Hugging Face's logo

Datasets:
LeeSek
/
dockerfiles-linted

Dataset card Data Studio Files
xet
Community
dockerfiles-linted
File size: 3,344 Bytes
6e7537c
 
 
 
 
 
 
 
 
 
 
 
 
 
c17cab9
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
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
 
---
datasets:
  - name: dockerfiles-linted
    license: cc-by-4.0
    language:
      - en
    tags:
      - dockerfile
      - hadolint
      - devops
      - github
      - software-engineering
---

# Dockerfiles Linted Dataset

This dataset contains 195,758 Dockerfiles collected from public container images found on Docker Hub. Each Dockerfile is enriched with metadata and statically analyzed using [Hadolint](https://github.com/hadolint/hadolint).

## Dataset Overview

Dockerfiles were collected by:
1. Enumerating Docker Hub images,
2. Extracting raw GitHub URLs pointing to Dockerfile locations in those repositories,
3. Downloading each file and extracting content-level metadata (e.g. line count, validity),
4. Fetching GitHub-level metadata such as license, stars, and path,
5. Running Hadolint to detect Dockerfile best-practice violations.

Each data entry includes:
- Dockerfile content and metadata (hash, lines, valid),
- GitHub repository context (repo URL, license, stars, branch),
- Hadolint static analysis results.

## Structure

The dataset is split into 4 progressive stages:

```
resolved/       → Resolved image ↔ raw Dockerfile URL mappings
enriched/       → Dockerfile content + metadata (hash, lines, valid)
full/           → Enriched with GitHub repository metadata
linted/         → Final stage with Hadolint static analysis results
summary/        → Visual summaries and statistics
```

Each folder contains:
- 20 partial `.jsonl` chunks (10k entries each),
- 1 combined `*_all.jsonl` file for bulk use,
- a `dataset_info.json` describing the data structure.

## Bias Warning: CircleCI Dominance

A very large portion of the Dockerfiles originate from the `circleci/circleci-dockerfiles` repository (170,000+ entries).  
To ensure clarity, we include **two sets of visualizations**:

- **Full dataset statistics**
- **Statistics with CircleCI excluded**

This allows users to explore data either globally or more evenly across repositories.

## Repository Distribution

**Top 20 repositories by number of Dockerfiles (entire dataset):**

![Top 20 Repositories](summary/top_repositories.png)

**Top 20 repositories (excluding CircleCI):**

![Top 20 Repositories (No CircleCI)](summary/top_repositories_no_circleci.png)

## Hadolint Rule Violations

**Top 20 Hadolint rules triggered (entire dataset):**

![Top 20 Hadolint Rules](summary/top_hadolint_rules.png)

**Top 20 Hadolint rules (excluding CircleCI):**

![Top 20 Hadolint Rules (No CircleCI)](summary/top_hadolint_rules_no_circleci.png)

## Example Entry

Each row in the `.jsonl` files contains enriched information about a Dockerfile.  
The full field structure for each stage is documented in its respective `dataset_info.json`.

## License

This dataset is provided under the [Creative Commons Attribution 4.0 International (CC BY 4.0)](https://creativecommons.org/licenses/by/4.0/) license.

## Attribution Notice

This dataset contains Dockerfiles originally created by developers across many public GitHub repositories.  
We do not claim ownership of any individual Dockerfile.  
Instead, we provide this dataset as a collection of metadata, structure, and analysis, compiled and curated for research and educational use.

Please consult individual repository licenses if you plan to reuse specific Dockerfile content in derived works.