Datasets:

Modalities:
Tabular
Text
Formats:
json
ArXiv:
Libraries:
Datasets
pandas
File size: 4,888 Bytes
101bfae
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
d1e96e6
 
 
 
101bfae
 
 
cf5388d
101bfae
cf5388d
 
 
101bfae
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
cf5388d
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
---
configs:
- config_name: ghs_parameters
  data_files:
  - split: kotlin
    path: ghs/parameters/kotlin.json
  - split: java
    path: ghs/parameters/java.json
  - split: python
    path: ghs/parameters/python.json
- config_name: repo_metadata
  data_files:
  - split: jvm
    path: ghs/data/jvm.jsonl
  - split: python
    path: ghs/data/python.jsonl
- config_name: repo_metadata_permissive
  data_files:
  - split: jvm
    path: ghs/data/jvm_permissive.jsonl
  - split: python
    path: ghs/data/python_permissive.jsonl
- config_name: readmes
  data_files:
  - split: jvm
    path: readmes/jvm.jsonl
  - split: python
    path: readmes/python.jsonl
- config_name: workflows
  data_files:
  - split: jvm
    path: workflows/jvm.jsonl
  - split: python
    path: workflows/python.jsonl
- config_name: repo_contents
  data_files:
  - split: python
    path: repo_contents/python.jsonl
  - split: jvm
    path: repo_contents/jvm.jsonl
- config_name: splits
  data_files:
  - split: python_single_dm_in_root_folder
    path: splits/python_single_dm_in_root_folder.jsonl
  - split: python_single_dm_in_root_folder_no_docker
    path: splits/python_single_dm_in_root_folder_no_docker.jsonl
  - split: python_baseline_failure
    path: splits/python_baseline_failure.jsonl
  - split: jvm_single_dm_in_root_folder
    path: splits/jvm_single_dm_in_root_folder.jsonl
  - split: jvm_single_dm_in_root_folder_no_docker
    path: splits/jvm_single_dm_in_root_folder_no_docker.jsonl
  - split: jvm_baseline_failure
    path: splits/jvm_baseline_failure.jsonl
  - split: python_baseline_failure_test
    path: splits/python_baseline_failure_test.jsonl
  - split: python_baseline_failure_train
    path: splits/python_baseline_failure_train.jsonl
---


# 🌱⚙️ EnvBench

This repository contains data associated with EnvBench benchmark from [**EnvBench: A Benchmark for Automated Environment Setup**](https://arxiv.org/abs/2503.14443).

It contains:

* statistics about repositories from GitHub Search under `ghs/data` folder;
* several data splits under `splits` folder;
* Git repositories under `repos` folder;
* READMEs under `readmes` folder (available as `readmes` config);
* GitHub Actions workflows under `workflows` folder (available as `workflows` config);
* list of files in the repositories (available as `repo_contents` config).

## Data splits

We provide the repositories at each step of our filtering process.

### JVM

1. We obtain the list of repositories from GitHub Search: `ghs/data/jvm_permissive.jsonl`
2. We clone the repositories to `repos/full/jvm`
3. We filter repositories from previous step to keep only those with configuration files related to a single build tool in root directory: `splits/jvm_single_dm_in_root_folder.jsonl` (`splits/jvm_single_dm_in_root_folder.jsonl` additionally contains 'dm' field with build tool for each repository)
4. We filter the repositories from previous step to keep only those without Docker-related configuration files: `splits/jvm_single_dm_in_root_folder_no_docker.jsonl`
5. We filter the repositories from previous step to keep only those that our deterministic script couldn't configure correctly: `splits/jvm_baseline_failure.jsonl`

The list of 665 repositories from our benchmark is available in `splits/jvm_baseline_failure.jsonl`. For convenience, the final set of Git repositories is also stored separately under `repos/final/jvm`.

### Python

1. We obtain the list of repositories from GitHub Search: `ghs/data/python_permissive.jsonl`
2. We clone the repositories to `repos/full/python`
3. We filter repositories from previous step to keep only those with configuration files related to a single build tool in root directory: `splits/python_single_dm_in_root_folder.jsonl` (`splits/python_single_dm_in_root_folder.jsonl` additionally contains 'dm' field with dependency manager for each repository)
4. We filter the repositories from previous step to keep only those without Docker-related configuration files: `splits/python_single_dm_in_root_folder_no_docker.jsonl`
5. We filter the repositories from previous step to keep only those that our deterministic script couldn't configure correctly: `splits/python_baseline_failure.jsonl`

The list of 329 repositories from our benchmark is available in `splits/python_baseline_failure.jsonl`. For convenience, the final set of Git repositories is also stored separately under `repos/final/python`.

## 📚 Citation

If you find this work useful for your research, please consider the following citation:

```
@inproceedings{eliseeva2025envbench,
    title={EnvBench: A Benchmark for Automated Environment Setup},
    author={Aleksandra Eliseeva and Alexander Kovrigin and Ilia Kholkin and Egor Bogomolov and Yaroslav Zharov},
    booktitle={ICLR 2025 Third Workshop on Deep Learning for Code},
    year={2025},
    url={https://openreview.net/forum?id=izy1oaAOeX}
}
```