Identifier Suggestion

Models for source code identifier suggestion built by learning from Big Code.

Architecture

VSCode Extension Demo

Setup

Requirements

1. Have Python 3.7 or above installed.

2. Have a Unix-like terminal with bash, coreutils and some other utilities like wget (required by some .sh scripts).

Installation

1. Run pip install -r requirements/dev.txt to install all Python package dependencies.

Installation with a virtual environment

For example, you can use the standard venv:

1. Run python3 -m venv .venv to create a virtual environment.

2. Run source .venv/bin/activate to activate the virtual environment. This step must be executed on every new terminal session.

3. Run pip install -r requirements/dev.txt to install all Python package dependencies.

Training the model

1. Run ./src/scripts/baseline/train.sh from this directory with all necessary packages pre-installed.

Serving the model

1. Run ./src/scripts/baseline/serve.sh from this directory with all necessary packages pre-installed.

Project Structure

Follows the Cookiecutter Data Science project structure.

├── LICENSE
├── README.md          <- The top-level README for developers using this project.
├── data
│   ├── external       <- Data from third party sources.
│   ├── interim        <- Intermediate data that has been transformed.
│   ├── meta           <- Data about the data.
│   ├── processed      <- The final, canonical data sets for modeling.
│   └── raw            <- The original, immutable data dump.
│
├── experiments        <- Generated JSON files with hyperparameters for random search
│
├── models             <- Trained and serialized models, model predictions, or model summaries
│
├── notebooks          <- Jupyter notebooks. Naming convention is a number (for ordering),
│                         the creator's initials, and a short `-` delimited description, e.g.
│                         `1.0-jqp-initial-data-exploration`.
│
├── reports            <- Generated analysis as HTML, PDF, LaTeX, etc.
│   └── thesis         <- Generated graphics and tables used in the thesis document
│
├── requirements       <- The requirements files for reproducing this environment.
│
├── src                <- Source code of the project.
│   ├── common         <- Module with common values and procedures
│   ├── data           <- Scripts for Java method name parsing and data preprocessing
│   ├── evaluation     <- Logic for evaluating model performance
│   ├── metrics        <- Metrics for evaluation of models
│   ├── models         <- Scripts to train models and then use trained models to make predictions
│   ├── pipelines      <- Training pipelines
│   ├── preprocessing  <- Data preprocessing logic
│   ├── scripts        <- Scripts for running model training or serving
│   ├── server         <- Model server
│   ├── utils          <- Utility functions
│   └── visualization  <- Scripts to create exploratory and results oriented visualizations
│
└── vscode-extension   <- Extension for VSCode