DieWeaver quickstart

Start with the playground

Open /demos/dieweaver, load the supplied example, change a scenario constraint and run. Download the resulting evidence JSON. The hosted adapter uses a bounded subprocess, not an animation.

Input contract

requirements sets capability bits, minimum health, selected resource count, power and latency budgets, and minimum failure-domain diversity. resources supplies candidate IDs and numeric attributes. The public adapter caps resources at 16 and selected count at 6. Score and power units are scenario-defined, not hardware measurements.

{
  "requirements": {
    "caps": 3,
    "health_min": 70,
    "count": 2,
    "max_total_power": 20,
    "max_latency": 10,
    "min_failure_domains": 2,
    "weights": {
      "diversity": 25
    }
  },
  "resources": [
    {
      "id": "a",
      "valid": true,
      "health": 95,
      "caps": 7,
      "score": 90,
      "latency": 4,
      "power": 8,
      "failure_domain": "die0"
    },
    {
      "id": "b",
      "valid": true,
      "health": 88,
      "caps": 3,
      "score": 87,
      "latency": 5,
      "power": 9,
      "failure_domain": "die1"
    },
    {
      "id": "c",
      "valid": true,
      "health": 96,
      "caps": 7,
      "score": 99,
      "latency": 3,
      "power": 9,
      "failure_domain": "die0"
    }
  ]
}

Expected result

The published sample selects resources b and c. They span two failure domains, use 18 scenario power units, and produce an objective of 394. An infeasible response is a valid result when constraints cannot be met.

{
  "status": "optimized",
  "assignments": [
    {
      "logical": 0,
      "resource": "b",
      "failure_domain": "die1"
    },
    {
      "logical": 1,
      "resource": "c",
      "failure_domain": "die0"
    }
  ],
  "objective": 394.0,
  "total_power": 18.0,
  "failure_domains": 2,
  "constraints": {
    "max_total_power": 20.0,
    "max_latency": 10.0,
    "min_failure_domains": 2
  }
}

Use the licensed package

After entitled delivery, extract the named commercial release and run its CLI in an isolated Python environment. These commands refer to the actual archived release, not a package invented for the website.

python -m pip install ./dist/garmo_dieweaver-1.2.0-py3-none-any.whl
python -m dieweaver advanced-demo
python -m dieweaver serve --port 8787

What this demonstrates

The algorithm is executing real software on synthetic input. Production datasets, calibration, integration performance, service levels and customer ROI require separate evaluation. The website adapter limits requests and does not expose the release source.