try:
+
+            data = await request.json()
+
+            q = await self.sub.subscribe(id)
+
+            print(data)
+
+            await self.pub.send(
+                id,
+                self.to_request(data),
+            )
+
+            try:
+                resp = await asyncio.wait_for(q.get(), self.timeout)
+            except:
+                raise RuntimeError("Timeout waiting for response")
+
+            print(resp)
+
+            if resp.error:
+                return web.json_response(
+                    { "error": resp.error.message }
+                )
+
+            return web.json_response(
+                self.from_response(resp)
+            )
+
+        except Exception as e:
+            logging.error(f"Exception: {e}")
+
+            return web.json_response(
+                { "error": str(e) }
+            )
+
+        finally:
+            await self.sub.unsubscribe(id)
+
+
+class MultiResponseServiceEndpoint(ServiceEndpoint):
+
+    async def handle(self, request):
+
+        id = str(uuid.uuid4())
+
+        try:
+
+            data = await request.json()
+
+            q = await self.sub.subscribe(id)
+
+            print(data)
+
+            await self.pub.send(
+                id,
+                self.to_request(data),
+            )
+
+            # Keeps looking at responses...
+
+            while True:
+
+                try:
+                    resp = await asyncio.wait_for(q.get(), self.timeout)
+                except:
+                    raise RuntimeError("Timeout waiting for response")
+
+                print(resp)
+
+                if resp.error:
+                    return web.json_response(
+                        { "error": resp.error.message }
+                    )
+
+                # Until from_response says we have a finished answer
+                resp, fin = self.from_response(resp)
+
+
+                if fin:
+                    return web.json_response(resp)
+
+                # Not finished, so loop round and continue
+
+        except Exception as e:
+            logging.error(f"Exception: {e}")
+
+            return web.json_response(
+                { "error": str(e) }
+            )
+
+        finally:
+            await self.sub.unsubscribe(id)

trustgraph-flow-0.17.0/trustgraph/api/gateway/graph_embeddings_load.py

@@ -0,0 +1,60 @@
+
+import asyncio
+from pulsar.schema import JsonSchema
+import uuid
+from aiohttp import WSMsgType
+
+from ... schema import Metadata
+from ... schema import GraphEmbeddings
+from ... schema import graph_embeddings_store_queue
+
+from . publisher import Publisher
+from . socket import SocketEndpoint
+from . serialize import to_subgraph, to_value
+
+class GraphEmbeddingsLoadEndpoint(SocketEndpoint):
+
+    def __init__(self, pulsar_host, path="/api/v1/load/graph-embeddings"):
+
+        super(GraphEmbeddingsLoadEndpoint, self).__init__(
+            endpoint_path=path
+        )
+
+        self.pulsar_host=pulsar_host
+
+        self.publisher = Publisher(
+            self.pulsar_host, graph_embeddings_store_queue,
+            schema=JsonSchema(GraphEmbeddings)
+        )
+
+    async def start(self):
+
+        self.task = asyncio.create_task(
+            self.publisher.run()
+        )
+
+    async def listener(self, ws, running):
+        
+        async for msg in ws:
+            # On error, finish
+            if msg.type == WSMsgType.ERROR:
+                break
+            else:
+
+                data = msg.json()
+
+                elt = GraphEmbeddings(
+                    metadata=Metadata(
+                        id=data["metadata"]["id"],
+                        metadata=to_subgraph(data["metadata"]["metadata"]),
+                        user=data["metadata"]["user"],
+                        collection=data["metadata"]["collection"],
+                    ),
+                    entity=to_value(data["entity"]),
+                    vectors=data["vectors"],
+                )
+
+                await self.publisher.send(None, elt)
+
+
+        running.stop()

trustgraph-flow-0.17.0/trustgraph/api/gateway/graph_embeddings_stream.py

@@ -0,0 +1,56 @@
+
+import asyncio
+from pulsar.schema import JsonSchema
+import uuid
+
+from ... schema import GraphEmbeddings
+from ... schema import graph_embeddings_store_queue
+
+from . subscriber import Subscriber
+from . socket import SocketEndpoint
+from . serialize import serialize_graph_embeddings
+
+class GraphEmbeddingsStreamEndpoint(SocketEndpoint):
+
+    def __init__(self, pulsar_host, path="/api/v1/stream/graph-embeddings"):
+
+        super(GraphEmbeddingsStreamEndpoint, self).__init__(
+            endpoint_path=path
+        )
+
+        self.pulsar_host=pulsar_host
+
+        self.subscriber = Subscriber(
+            self.pulsar_host, graph_embeddings_store_queue,
+            "api-gateway", "api-gateway",
+            schema=JsonSchema(GraphEmbeddings)
+        )
+
+    async def start(self):
+
+        self.task = asyncio.create_task(
+            self.subscriber.run()
+        )
+
+    async def async_thread(self, ws, running):
+
+        id = str(uuid.uuid4())
+
+        q = await self.subscriber.subscribe_all(id)
+
+        while running.get():
+            try:
+                resp = await asyncio.wait_for(q.get(), 0.5)
+                await ws.send_json(serialize_graph_embeddings(resp))
+
+            except TimeoutError:
+                continue
+
+            except Exception as e:
+                print(f"Exception: {str(e)}", flush=True)
+                break
+
+        await self.subscriber.unsubscribe_all(id)
+
+        running.stop()
+

trustgraph-flow-0.17.0/trustgraph/api/gateway/graph_rag.py

@@ -0,0 +1,30 @@
+
+from ... schema import GraphRagQuery, GraphRagResponse
+from ... schema import graph_rag_request_queue
+from ... schema import graph_rag_response_queue
+
+from . endpoint import ServiceEndpoint
+
+class GraphRagEndpoint(ServiceEndpoint):
+    def __init__(self, pulsar_host, timeout):
+
+        super(GraphRagEndpoint, self).__init__(
+            pulsar_host=pulsar_host,
+            request_queue=graph_rag_request_queue,
+            response_queue=graph_rag_response_queue,
+            request_schema=GraphRagQuery,
+            response_schema=GraphRagResponse,
+            endpoint_path="/api/v1/graph-rag",
+            timeout=timeout,
+        )
+
+    def to_request(self, body):
+        return GraphRagQuery(
+            query=body["query"],
+            user=body.get("user", "trustgraph"),
+            collection=body.get("collection", "default"),
+        )
+
+    def from_response(self, message):
+        return { "response": message.response }
+

trustgraph-flow-0.17.0/trustgraph/api/gateway/internet_search.py

@@ -0,0 +1,29 @@
+
+from ... schema import LookupRequest, LookupResponse
+from ... schema import internet_search_request_queue
+from ... schema import internet_search_response_queue
+
+from . endpoint import ServiceEndpoint
+
+class InternetSearchEndpoint(ServiceEndpoint):
+    def __init__(self, pulsar_host, timeout):
+
+        super(InternetSearchEndpoint, self).__init__(
+            pulsar_host=pulsar_host,
+            request_queue=internet_search_request_queue,
+            response_queue=internet_search_response_queue,
+            request_schema=LookupRequest,
+            response_schema=LookupResponse,
+            endpoint_path="/api/v1/internet-search",
+            timeout=timeout,
+        )
+
+    def to_request(self, body):
+        return LookupRequest(
+            term=body["term"],
+            kind=body.get("kind", None),
+        )
+
+    def from_response(self, message):
+        return { "text": message.text }
+

trustgraph-flow-0.17.0/trustgraph/api/gateway/prompt.py

@@ -0,0 +1,41 @@
+
+import json
+
+from ... schema import PromptRequest, PromptResponse
+from ... schema import prompt_request_queue
+from ... schema import prompt_response_queue
+
+from . endpoint import ServiceEndpoint
+
+class PromptEndpoint(ServiceEndpoint):
+    def __init__(self, pulsar_host, timeout):
+
+        super(PromptEndpoint, self).__init__(
+            pulsar_host=pulsar_host,
+            request_queue=prompt_request_queue,
+            response_queue=prompt_response_queue,
+            request_schema=PromptRequest,
+            response_schema=PromptResponse,
+            endpoint_path="/api/v1/prompt",
+            timeout=timeout,
+        )
+
+    def to_request(self, body):
+        return PromptRequest(
+            id=body["id"],
+            terms={
+                k: json.dumps(v)
+                for k, v in body["variables"].items()
+            }
+        )
+
+    def from_response(self, message):
+        if message.object:
+            return {
+                "object": message.object
+            }
+        else:
+            return {
+                "text": message.text
+            }
+

trustgraph-flow-0.17.0/trustgraph/api/gateway/publisher.py

@@ -0,0 +1,41 @@
+
+import asyncio
+import aiopulsar
+
+class Publisher:
+
+    def __init__(self, pulsar_host, topic, schema=None, max_size=10,
+                 chunking_enabled=False):
+        self.pulsar_host = pulsar_host
+        self.topic = topic
+        self.schema = schema
+        self.q = asyncio.Queue(maxsize=max_size)
+        self.chunking_enabled = chunking_enabled
+
+    async def run(self):
+
+        while True:
+
+            try:
+                async with aiopulsar.connect(self.pulsar_host) as client:
+                    async with client.create_producer(
+                            topic=self.topic,
+                            schema=self.schema,
+                            chunking_enabled=self.chunking_enabled,
+                    ) as producer:
+                        while True:
+                            id, item = await self.q.get()
+
+                            if id:
+                                await producer.send(item, { "id": id })
+                            else:
+                                await producer.send(item)
+
+            except Exception as e:
+                print("Exception:", e, flush=True)
+
+            # If handler drops out, sleep a retry
+            await asyncio.sleep(2)
+
+    async def send(self, id, msg):
+        await self.q.put((id, msg))

trustgraph-flow-0.17.0/trustgraph/api/gateway/running.py

@@ -0,0 +1,5 @@
+
+class Running:
+    def __init__(self): self.running = True
+    def get(self): return self.running
+    def stop(self): self.running = False

trustgraph-flow-0.17.0/trustgraph/api/gateway/serialize.py

@@ -0,0 +1,57 @@
+from ... schema import Value, Triple
+
+def to_value(x):
+    return Value(value=x["v"], is_uri=x["e"])
+
+def to_subgraph(x):
+    return [
+        Triple(
+            s=to_value(t["s"]),
+            p=to_value(t["p"]),
+            o=to_value(t["o"])
+        )
+        for t in x
+    ]
+
+def serialize_value(v):
+    return {
+        "v": v.value,
+        "e": v.is_uri,
+    }
+
+def serialize_triple(t):
+    return {
+        "s": serialize_value(t.s),
+        "p": serialize_value(t.p),
+        "o": serialize_value(t.o)
+    }
+
+def serialize_subgraph(sg):
+    return [
+        serialize_triple(t)
+        for t in sg
+    ]
+
+def serialize_triples(message):
+    return {
+        "metadata": {
+            "id": message.metadata.id,
+            "metadata": serialize_subgraph(message.metadata.metadata),
+            "user": message.metadata.user,
+            "collection": message.metadata.collection,
+        },
+        "triples": serialize_subgraph(message.triples),
+    }
+    
+def serialize_graph_embeddings(message):
+    return {
+        "metadata": {
+            "id": message.metadata.id,
+            "metadata": serialize_subgraph(message.metadata.metadata),
+            "user": message.metadata.user,
+            "collection": message.metadata.collection,
+        },
+        "vectors": message.vectors,
+        "entity": serialize_value(message.entity),
+    }
+