Skip to content

STREAM_REFERENCE.md

Latest commit

 

History

History
182 lines (141 loc) · 5.65 KB

STREAM_REFERENCE.md

File metadata and controls

182 lines (141 loc) · 5.65 KB

Live Scene Graph Stream Reference (QUICK START)

🔴 Stream Status: ACTIVE ✅

Last Updated: 3/8/2026 11:04 PM (continuously updating)

Current Metrics:

  • Events Generated: 1790 (growing every ~1 second)
  • Snapshots Archived: 1771
  • Active Vehicles: 5 (IDs: 26, 31, 32, 33, 34)
  • Server: CARLA 0.9.16 running on 127.0.0.1:2000

📁 Files You Need for VIATRA

Immediate (Load These Into Eclipse)

  1. Metamodel Definition:

    queries/SceneGraph.ecore
    • Namespace: http://cas782/scenegraph
    • Classes: Scene, Vehicle, Pedestrian, Node, Edge
    • Status: ✅ Ready to import into Eclipse

  2. Query Patterns:

    queries/scenegraph.vql
    • Pattern 1: fastVehicle — vehicles with speed > 13.9 m/s
    • Pattern 2: connected — edges linking two nodes
    • Pattern 3: vehiclePedestrianSharedRoad — pedestrians on vehicle paths
    • Status: ✅ Copy-paste ready into VIATRA Query Explorer

  3. Live Model Instance (Latest):

    data/stream/latest_snapshot.xmi
    • Format: EMF XMI (auto-synced every tick)
    • Content: 5 vehicles with positions (x, y, z) and speeds
    • Last Synced: 3/8/2026 11:04 PM (5 seconds ago)
    • Example: 
      <scenegraph:Scene xmlns:scenegraph="http://cas782/scenegraph" ...>
  <nodes xsi:type="scenegraph:Vehicle" id="32" x="120.27" y="28.58" z="0.13" speed="0.0036" />
  <nodes xsi:type="scenegraph:Vehicle" id="33" x="-48.83" y="-8.58" z="-0.00" speed="0.00" />
  <!-- ... 3 more vehicles -->
</scenegraph:Scene>
    • Status: ✅ Ready to load as model instance

Advanced (Incremental Updates)

  1. Event Stream (JSONL):

    data/stream/events.jsonl
    • 1790 lines (1 per tick)
    • Format: JSON Lines (newline-delimited JSON)
    • Usage: Programmatic consumption for real-time delta updates
    • Example line: 
      {"timestamp": "2026-03-09T02:33:12.935688+00:00", "tick": 1, "snapshot": "data/stream/snapshots/snapshot_000001.xmi", "node_count": 9, "edge_count": 0, "node_changes": {"added": [{"node_type": "Vehicle", "external_id": "26", "x": -64.6446, "y": 24.471, "z": -0.0075, "speed": 0.0}], "removed": [], "updated": []}, "edge_changes": {"added": [], "removed": []}}

  2. Snapshot Archive:

    data/stream/snapshots/snapshot_000001.xmi ... snapshot_001771.xmi
    • 1771 historical snapshots (~29 minutes of continuous operation)
    • Each ~3-5 KB
    • For batch analysis or time-series queries
    • Status: ✅ Ready for batch loading

  3. JSON State Mirror:

    data/stream/current_state.json
    • Same data as latest_snapshot.xmi but in JSON format
    • Useful for programmatic consumption in Java/Python
    • Status: ✅ Available

⚡ Getting Started (3 Steps)

Step 1: Install Eclipse + VIATRA

Follow the detailed steps in VIATRA_STREAM_GUIDE.md section "Eclipse + VIATRA Setup"

Step 2: Import Metamodel

  1. Open Eclipse
  2. File → New → EMF Project
  3. Import → File System → Browse to queries/SceneGraph.ecore
  4. Validate namespace: http://cas782/scenegraph ✓

Step 3: Load Model + Run Queries

  1. Right-click ecore file → Run As → VIATRA Query Tooling
  2. Load instance: data/stream/latest_snapshot.xmi
  3. Paste patterns from queries/scenegraph.vql
  4. Execute in VIATRA Query Explorer
  5. See live results!

🔄 Real-Time Streaming (Optional)

The stream is continuously running in the background.

  • New snapshots written every ~1 second
  • Events logged to events.jsonl with deltas
  • latest_snapshot.xmi auto-updated

To see live updates in Eclipse:

  • Periodically reload latest_snapshot.xmi in your model
  • Or write a Java/Xtend program that polls events.jsonl every 1 second
  • See VIATRA_STREAM_GUIDE.md section "Real-Time Polling Loop (Pseudocode)" for Java example

📊 Example: Load and Query

In VIATRA Query Explorer:

Load your model:

File → Open Model Instance
Path: data/stream/latest_snapshot.xmi

Execute pattern (paste into VQL editor):

pattern fastVehicle(vehicle : Vehicle) {
    Vehicle.speed(vehicle, speed);
    check(speed > 13.9);
}

Expected Output:

Results (0 found in current snapshot)
-- All vehicles currently stationary or slow-moving

🎯 What You Have

Item Location Status Size
Metamodel queries/SceneGraph.ecore ✅ Ready ~2 KB
Queries queries/scenegraph.vql ✅ Ready ~1 KB
Latest State data/stream/latest_snapshot.xmi ✅ Live ~1 KB
Event Stream data/stream/events.jsonl ✅ Live ~1.8 MB
Snapshots data/stream/snapshots/ ✅ Live ~5-10 MB
Docs VIATRA_STREAM_GUIDE.md ✅ Ready ~12 KB

🔧 Troubleshooting Quick Links

  • "Cannot import metamodel" → See VIATRA_STREAM_GUIDE.md § "Step 4: Register Metamodel"
  • "No results for query" → All vehicles are stationary now; try loading an older snapshot or wait for traffic
  • "XMI parsing error" → Ensure you're loading from data/stream/latest_snapshot.xmi (not symlink)
  • "VIATRA not installed" → Follow Eclipse Marketplace steps in VIATRA_STREAM_GUIDE.md § "Installation"

ℹ️ System Info

  • CARLA Version: 0.9.16
  • Python: 3.12.10
  • Metamodel Namespace: http://cas782/scenegraph
  • XMI Version: 2.0
  • Server: 127.0.0.1:2000
  • Uptime: 29+ minutes (continuous)

📖 For detailed setup & integration, see: VIATRA_STREAM_GUIDE.md

🚀 Next: Follow Step 1 above to install Eclipse + VIATRA, then load the metamodel and model instance!