Logo
ScienceStack
Table of Contents
Fetching ...
Sign In

Code4MeV2: a Research-oriented Code-completion Platform

Roham Koohestani, Parham Bateni, Aydin Ebrahimi, Behdad Etezadi, Kiarash Karimi, Maliheh Izadi

TL;DR

The paper tackles the barrier of proprietary telemetry in AI-assisted software development by proposing Code4MeV2, an open-source, modular, client-server platform for code completion and context-aware chat within JetBrains IDEs. It emphasizes transparency, configurability, and extensibility to enable reproducible, large-scale human–AI interaction studies, backed by a Python FastAPI backend, Celery task queue, and PostgreSQL analytics store. Preliminary evaluations show competitive code-completion latency (~$186.31$ ms) and informative user feedback from both researchers and daily users, while highlighting areas for improvement such as UI configurability and project-wide context retrieval. By providing a shared infrastructure with an analytics dashboard and modular telemetry, Code4MeV2 aims to become foundational infrastructure for AI4SE research and invites broad community adoption and contribution.

Abstract

The adoption of AI-powered code completion tools in software development has increased substantially, yet the user interaction data produced by these systems remain proprietary within large corporations. This creates a barrier for the academic community, as researchers must often develop dedicated platforms to conduct studies on human--AI interaction, making reproducible research and large-scale data analysis impractical. In this work, we introduce Code4MeV2, a research-oriented, open-source code completion plugin for JetBrains IDEs, as a solution to this limitation. Code4MeV2 is designed using a client--server architecture and features inline code completion and a context-aware chat assistant. Its core contribution is a modular and transparent data collection framework that gives researchers fine-grained control over telemetry and context gathering. Code4MeV2 achieves industry-comparable performance in terms of code completion, with an average latency of 200~ms. We assess our tool through a combination of an expert evaluation and a user study with eight participants. Feedback from both researchers and daily users highlights its informativeness and usefulness. We invite the community to adopt and contribute to this tool. More information about the tool can be found at https://app.code4me.me.

Code4MeV2: a Research-oriented Code-completion Platform

TL;DR

The paper tackles the barrier of proprietary telemetry in AI-assisted software development by proposing Code4MeV2, an open-source, modular, client-server platform for code completion and context-aware chat within JetBrains IDEs. It emphasizes transparency, configurability, and extensibility to enable reproducible, large-scale human–AI interaction studies, backed by a Python FastAPI backend, Celery task queue, and PostgreSQL analytics store. Preliminary evaluations show competitive code-completion latency (~ ms) and informative user feedback from both researchers and daily users, while highlighting areas for improvement such as UI configurability and project-wide context retrieval. By providing a shared infrastructure with an analytics dashboard and modular telemetry, Code4MeV2 aims to become foundational infrastructure for AI4SE research and invites broad community adoption and contribution.

Abstract

The adoption of AI-powered code completion tools in software development has increased substantially, yet the user interaction data produced by these systems remain proprietary within large corporations. This creates a barrier for the academic community, as researchers must often develop dedicated platforms to conduct studies on human--AI interaction, making reproducible research and large-scale data analysis impractical. In this work, we introduce Code4MeV2, a research-oriented, open-source code completion plugin for JetBrains IDEs, as a solution to this limitation. Code4MeV2 is designed using a client--server architecture and features inline code completion and a context-aware chat assistant. Its core contribution is a modular and transparent data collection framework that gives researchers fine-grained control over telemetry and context gathering. Code4MeV2 achieves industry-comparable performance in terms of code completion, with an average latency of 200~ms. We assess our tool through a combination of an expert evaluation and a user study with eight participants. Feedback from both researchers and daily users highlights its informativeness and usefulness. We invite the community to adopt and contribute to this tool. More information about the tool can be found at https://app.code4me.me.

Paper Structure

This paper contains 10 sections, 3 figures.

Table of Contents

  1. Introduction
  2. Domain Analysis
  3. Code4Me V2
  4. General Architecture
  5. Backend
  6. Frontend
  7. Analytics Platform
  8. Preliminary Evaluation
  9. Limitations and Future Work
  10. Conclusion

Figures (3)

  • Figure 1: Graphical depiction of the high-level architecture of Code4Me V2. (a) represents the backend processes of the server. (b) gives an architectural view of the implemented IntelliJ Plugin for Code4Me V2, and (c) represents the admin dashboard.
  • Figure 2: Plugin screenshot (inside IntelliJ IDEA)
  • Figure 3: Model calibration pane in the analytics dashboard