Using Exported ZFP Settings with ria-toolkit-oss

Each of the three Zone Fingerprinting screens in RIA Hub — Capture Setup, Campaign Control (Enrollment), and Zone Fingerprinting Demo (Inference) — has an Export Settings JSON button. The downloaded file captures all the configuration choices you made in the UI. This guide shows you how to take those files and reproduce or automate the same workflow from the command line using ria-toolkit-oss.

Why export settings?

• Reproducibility — share your exact SDR and capture configuration with teammates.
• Automation — feed the exported JSON directly into CI/CD pipelines or scheduled jobs.
• OSS portability — run the same fingerprinting pipeline on a server without a RIA Hub installation.

The three export files

Screen	Filename	screen field
Screen 1 — Capture Setup	zfp-capture-settings.json	capture_setup
Screen 2 — Enrollment (Campaign)	zfp-enrollment-settings.json	enrollment_campaign
Screen 3 — Inference	zfp-inference-settings.json	zone_fingerprinting_inference

All files follow the same envelope:

{
  "screen": "<screen_id>",
  "version": "1.0",
  "settings": { ... },
  "docs": "https://docs.ria.qoherent.ai/guides/conductor/zfp-settings-export"
}

---

Screen 1 — Capture Setup

The settings block mirrors the SDR capture session configuration:

{
  "protocol": "wifi",
  "device": "plutosdr",
  "centerFreq": 2437000000,
  "sampleRate": 10000000,
  "gain": 40,
  "durationSec": 30,
  "recordingsPerDevice": 5
}

Using the capture YAML workflow

Screen 1 also exports a YAML workflow file (capture_workflow.yml). Commit it to your campaign repo under .ria/capture_workflow.yml and run:

qdm capture run --workflow .ria/capture_workflow.yml

To drive the capture directly from the JSON settings file:

import json
from ria_toolkit_oss.capture import CaptureSession

with open("zfp-capture-settings.json") as f:
    cfg = json.load(f)["settings"]

session = CaptureSession(
    device=cfg["device"],
    center_freq=cfg["centerFreq"],
    sample_rate=cfg["sampleRate"],
    gain=cfg["gain"],
    duration_sec=cfg["durationSec"],
)
session.run()

---

Screen 2 — Enrollment (Campaign Control)

The settings block contains the full campaign configuration object — the same JSON you would type into the “Edit as JSON” view in the UI. Pass it directly to the qdm campaign run command:

qdm campaign run --config zfp-enrollment-settings.json

Or use it with the Python API:

import json
from ria_toolkit_oss.conductor import Campaign

with open("zfp-enrollment-settings.json") as f:
    payload = json.load(f)

# payload["settings"] is the campaign configuration dict
campaign = Campaign.from_dict(payload["settings"])
campaign.run()

Protocol-specific notes

Protocol	What you need to do
WiFi	Connect the device to a network and run a ping test or video stream to generate traffic.
Bluetooth	Put the device into pairing mode so it broadcasts advertisement packets continuously. It does not need to complete pairing.
Zigbee	Use a coordinator to put the device into join/permit-join mode, or trigger a sensor event.

---

Screen 3 — Inference (Zone Fingerprinting Demo)

The settings block captures the inference configuration:

{
  "sdr": { "center_freq": 2437000000, "sample_rate": 10000000, "gain": 40 },
  "inference": {
    "model_repo": "owner/repo",
    "model_file": "models/zfp_v1.onnx",
    "sdr_device": "plutosdr",
    "center_freq": 2437000000,
    "sample_rate": 10000000,
    "gain": 40,
    "detector_threshold": 0.7
  },
  "enrolled_repo": "owner/campaign-repo",
  "scan_mode": "continuous",
  "confidence_threshold": 0.8
}

Run inference from the command line:

qdm infer run \
  --model models/zfp_v1.onnx \
  --device plutosdr \
  --center-freq 2437000000 \
  --sample-rate 10000000 \
  --gain 40 \
  --confidence-threshold 0.8

Or from Python:

import json
from ria_toolkit_oss.inference import FingerprintInference

with open("zfp-inference-settings.json") as f:
    cfg = json.load(f)["settings"]

inf = FingerprintInference(
    model_path=cfg["inference"]["model_file"],
    device=cfg["inference"]["sdr_device"],
    center_freq=cfg["inference"]["center_freq"],
    sample_rate=cfg["inference"]["sample_rate"],
    gain=cfg["inference"]["gain"],
    confidence_threshold=cfg["confidence_threshold"],
)

for result in inf.stream():
    if result.authorized:
        print(f"✓ {result.device_id} ({result.confidence:.0%})")
    else:
        print(f"✗ Unknown — {result.device_id}")

---

Enrolled devices file

Screen 3 stores enrolled devices in your campaign repo at .ria/enrolled_devices.json. You can load and use it independently:

import json

with open(".ria/enrolled_devices.json") as f:
    devices = json.load(f)

enrolled_ids = {d["id"] for d in devices}

---

Combining all three exports

A common automation pattern is to chain all three settings files into a single pipeline script:

# 1. Capture training data
qdm capture run --workflow .ria/capture_workflow.yml

# 2. Run enrollment campaign
qdm campaign run --config zfp-enrollment-settings.json

# 3. Start live inference
qdm infer run --settings zfp-inference-settings.json

For questions, see the Zone Fingerprinting Demo guide or the Campaign Control guide.