This pages introduces some of the tools, datasets, and algorithms we developed.
Semantic Graph Dataset
We are continously building a dataset of semantics graph. Currently,
each graph might represent a single program file or the diff between two
versions of it. Each graph is generated by processing a “task” item, which
specified where the graph comes from and how it is generated etc.
The dataset is used for internal research but also available to wider community
upon request (contact author of this page). Here are some basic stats about it:
Summary
- All tasks: 106436
Task status
| success |
failure |
| 97205 |
9231 |
Graph labels