Datasets:

Modalities:
Tabular
Text
Formats:
parquet
ArXiv:
License:
jonan-richards commited on
Commit
04013a4
·
verified ·
1 Parent(s): d38e010

Add preprint, update license

Browse files
Files changed (2) hide show
  1. DATA_LICENSE +2 -2
  2. README.md +38 -10
DATA_LICENSE CHANGED
@@ -47,8 +47,8 @@ https://seart-ghs.si.usi.ch/
47
  https://github.com/seart-group/ghs
48
  https://doi.org/10.5281/zenodo.4588464
49
 
50
- The seed CSV is not distributed with AgentLogs. Repository records in this dataset
51
- were selected using that seed and collected from the GitHub and Copilot APIs.
52
 
53
  GitHub Search is used under the MIT License below. Redistributors must include
54
  this copyright and permission notice for that component.
 
47
  https://github.com/seart-group/ghs
48
  https://doi.org/10.5281/zenodo.4588464
49
 
50
+ The seed CSV is not distributed with AgentLogs. The repositories table is built
51
+ from that CSV.
52
 
53
  GitHub Search is used under the MIT License below. Redistributors must include
54
  this copyright and permission notice for that component.
README.md CHANGED
@@ -17,13 +17,16 @@ configs:
17
 
18
  # AgentLogs
19
 
20
- [![version 0.2](https://img.shields.io/badge/version-0.2-blue)](https://huggingface.co/datasets/risenlab/agentlogs/tree/v0.2)
21
- [![PyPI](https://img.shields.io/pypi/v/risenlab-agentlogs)](https://pypi.org/project/risenlab-agentlogs/)
22
- [![GitHub](https://img.shields.io/badge/GitHub-code-black?logo=github)](https://github.com/risenlab/agentlogs)
 
23
 
24
  AgentLogs is a dataset of activity related to the [GitHub agents functionality](https://github.com/features/copilot/agents): repository metadata, agent tasks, sessions, session logs (messages, tool calls, usage details, etc.), and user records.
25
 
26
- Stay tuned for the preprint!
 
 
27
 
28
  | | # Records | Size | Table | Content |
29
  | --- | ---: | ---: | --- | --- |
@@ -81,17 +84,42 @@ from agentlogs.schema import assert_version
81
 
82
  assert_version("v0.2")
83
 
84
- repositories = load_dataset("risenlab/agentlogs", "repositories", revision="v0.2")
85
- tasks = load_dataset("risenlab/agentlogs", "agent_tasks", revision="v0.2")
86
- sessions = load_dataset("risenlab/agentlogs", "agent_sessions", revision="v0.2")
87
- users = load_dataset("risenlab/agentlogs", "users", revision="v0.2")
88
  ```
89
 
90
  `agent_session_logs` is large. Prefer a snapshot. Do not stream that config with `load_dataset` (full rows, including payloads, are downloaded to your machine).
91
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
92
  ## License
93
 
94
- [CC BY 4.0](DATA_LICENSE), with the GitHub Search (GHS) MIT notice. Comply with
95
- GitHub's terms when using this data. Suggested attribution:
96
 
97
  > AgentLogs dataset by RISEN Lab, licensed under CC BY 4.0.
 
 
 
 
 
 
 
 
 
17
 
18
  # AgentLogs
19
 
20
+ <a href="https://huggingface.co/datasets/risenlab/agentlogs" target="_blank" rel="noopener noreferrer"><img src="https://img.shields.io/badge/Version-v0.2-blue?logo=huggingface&amp;logoColor=whitesmoke" alt="Version"></a>
21
+ <a href="https://pypi.org/project/risenlab-agentlogs/" target="_blank" rel="noopener noreferrer"><img src="https://img.shields.io/pypi/v/risenlab-agentlogs?label=Package&amp;logo=pypi&amp;logoColor=whitesmoke" alt="Package"></a>
22
+ <a href="https://github.com/risenlab/agentlogs/tree/main" target="_blank" rel="noopener noreferrer"><img src="https://img.shields.io/badge/Code-v0.2-181717?logo=github&amp;logoColor=whitesmoke" alt="Code"></a>
23
+ <a href="https://arxiv.org/abs/2608.29204" target="_blank" rel="noopener noreferrer"><img src="https://img.shields.io/badge/Preprint-2608.29204-b31b1b.svg?logo=arxiv&amp;logoColor=whitesmoke" alt="Preprint"></a>
24
 
25
  AgentLogs is a dataset of activity related to the [GitHub agents functionality](https://github.com/features/copilot/agents): repository metadata, agent tasks, sessions, session logs (messages, tool calls, usage details, etc.), and user records.
26
 
27
+ This dataset is described in:
28
+
29
+ > Jonan Richards, Kosei Horikawa, Youmei Fan, Yutaro Kashiwa, and Mairieli Wessel (2026), *AgentLogs: A Dataset for Opening the Black Box of GitHub's Cloud Agent*. arXiv: [2608.29204](https://arxiv.org/abs/2608.29204) (preprint).
30
 
31
  | | # Records | Size | Table | Content |
32
  | --- | ---: | ---: | --- | --- |
 
84
 
85
  assert_version("v0.2")
86
 
87
+ repositories = load_dataset("risenlab/agentlogs", "repositories", split="train", revision="v0.2")
88
+ tasks = load_dataset("risenlab/agentlogs", "agent_tasks", split="train", revision="v0.2")
89
+ sessions = load_dataset("risenlab/agentlogs", "agent_sessions", split="train", revision="v0.2")
90
+ users = load_dataset("risenlab/agentlogs", "users", split="train", revision="v0.2")
91
  ```
92
 
93
  `agent_session_logs` is large. Prefer a snapshot. Do not stream that config with `load_dataset` (full rows, including payloads, are downloaded to your machine).
94
 
95
+ ## Citation
96
+
97
+ If you use AgentLogs in an academic publication, please cite this preprint:
98
+
99
+ ```bibtex
100
+ @misc{richards2026AgentLogsDatasetOpening,
101
+ title = {{AgentLogs: A Dataset for Opening the Black Box of GitHub's Cloud Agent}},
102
+ author = {Richards, Jonan and Horikawa, Kosei and Fan, Youmei and Kashiwa, Yutaro and Wessel, Mairieli},
103
+ year = 2026,
104
+ month = aug,
105
+ eprint = {2608.29204},
106
+ primaryclass = {cs.SE},
107
+ doi = {10.48550/arXiv.2608.29204},
108
+ archiveprefix = {arXiv},
109
+ url = {https://arxiv.org/abs/2608.29204}
110
+ }
111
+ ```
112
+
113
  ## License
114
 
115
+ [CC BY 4.0](DATA_LICENSE). Suggested attribution:
 
116
 
117
  > AgentLogs dataset by RISEN Lab, licensed under CC BY 4.0.
118
+
119
+ Repository sampling uses [GitHub Search](https://seart-ghs.si.usi.ch/)
120
+ ([GitHub](https://github.com/seart-group/ghs), [Zenodo](https://doi.org/10.5281/zenodo.4588464))
121
+ &copy; SEART Research Group and Contributors, used under the
122
+ [MIT License](https://github.com/seart-group/ghs/blob/master/LICENSE).
123
+
124
+ The GitHub Search seed CSV is not distributed with this dataset. The
125
+ `repositories` table is built from that CSV.