Understanding and Customizing the Abstraction Layer

The abstraction layer bridges outgoing streams with connections from other projects, ensuring a seamless and consistent way to forward WebSocket (WS) connections. This design decouples stream handling from client connections, improving scalability and maintainability.

---

Features

• Stream Forwarding: Automatically forwards outgoing streams via WebSocket connections.
• Decoupled Architecture: Abstracts stream generation from client handling, enabling easier integration with external projects.
• Source Obfuscation: Stream sources are hidden from connecting clients. Client applications only interact with the abstraction layer via its exposed port, ensuring that the actual stream source locations remain concealed. (see benefits)
• Scalable: Supports multiple simultaneous connections without direct dependencies on stream sources.
• Customizable: Configurable behaviour to meet specific project (functional or security) needs.

---

Configuration

Currently there is no seperate configuration file, instead configurations related to the Abstraction-Layer (i.e. Running port) can be made through the root configuration file.

---

Adding/Removing Streams

There are two ways to add/remove streams to/from the Abstraction-Layer, depending on where your stream is running/being added from.

As an Argussight process

One option is to run a process within Argussight. In this case, ensure that your process inherits from the Streamer class. This allows Argussight's Spawner to manage the stream, ensuring it is correctly added and removed when your process is created and terminated. The stream can be accessed by using the unique name of the process. For more details on accessing the streams from the Abstraction-Layer, see below.

From outside the Argussight architecture

Argussight supports adding any stream to the Abstraction-Layer. Streams can be added through the corresponding gRPC request (AddStreamRequest). While any MJPEG or MPEG1 stream can be added, we highly recommend using MXCuBE's Video-Streamer to generate the stream. Below is an example of how to generate and add a stream to the Abstraction-Layer via the gRPC connection.

import argussight.grpc.argus_service_pb2 as pb2
import argussight.grpc.argus_service_pb2_grpc as pb2_grpc
import grpc
import uuid

# Create channel connection
channel = grpc.insecure_channel("localhost:50051") # replace with the location of your GRPC channel
stub = pb2_grpc.SpawnerServiceStub(channel)

# Create a test video-stream
port = "8000"
stream_hash = str(uuid.uuid1())
subprocess.Popen(
                [
                    "video-streamer",
                    "-uri",
                    "test",
                    "-hs",
                    "localhost",
                    "-p",
                    port,
                    "-of",
                    "MPEG1",
                    "-id",
                    stream_hash,
                ],
                close_fds=True,
            )

# Add Stream to the Abstraction Layer
try:
    stub.AddStream(
        pb2.AddStreamRequest(
            name="TestStream",
            port=port,
            stream_id=stream_hash,
        )
    )
except Exception:
    print("Couldn't add stream")

Note: The name argument serves as the unique identifier for a stream. If you add two streams with the same name, the first stream will be removed and replaced by the second.

---

Access Streams

To access an added stream, you must establish a WebSocket connection to ws://<host>:<port>/<name>, where <host> and <port> refer to the host and port of the abstraction-layer, and <name> is the unique identifier of the stream you wish to access.

Get a list of all available streams

To get a list of all accessable streams, a request (GetProcessesRequest) can be send to Argussight via gRPC. The response to that request contains information about all the processes and streams registered within the running Argussight server. To extract the identifiers of the streams from the responses body, you can do as follows:

import argussight.grpc.argus_service_pb2 as pb2
import argussight.grpc.argus_service_pb2_grpc as pb2_grpc
import grpc

# Create channel connection
channel = grpc.insecure_channel("localhost:50051") # replace with the location of your GRPC channel
stub = pb2_grpc.SpawnerServiceStub(channel)

# Fetch the stream names
try:
    response = stub.GetProcesses(pb2.GetProcessesRequest())
    if response.status = "success":
        stream_names = list(response.streams)
    else:
        print(f"Error occured when fetching processes {response.error_message}")
except Exception:
    print("Error occured when connecting to the server")