Add initial dataset card
Browse files
README.md
CHANGED
|
@@ -72,6 +72,108 @@ configs:
|
|
| 72 |
- split: test
|
| 73 |
path: commitchronicle-py-long-labels/test-*
|
| 74 |
---
|
| 75 |
-
#
|
| 76 |
|
| 77 |
-
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 72 |
- split: test
|
| 73 |
path: commitchronicle-py-long-labels/test-*
|
| 74 |
---
|
| 75 |
+
# LCA (CMG)
|
| 76 |
|
| 77 |
+
This is the data for CMG benchmark as part of LCA.
|
| 78 |
+
|
| 79 |
+
## How-to
|
| 80 |
+
|
| 81 |
+
1. Since the dataset is private, if you haven't used HF Hub before, add your token via `huggingface-cli` first:
|
| 82 |
+
|
| 83 |
+
```
|
| 84 |
+
huggingface-cli login
|
| 85 |
+
```
|
| 86 |
+
|
| 87 |
+
2. List all the available configs via [`datasets.get_dataset_config_names`](https://huggingface.co/docs/datasets/v2.14.3/en/package_reference/loading_methods#datasets.get_dataset_config_names) and choose an appropriate one
|
| 88 |
+
|
| 89 |
+
|
| 90 |
+
3. Load the data via [`load_dataset`](https://huggingface.co/docs/datasets/v2.14.3/en/package_reference/loading_methods#datasets.load_dataset):
|
| 91 |
+
|
| 92 |
+
```
|
| 93 |
+
from datasets import load_dataset
|
| 94 |
+
|
| 95 |
+
configuration = "TODO" # select a configuration
|
| 96 |
+
dataset = load_dataset("JetBrains-Research/lca-cmg", configuration, split="test")
|
| 97 |
+
```
|
| 98 |
+
|
| 99 |
+
Some notes:
|
| 100 |
+
* All the data we have is considered to be in test split
|
| 101 |
+
|
| 102 |
+
## Dataset Structure
|
| 103 |
+
|
| 104 |
+
This dataset contains two kinds of data: either *full data* about each commit (including modifications) or metadata with quality *labels*.
|
| 105 |
+
|
| 106 |
+
|
| 107 |
+
### Full data
|
| 108 |
+
|
| 109 |
+
This section concerns configuration with *full data* about each commit (no `-labels` suffix).
|
| 110 |
+
|
| 111 |
+
Each example has the following fields:
|
| 112 |
+
|
| 113 |
+
| **Field** | **Description** |
|
| 114 |
+
|:------------------:|:----------------------------------------:|
|
| 115 |
+
| `repo` | Commit repository. |
|
| 116 |
+
| `hash` | Commit hash. |
|
| 117 |
+
| `date` | Commit date. |
|
| 118 |
+
| `license` | Commit repository's license. |
|
| 119 |
+
| `message` | Commit message. |
|
| 120 |
+
| `mods` | Modifications from commit (see below for details). |
|
| 121 |
+
|
| 122 |
+
We consider two possible commit modifications formats:
|
| 123 |
+
* lists of file diffs
|
| 124 |
+
* full file sources before and after commit
|
| 125 |
+
|
| 126 |
+
|
| 127 |
+
Currently, only [CommitChronicle](https://huggingface.co/datasets/JetBrains-Research/commit-chronicle) is considered as an input source, so only the first commit modifications format is used.
|
| 128 |
+
|
| 129 |
+
### List of diffs (CommitChronicle)
|
| 130 |
+
|
| 131 |
+
Data point example:
|
| 132 |
+
|
| 133 |
+
```
|
| 134 |
+
{'hash': 'b76ed0db81b3123ede5dc5e5f1bddf36336f3722',
|
| 135 |
+
'repo': 'apache/libcloud',
|
| 136 |
+
'date': '05.03.2022 17:52:34',
|
| 137 |
+
'license': 'Apache License 2.0',
|
| 138 |
+
'message': 'Add tests which verify that all OpenStack driver can be instantiated\nwith all the supported auth versions.\nNOTE: Those tests will fail right now due to the regressions being\nintroduced recently which breaks auth for some versions.',
|
| 139 |
+
'mods': {'diff': [{'change_type': 'MODIFY',
|
| 140 |
+
'new_path': 'libcloud/test/compute/test_openstack.py',
|
| 141 |
+
'old_path': 'libcloud/test/compute/test_openstack.py',
|
| 142 |
+
'diff': '@@ -39,6 +39,7 @@ from libcloud.utils.py3 import u\n<...>'}],
|
| 143 |
+
'file_src': None}}
|
| 144 |
+
```
|
| 145 |
+
|
| 146 |
+
### Labels
|
| 147 |
+
|
| 148 |
+
This section concerns configuration with metadata and *labels* (with `-labels` suffix).
|
| 149 |
+
|
| 150 |
+
Each example has the following fields:
|
| 151 |
+
|
| 152 |
+
| **Field** | **Description** |
|
| 153 |
+
|:------------------:|:----------------------------------------:|
|
| 154 |
+
| `repo` | Commit repository. |
|
| 155 |
+
| `hash` | Commit hash. |
|
| 156 |
+
| `date` | Commit date. |
|
| 157 |
+
| `license` | Commit repository's license. |
|
| 158 |
+
| `message` | Commit message. |
|
| 159 |
+
| `label` | Label of current commit as a target for CMG task. |
|
| 160 |
+
| `comment` | Comment for a label for current commit (optional, might be empty). |
|
| 161 |
+
|
| 162 |
+
Labels are in 1-5 scale, where:
|
| 163 |
+
* 1 – strong no
|
| 164 |
+
* 2 – weak no
|
| 165 |
+
* 3 – unsure
|
| 166 |
+
* 4 – weak yes
|
| 167 |
+
* 5 – strong yes
|
| 168 |
+
|
| 169 |
+
Data point example:
|
| 170 |
+
|
| 171 |
+
```
|
| 172 |
+
{'hash': '1559a4c686ddc2947fc3606e1c4279062cc9480f',
|
| 173 |
+
'repo': 'appscale/gts',
|
| 174 |
+
'date': '15.07.2018 21:00:39',
|
| 175 |
+
'license': 'Apache License 2.0',
|
| 176 |
+
'message': 'Add auto_id_policy and logs_path flags\n\nThese changes were introduced in the 1.7.5 SDK.',
|
| 177 |
+
'label': 1,
|
| 178 |
+
'comment': 'no way to know the version'}
|
| 179 |
+
```
|