trustgraph-flow 0.17.2__tar.gz → 0.17.4__tar.gz

{trustgraph-flow-0.17.2 → trustgraph-flow-0.17.4}/PKG-INFO

@@ -1,12 +1,12 @@
 Metadata-Version: 2.1
 Name: trustgraph-flow
-Version: 0.17.2
+Version: 0.17.4
 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/v0.17.2.tar.gz
+Download-URL: https://github.com/trustgraph-ai/trustgraph/archive/refs/tags/v0.17.4.tar.gz
 Platform: UNKNOWN
 Classifier: Programming Language :: Python :: 3
 Classifier: License :: OSI Approved :: GNU General Public License v3 or later (GPLv3+)

{trustgraph-flow-0.17.2 → trustgraph-flow-0.17.4}/setup.py

@@ -59,7 +59,6 @@ setuptools.setup(
         "ibis",
         "jsonschema",
         "aiohttp",
-        "aiopulsar-py",
         "pinecone[grpc]",
     ],
     scripts=[

{trustgraph-flow-0.17.2 → trustgraph-flow-0.17.4}/trustgraph/api/gateway/endpoint.py

@@ -43,8 +43,8 @@ class ServiceEndpoint:
 
     async def start(self):
 
-        self.pub_task = asyncio.create_task(self.pub.run())
-        self.sub_task = asyncio.create_task(self.sub.run())
+        self.pub.start()
+        self.sub.start()
 
     def add_routes(self, app):
 
@@ -82,20 +82,18 @@ class ServiceEndpoint:
 
             print(data)
 
-            q = await self.sub.subscribe(id)
+            q = self.sub.subscribe(id)
 
-            await self.pub.send(
-                id,
-                self.to_request(data),
+            await asyncio.to_thread(
+                self.pub.send, id, self.to_request(data)
             )
-            print("Request sent")
 
             try:
-                resp = await asyncio.wait_for(q.get(), self.timeout)
-            except:
-                raise RuntimeError("Timeout waiting for response")
+                resp = await asyncio.to_thread(q.get, timeout=self.timeout)
+            except Exception as e:
+                raise RuntimeError("Timeout")
 
-            print("Response got")
+            print(resp)
 
             if resp.error:
                 print("Error")
@@ -103,8 +101,6 @@ class ServiceEndpoint:
                     { "error": resp.error.message }
                 )
 
-            print("Send response")
-
             return web.json_response(
                 self.from_response(resp)
             )
@@ -117,7 +113,7 @@ class ServiceEndpoint:
             )
 
         finally:
-            await self.sub.unsubscribe(id)
+            self.sub.unsubscribe(id)
 
 
 class MultiResponseServiceEndpoint(ServiceEndpoint):
@@ -130,11 +126,10 @@ class MultiResponseServiceEndpoint(ServiceEndpoint):
 
             data = await request.json()
 
-            q = await self.sub.subscribe(id)
+            q = self.sub.subscribe(id)
 
-            await self.pub.send(
-                id,
-                self.to_request(data),
+            await asyncio.to_thread(
+                self.pub.send, id, self.to_request(data)
             )
 
             # Keeps looking at responses...
@@ -142,8 +137,8 @@ class MultiResponseServiceEndpoint(ServiceEndpoint):
             while True:
 
                 try:
-                    resp = await asyncio.wait_for(q.get(), self.timeout)
-                except:
+                    resp = await asyncio.to_thread(q.get, timeout=self.timeout)
+                except Exception as e:
                     raise RuntimeError("Timeout waiting for response")
 
                 if resp.error:
@@ -168,4 +163,4 @@ class MultiResponseServiceEndpoint(ServiceEndpoint):
             )
 
         finally:
-            await self.sub.unsubscribe(id)
+            self.sub.unsubscribe(id)

{trustgraph-flow-0.17.2 → trustgraph-flow-0.17.4}/trustgraph/api/gateway/graph_embeddings_load.py

@@ -31,9 +31,7 @@ class GraphEmbeddingsLoadEndpoint(SocketEndpoint):
 
     async def start(self):
 
-        self.task = asyncio.create_task(
-            self.publisher.run()
-        )
+        self.publisher.start()
 
     async def listener(self, ws, running):
         
@@ -56,7 +54,7 @@ class GraphEmbeddingsLoadEndpoint(SocketEndpoint):
                     vectors=data["vectors"],
                 )
 
-                await self.publisher.send(None, elt)
+                self.publisher.send(None, elt)
 
 
         running.stop()

{trustgraph-flow-0.17.2 → trustgraph-flow-0.17.4}/trustgraph/api/gateway/graph_embeddings_stream.py

@@ -1,5 +1,6 @@
 
 import asyncio
+import queue
 from pulsar.schema import JsonSchema
 import uuid
 
@@ -30,29 +31,27 @@ class GraphEmbeddingsStreamEndpoint(SocketEndpoint):
 
     async def start(self):
 
-        self.task = asyncio.create_task(
-            self.subscriber.run()
-        )
+        self.subscriber.start()
 
     async def async_thread(self, ws, running):
 
         id = str(uuid.uuid4())
 
-        q = await self.subscriber.subscribe_all(id)
+        q = self.subscriber.subscribe_all(id)
 
         while running.get():
             try:
-                resp = await asyncio.wait_for(q.get(), 0.5)
+                resp = await asyncio.to_thread(q.get, timeout=0.5)
                 await ws.send_json(serialize_graph_embeddings(resp))
 
-            except TimeoutError:
+            except queue.Empty:
                 continue
 
             except Exception as e:
                 print(f"Exception: {str(e)}", flush=True)
                 break
 
-        await self.subscriber.unsubscribe_all(id)
+        self.subscriber.unsubscribe_all(id)
 
         running.stop()
 

trustgraph-flow-0.17.4/trustgraph/api/gateway/publisher.py

@@ -0,0 +1,53 @@
+
+import queue
+import time
+import pulsar
+import threading
+
+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 = queue.Queue(maxsize=max_size)
+        self.chunking_enabled = chunking_enabled
+
+    def start(self):
+        self.task = threading.Thread(target=self.run)
+        self.task.start()
+
+    def run(self):
+
+        while True:
+
+            try:
+
+                client = pulsar.Client(
+                    self.pulsar_host,
+                )
+
+                producer = client.create_producer(
+                    topic=self.topic,
+                    schema=self.schema,
+                    chunking_enabled=self.chunking_enabled,
+                )
+
+                while True:
+
+                    id, item = self.q.get()
+
+                    if id:
+                        producer.send(item, { "id": id })
+                    else:
+                        producer.send(item)
+
+            except Exception as e:
+                print("Exception:", e, flush=True)
+
+            # If handler drops out, sleep a retry
+            time.sleep(2)
+
+    def send(self, id, msg):
+        self.q.put((id, msg))

{trustgraph-flow-0.17.2 → trustgraph-flow-0.17.4}/trustgraph/api/gateway/service.py

@@ -166,7 +166,8 @@ class Api:
             # content is valid base64
             doc = base64.b64decode(data["data"])
 
-            resp = await self.document_out.send(
+            resp = await asyncio.to_thread(
+                self.document_out.send,
                 None,
                 Document(
                     metadata=Metadata(
@@ -211,7 +212,8 @@ class Api:
             # Text is base64 encoded
             text = base64.b64decode(data["text"]).decode(charset)
 
-            resp = await self.text_out.send(
+            resp = asyncio.to_thread(
+                self.text_out.send,
                 None,
                 TextDocument(
                     metadata=Metadata(
@@ -242,8 +244,8 @@ class Api:
         for ep in self.endpoints:
             await ep.start()
 
-        self.doc_ingest_pub_task = asyncio.create_task(self.document_out.run())
-        self.text_ingest_pub_task = asyncio.create_task(self.text_out.run())
+        self.document_out.start()
+        self.text_out.start()
 
         return self.app
 

trustgraph-flow-0.17.4/trustgraph/api/gateway/subscriber.py

@@ -0,0 +1,109 @@
+
+import queue
+import pulsar
+import threading
+import time
+
+class Subscriber:
+
+    def __init__(self, pulsar_host, topic, subscription, consumer_name,
+                 schema=None, max_size=100):
+        self.pulsar_host = pulsar_host
+        self.topic = topic
+        self.subscription = subscription
+        self.consumer_name = consumer_name
+        self.schema = schema
+        self.q = {}
+        self.full = {}
+        self.max_size = max_size
+        self.lock = threading.Lock()
+
+    def start(self):
+        self.task = threading.Thread(target=self.run)
+        self.task.start()
+
+    def run(self):
+
+        while True:
+
+            try:
+
+                client = pulsar.Client(
+                    self.pulsar_host,
+                )
+
+                consumer = client.subscribe(
+                    topic=self.topic,
+                    subscription_name=self.subscription,
+                    consumer_name=self.consumer_name,
+                    schema=self.schema,
+                )
+
+                while True:
+
+                    msg = consumer.receive()
+
+                    # Acknowledge successful reception of the message
+                    consumer.acknowledge(msg)
+
+                    try:
+                        id = msg.properties()["id"]
+                    except:
+                        id = None
+
+                    value = msg.value()
+
+                    with self.lock:
+
+                        if id in self.q:
+                            try:
+                                self.q[id].put(value, timeout=0.5)
+                            except:
+                                pass
+
+                        for q in self.full.values():
+                            try:
+                                q.put(value, timeout=0.5)
+                            except:
+                                pass
+
+            except Exception as e:
+                print("Exception:", e, flush=True)
+         
+            # If handler drops out, sleep a retry
+            time.sleep(2)
+
+    def subscribe(self, id):
+
+        with self.lock:
+
+            q = queue.Queue(maxsize=self.max_size)
+            self.q[id] = q
+
+        return q
+
+    def unsubscribe(self, id):
+        
+        with self.lock:
+
+            if id in self.q:
+#                self.q[id].shutdown(immediate=True)
+                del self.q[id]
+    
+    def subscribe_all(self, id):
+
+        with self.lock:
+
+            q = queue.Queue(maxsize=self.max_size)
+            self.full[id] = q
+
+        return q
+
+    def unsubscribe_all(self, id):
+
+        with self.lock:
+
+            if id in self.full:
+#                self.full[id].shutdown(immediate=True)
+                del self.full[id]
+

{trustgraph-flow-0.17.2 → trustgraph-flow-0.17.4}/trustgraph/api/gateway/triples_load.py

@@ -29,9 +29,7 @@ class TriplesLoadEndpoint(SocketEndpoint):
 
     async def start(self):
 
-        self.task = asyncio.create_task(
-            self.publisher.run()
-        )
+        self.publisher.start()
 
     async def listener(self, ws, running):
         
@@ -53,7 +51,7 @@ class TriplesLoadEndpoint(SocketEndpoint):
                     triples=to_subgraph(data["triples"]),
                 )
 
-                await self.publisher.send(None, elt)
+                self.publisher.send(None, elt)
 
 
         running.stop()

{trustgraph-flow-0.17.2 → trustgraph-flow-0.17.4}/trustgraph/api/gateway/triples_stream.py

@@ -1,5 +1,6 @@
 
 import asyncio
+import queue
 from pulsar.schema import JsonSchema
 import uuid
 
@@ -28,29 +29,27 @@ class TriplesStreamEndpoint(SocketEndpoint):
 
     async def start(self):
 
-        self.task = asyncio.create_task(
-            self.subscriber.run()
-        )
+        self.subscriber.start()
 
     async def async_thread(self, ws, running):
 
         id = str(uuid.uuid4())
 
-        q = await self.subscriber.subscribe_all(id)
+        q = self.subscriber.subscribe_all(id)
 
         while running.get():
             try:
-                resp = await asyncio.wait_for(q.get(), 0.5)
+                resp = await asyncio.to_thread(q.get, timeout=0.5)
                 await ws.send_json(serialize_triples(resp))
 
-            except TimeoutError:
+            except queue.Empty:
                 continue
 
             except Exception as e:
                 print(f"Exception: {str(e)}", flush=True)
                 break
 
-        await self.subscriber.unsubscribe_all(id)
+        self.subscriber.unsubscribe_all(id)
 
         running.stop()
 

trustgraph-flow-0.17.4/trustgraph/flow_version.py

@@ -0,0 +1 @@
+__version__ = "0.17.4"

{trustgraph-flow-0.17.2 → trustgraph-flow-0.17.4}/trustgraph_flow.egg-info/PKG-INFO

@@ -1,12 +1,12 @@
 Metadata-Version: 2.1
 Name: trustgraph-flow
-Version: 0.17.2
+Version: 0.17.4
 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/v0.17.2.tar.gz
+Download-URL: https://github.com/trustgraph-ai/trustgraph/archive/refs/tags/v0.17.4.tar.gz
 Platform: UNKNOWN
 Classifier: Programming Language :: Python :: 3
 Classifier: License :: OSI Approved :: GNU General Public License v3 or later (GPLv3+)

{trustgraph-flow-0.17.2 → trustgraph-flow-0.17.4}/trustgraph_flow.egg-info/requires.txt

@@ -1,5 +1,4 @@
 aiohttp
-aiopulsar-py
 anthropic
 cassandra-driver
 cohere

trustgraph-flow-0.17.2/trustgraph/api/gateway/publisher.py

@@ -1,41 +0,0 @@
-
-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-0.17.2/trustgraph/api/gateway/subscriber.py

@@ -1,68 +0,0 @@
-
-import asyncio
-import aiopulsar
-
-class Subscriber:
-
-    def __init__(self, pulsar_host, topic, subscription, consumer_name,
-                 schema=None, max_size=10):
-        self.pulsar_host = pulsar_host
-        self.topic = topic
-        self.subscription = subscription
-        self.consumer_name = consumer_name
-        self.schema = schema
-        self.q = {}
-        self.full = {}
-
-    async def run(self):
-        while True:
-            try:
-                async with aiopulsar.connect(self.pulsar_host) as client:
-                    async with client.subscribe(
-                        topic=self.topic,
-                        subscription_name=self.subscription,
-                        consumer_name=self.consumer_name,
-                        schema=self.schema,
-                    ) as consumer:
-                        while True:
-                            msg = await consumer.receive()
-
-                            # Acknowledge successful reception of the message
-                            await consumer.acknowledge(msg)
-
-                            try:
-                                id = msg.properties()["id"]
-                            except:
-                                id = None
-
-                            value = msg.value()
-                            if id in self.q:
-                                await self.q[id].put(value)
-
-                            for q in self.full.values():
-                                await q.put(value)
-
-            except Exception as e:
-                print("Exception:", e, flush=True)
-         
-            # If handler drops out, sleep a retry
-            await asyncio.sleep(2)
-
-    async def subscribe(self, id):
-        q = asyncio.Queue()
-        self.q[id] = q
-        return q
-
-    async def unsubscribe(self, id):
-        if id in self.q:
-            del self.q[id]
-    
-    async def subscribe_all(self, id):
-        q = asyncio.Queue()
-        self.full[id] = q
-        return q
-
-    async def unsubscribe_all(self, id):
-        if id in self.full:
-            del self.full[id]
-

trustgraph-flow-0.17.2/trustgraph/flow_version.py

@@ -1 +0,0 @@
-__version__ = "0.17.2"