Export Geant4 Simulation Events to JSON for 3D Web Visualization - OpenGL

[jhajduga]

✨ Issue: Export Geant4 Simulation Events to JSON for 3D Web Visualization

🎯 Goal

Enable the visualization of selected Geant4 simulation events (including full secondary particle showers) in a modern 3D web environment (e.g., using three.js).
This requires exporting complete event-level tracking data (trajectory points per particle) from ROOT trees into a single structured JSON file.

---

📋 Current State

• Event data is stored in TTree structures, using G4AnalysisManager.

• Per-step data (position, energy, time, track info) is already collected in vectors:

  • e.g., m_VoxelHitsTrkId, m_VoxelHitsTrkEnergy, m_VoxelHitsTrkPosX/Y/Z, etc.

• However, ParentID information is missing, which is needed to reconstruct particle shower trees (γ → e⁻/e⁺ → ...).

---

📦 Task Overview

🔧 [C++] Extend Tracking and NTuple Output

• In SteppingAction or hit creation logic, capture G4Track::GetParentID() for each step.
• Add a new vector (e.g., m_VoxelHitsParentTrkId) and include it in the NTuple output.
• Optionally, capture G4Track::GetDefinition()->GetPDGEncoding() and/or name for particle typing.

🧪 [Python] Write PyROOT-based Export Script

• Use PyROOT to load the .root file and access TTrees.

• Group all hits/steps by (G4EvtId, TrackId) to form trajectories.

• For each trajectory:

  • Store ordered 3D positions and times.
  • Include TrackId, ParentId, ParticleType.

• Dump all selected events into a single events.json file with a structure like:

  {
    "event_id": 42,
    "particles": [
      {
        "track_id": 1,
        "parent_id": 0,
        "pdg": 22,
        "name": "gamma",
        "trajectory": [
          { "x": ..., "y": ..., "z": ..., "t": ... },
          ...
        ]
      },
      ...
    ]
  }

🌐 [Frontend] Basic three.js Renderer (later step)

• Load events.json and render each particle as a colored trajectory line.
• Add basic interactivity (event selection, toggling particle types, play/pause time).
• Optional: visualize parent_id relations as shower trees or via color-coding.
• Optional: visualize process_type_id and 'electron_origin_type_id'/'particle_origin_type_id' in shower trees or via color-coding. (

---

📁 Output Format (Single File)

Output will be a JSON file:

• all_events.json (list of events with particle trajectories)

---

🧠 Notes / Considerations

• The filtering (e.g. “only events with primary γ > 5 MeV”) will be handled in Python, prior to JSON export.
• Parent tracking must be consistent: if TrackId == 1 has ParentId == 0, we assume it's a primary.
• No need to visualize voxels or energy deposition – focus is on geometry + causality of particle motion.
• Firstly - pull changes from dose3d/g4rt:51-the-process-of-recording-the-type-of-physical-event-in-g4 to obtain all required info about particles.

---

🛠 Tools

• Simulation: Geant4, ROOT TTree (G4AnalysisManager)
• Export: PyROOT
• Frontend: JavaScript + three.js
• (optional) ROOT.js considered, but not selected for now