NJR-1 Dataset

NJR is a Normalized Java Resource. The <em>NJR-1</em> dataset consists of 293 Java bytecode programs, each of which executes at least 100 unique application methods at runtime. Additionally, 5 static analysis tools (SpotBugs, Wala, Doop, Soot, Petablox) successfully run on these programs. <br> These programs are repositories picked from the set of Java-8 projects on Github that compile and run successfully. <br> Each of these programs comes with an executable jar file, the compiled bytecode file, and the Java source code. There are 3 files available for download: <em>njr-1_dataset.zip, scripts.zip, benchmark_stats.csv.</em> <em>njr-1_dataset.zip</em> has the actual dataset programs. <em>scripts.zip</em> contains Python3 scripts to run analysis tools (SpotBugs, Wala, Doop, Soot, Petablox) on the entire dataset. The benchmark_stats.csv file lists, for each benchmark, the number of nodes and edges in its dynamic application call-graph, as well as the number of edges in its static application call-graph (as computed by Wala). <br> A summary of the same is listed here: <strong><em>Statistics Dynamic-Nodes Dynamic-Edges Static-Edges</em></strong><br> Mean 205 469 1404<br> St.Dev 199 464 2523<br> Median 149 327 610 To cite the dataset, please cite the following paper:<br> Jens Palsberg and Cristina V. Lopes, NJR: a Normalized Java Resource. <br> In Proceedings of ACM SIGPLAN International Workshop on State Of the Art in Program Analysis (SOAP), 2018.