Welcome to LUPE page
LUPE is a novel approach to API and code recommendation, exploiting cutting-edge deep learning techniques. Thanks to the underlying Encoder-Decoder architecture specialized in transforming sequences, LUPE can effectively learn the order in which APIs occur.
LUPE Overview
The LUPE architecture is depicted in figure above. Data is fetched from various open-source sources 1 using the DATA CURATOR 2 component. The collected data is then transformed into a suitable format to store in CSV files by the DATA CONVERTER 3 , which then builds a universal dictionary
containing APIs and their corresponding IDs. The SEQUENCE BUILDER 4 component uses the dictionary to extract APIs from Rascal M3 files. It also generates input for RECOMMENDER, which learns from data in the training phase to provide recommendations in the testing/deployment phase.
Repository Structure
This repository is organized as follows:
- The tools directory contains the implementation of the different tools we developed:
- The dataset directory contains the datasets described in the paper that we use to evaluate LUPE:
- D200_SMALL: meta-data from 200 small projects
- D200_BIG: meta-data from 200 small projects
- D1200: meta-data from 1200 projects
Venue acronym table
| Acronym | Venue |
|---|---|
| ESEC/FSE | The ACM Joint European Software Engineering Conference and Symposium on the Foundations of Software Engineering |
| JSS | Journal of Systems and Software |
| ISSRE | International Symposium on Software Reliability Engineering |
| SANER | IEEE International Conference on Software Analysis, Evolution and Reengineering |
| TSE | IEEE Transaction on Software Engineering |
| KDD | Knowledge Discovery and Data Mining |
| OOPSLA | Object-Oriented Programming, Systems, Languages & Applications |
| ICSE | International Conference on Software Engineering |
| EMSE | International Symposium on Empirical Software Engineering and Measurement |