Description

This module tries to recognize programming language using neural network.

How to generate learning data

For each programming language create separate directory in dataset, where name of directory is a extension of file which should be analysed. Inside this directories put your learning data files.

Example:

+---dataset
|   \---java
|       |---testFile.java
|       \---test.java
|   \---js
|       \---test.js

Currently application supports only java, js, php, python but adding a new language is extremely easy:

Create a new file named as a programming language and save into src/keywords/langs (for example src/keywords/langs/c.js)
Export array of keywords which is used by given language.

To generate learning data:

node --harmony-async-await filesData.js - to generate data required to train neural network.

It generates filesData.json which is required to train neural network.

How to train neural network

To train neural network simply type:

node --harmony-async-await train.js

It generates netData.json which store neural network data and is used to determine programming language.

How to test results

To test your neural network type:

node --harmony-async-await tester.js filePath

Where filePath is path to test file.

Requirements

Node v7
yarn install to install dependencies