Eclipse Static Analysis - 10 Java projects

This dataset consists of Eclipse's static analysis performed on 10 Java projects.

For each .java file of a test project, we ran a static analysis using Eclipse JDT Core allows us to retrieve all the possible function calls based on typing/imports for a given completion site). Each java project has three files structured as follows:

• *.json file. The file contains all the method declarations of the project and the function calls in their body. For each function call, the file lists all the possible function call that could have been made at that place in the source code. For practical purposes, we splitted this file into two text files.
• *_sequences.txt file. This file consists of all the method declaration + function call sequences in the project. The last element of each line corresponds to a completion site.
• *_proposals.txt file. Each line is made of the function-call suggestions retrieved by static analysis for the corresponding line in the *_sequences.txt file.

The corpus was used for the experiments in the paper Combining Code Embedding with Static Analysis for Function-Call Completion.

Github repository to replicate the experiments: https://github.com/mweyssow/cse-saner