Commit
·
1f82590
1
Parent(s):
49c33f2
Update README.md
Browse files
README.md
CHANGED
|
@@ -89,3 +89,37 @@ dataset_info:
|
|
| 89 |
download_size: 229809783
|
| 90 |
dataset_size: 347089107
|
| 91 |
---
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 89 |
download_size: 229809783
|
| 90 |
dataset_size: 347089107
|
| 91 |
---
|
| 92 |
+
|
| 93 |
+
# About the Dataset
|
| 94 |
+
|
| 95 |
+
The source code used to generate the dataset can be found on
|
| 96 |
+
[GitHub](https://github.com/alexjercan/bug-detection/tree/master/bugnet)
|
| 97 |
+
|
| 98 |
+
The dataset is based on the [CodeNet project](https://github.com/IBM/Project_CodeNet)
|
| 99 |
+
and contains Python and C++ code submissions for online coding competitions. The data
|
| 100 |
+
is obtained by selecting consecutive attempts of a single user that resulted in fixing a
|
| 101 |
+
buggy submission. Thus the data is represented by code pairs and annotated by the diff
|
| 102 |
+
and error of each changed instruction. We have already tokenized all the source code
|
| 103 |
+
files and kept the same format as in the original dataset.
|
| 104 |
+
|
| 105 |
+
The upgrade made compared to CodeNetPy is that we only keep one line errors.
|
| 106 |
+
This means that the task of bug detection and repair will be easier to manage.
|
| 107 |
+
We also removed all the files that fail on linters, so that we are focusing only
|
| 108 |
+
on bugs that cannot be identified easily.
|
| 109 |
+
|
| 110 |
+
The resulting dataset file will be a csv with the following columns:
|
| 111 |
+
- `problem_id`: The id of the problem, matches with the id from Project_CodeNet
|
| 112 |
+
- `language`: The programming language of the submission (`Python` or `C++`)
|
| 113 |
+
- `original_status`: The status of the initial submission (`TLE`, `MLE`, anything that is not `Accepted`)
|
| 114 |
+
- `fail`: The initial (buggy) source code formatted (`black` or `clang-fromat`)
|
| 115 |
+
- `pass`: The modified (accepted) source code formatted(`black` or `clang-format`
|
| 116 |
+
- `change`: The change that was made (`replace`, `insert`, `delete`)
|
| 117 |
+
- `i1`: Start of the change in the buggy source (the line; starting with 1)
|
| 118 |
+
- `i2`: End of the change in the buggy source (not inclusive; for `insert` we have `i1 == i2`)
|
| 119 |
+
- `j1`: Start of the change in the accepted source (the line; starting with 1)
|
| 120 |
+
- `j2`: End of the change in the accepted source (not inclusive; for `delete` we have `j1 == j2`)
|
| 121 |
+
- `error`: The error that was obtained running the buggy source code on the input/output examples
|
| 122 |
+
- `stderr`: The full output of stderr of running the buggy source code on the input/output examples
|
| 123 |
+
- `description`: The problem statement in html format
|
| 124 |
+
- `input`: The input for the test case
|
| 125 |
+
- `output`: The output for the test case
|