juniper-cascor-client

Python HTTP and WebSocket client for the juniper-cascor training service.

juniper-cascor-client is the REST and async-WebSocket client for juniper-cascor, the Cascade-Correlation training service. The synchronous JuniperCascorClient covers network lifecycle, training control, snapshots, metrics, and distributed-worker monitoring; two async WebSocket streams cover live monitoring (/ws/training, inbound events) and runtime control (/ws/control, outbound commands). It's the same client juniper-canopy uses to drive and observe a cascor backend.

Part of the Juniper platform. juniper-cascor-client is the HTTP/WebSocket client for the juniper-cascor training service in Juniper — a multi-package ML research platform built around constructive (Cascade-Correlation) and recurrent neural networks.

Install

pip install juniper-cascor-client

Optional extras: [observability] adds Prometheus counters for WebSocket frame validation; [test] and [dev] install the test and lint/type-check toolchains.

Quick start

REST:

from juniper_cascor_client import JuniperCascorClient

with JuniperCascorClient("http://localhost:8200") as client:  # origin only; not /v1
    client.create_network(input_size=2, output_size=2)
    client.start_training(
        dataset={"source": "inline"},
        inline_data={"train_x": [[0, 0], [1, 0], [0, 1], [1, 1]],
                     "train_y": [[1, 0], [0, 1], [0, 1], [1, 0]]},
        epochs=100,
    )
    print(client.get_training_status()["data"]["training_active"])

Live training stream (async):

import asyncio
from juniper_cascor_client import CascorTrainingStream

async def monitor():
    async with CascorTrainingStream("ws://localhost:8200") as stream:
        async for msg in stream:
            if msg["type"] == "metrics":
                print(msg["data"]["epoch"], msg["data"]["train_loss"])

asyncio.run(monitor())

Control stream (async): CascorControlStream(...).command("start", {"epochs": 200}).

API

JuniperCascorClient (synchronous REST): health_check / is_alive / is_ready / wait_for_ready; create_network / get_network / delete_network / get_topology / get_statistics; start_training / stop_training / pause_training / resume_training / reset_training; get_training_status / get_training_params / update_params; get_metrics / get_metrics_history; get_dataset / get_dataset_data / get_decision_boundary; list_snapshots / get_snapshot / save_snapshot / load_snapshot; list_workers / get_worker / get_worker_stats.

CascorTrainingStream (async, /ws/training): yields inbound JSON frames; raw iteration (async for) or callbacks (on_metrics, on_state, on_topology, on_cascade_add, on_candidate_progress, on_event, on_disconnect).

CascorControlStream (async, /ws/control): command(command, params) / set_params(params) with per-request correlation.

Status

Live on PyPI. The current version is shown by the badge above; see CHANGELOG.md. The WebSocket streams do not auto-reconnect — wrap reconnection logic around the stream context manager for long-running consumers.

Documentation

• docs/QUICK_START.md — installation and verification guide
• docs/REFERENCE.md — full API reference, error model, WebSocket frame catalogue, and base_url normalisation (APD-CCLIENT-005)
• docs/DEVELOPER_CHEATSHEET.md — development quick-reference

License

MIT — see LICENSE.