Gsoc 2026: My Approach and understanding of the 1st project and how should i procced further?

[Adarshsharath]

Hello, I’m Adarsh Prakash, a second-year B.E./B.Tech student with a strong interest in real-time systems and scientific computing.

I’ve been studying Project 1 (Real-Time MD Simulation Dashboard) and tried to break it down into a clear implementation plan. I would really appreciate your feedback on whether my understanding is correct and if I am heading in the right direction.

---

Proposed Approach to Solve Project 1: Real-Time MD Simulation Dashboard

My approach is to build the system incrementally, starting from a simple real-time pipeline and progressively integrating MDAnalysis and streaming capabilities.

---

1. Overall Design

The system will follow a streaming pipeline:

Simulation (IMDv3 stream) → imdclient → MDAnalysis → Backend (FastAPI) → WebSocket → Browser Dashboard

The backend will act as the central coordinator, receiving streaming data, performing analysis, and pushing results to the frontend in real time.

---

2. Phase-wise Implementation Plan

Phase 1: Prototype Real-Time Dashboard (No MDAnalysis)

• Build a FastAPI backend with WebSocket support
• Simulate streaming data (mock frames)
• Create a browser-based dashboard (HTML/JS + Chart.js)
• Display live updates (frame number and computed values)

Goal: Establish a working real-time data pipeline

---

Phase 2: Integrate MDAnalysis with File-Based Trajectories

• Load sample datasets using MDAnalysis (e.g., PSF/DCD)

• Implement core operations:

  • Universe creation
  • Trajectory iteration
  • Atom selection

• Compute simple observables:

  • Distance
  • Radius of gyration

• Stream computed values to the frontend via WebSocket

Goal: Replace mock data with real analysis

---

Phase 3: Transition to Streaming Data (IMDv3)

• Integrate imdclient to receive live simulation data
• Use MDAnalysis IMDReader to treat the stream as a trajectory
• Replace file-based iteration with a streaming loop

Goal: Enable real-time processing of live simulations

---

Phase 4: Interactive Analysis Layer

• Add UI controls for:

  • Atom selection (MDAnalysis selection syntax)
  • Choice of analysis (distance, angles, etc.)

• Dynamically update backend computations based on user input

Goal: Make the dashboard interactive and user-driven

---

Phase 5: Time-Dependent Analysis

• Implement buffering of previous frames

• Support advanced analyses:

  • Moving averages
  • Autocorrelation-like metrics

Goal: Enable analysis beyond single-frame computations

---

Phase 6: Visualization and Alerts

• Add real-time plots (Chart.js or similar)

• Implement event detection:

  • Threshold-based alerts
  • Structural anomaly detection

• Display warnings in the UI

Goal: Provide actionable insights during simulation

---

3. Technical Stack

• Backend: FastAPI (async, WebSockets)
• Analysis: MDAnalysis
• Streaming: imdclient (IMDv3 protocol)
• Frontend: HTML/JS (or React) + Chart.js

---

4. Key Challenges & Solutions

• Real-time performance:
  Use asynchronous processing and efficient frame handling

• Streaming integration:
  Use the IMDReader abstraction to handle streams like trajectories

• Scalability:
  Start with a modular single-user system and extend to multi-user support

---

5. Expected Outcome

• A functional web-based dashboard
• Real-time monitoring of MD simulations
• Interactive analysis tools
• A modular and extensible architecture

---

6. Development Strategy

I will follow an iterative approach:

• Build a minimal working version first
• Gradually replace mock components with real integrations
• Continuously test with increasing complexity

---

This approach ensures steady progress, early validation, and a robust final system.

---

I would be grateful if you could let me know whether my understanding aligns with the project expectations, and if there are any areas I should improve or explore further.

[Adarshsharath]

@orbeckst @kurtmckee @jandom @Marcello-Sega Can anyone please Guide me i am really interested in this project as it connects Chemistry with Tech....

[kurtmckee]

@Adarshsharath I'm not an appropriate person to @-mention -- I've contributed only once to this project.

Good luck with your project!

[orbeckst]

Hello @Adarshsharath , thank you for your interest. As you see in https://www.mdanalysis.org/2026/02/19/gsoc2026/ we had a pre-proposal and video-interview stage for GSCO 2026 and the deadline for prepropsal submission has passed. If you have not been invited to submit a full proposal then we will not be able to consider your GSOC proposal this year.

You're still welcome to contribute to MDAnalysis, of course.