deep-features-for-plagiarism-detection

Prototype

In folder features-calculation there is a subfolder named prototype. It contains a prototype of a tool which can possibly detect code duplicates.

How to configure:

1. Inside features-calculation/prototype create an empty subfolder named snippets.
2. Inside features-calculation create a subfolder named networks. Prototype expects that a trained network named network resides in this folder. So, there must be a binary file features-calculation/networks/network.
3. In the same way prototype assumes existance of a binary file features-calculation/svm/standard which contains a trained instance of SVM.

How to run:

Prototype has two commands:

• Add pull-request from github to "database"

Command line arguments: --add https://github.com/Ivan-Veselov/pull-request-test/pull/3

This command will save features of all methods that can be found in a given PR to snippets folder.

• Find methods from a given pull-request that has similar pairs inside the "database"

Command line arguments: --find-similarities https://github.com/Ivan-Veselov/pull-request-test/pull/4

This command compares methods from given pull-request with methods that were inserted inside the "database" by the previous command and shows the results of comparison.