Hugging Face's logo

Datasets:
Skilln
/
devops-qa-dataset

Tasks:
Question Answering
Languages:
English
Size:
10K<n<100K
Tags:
devops
kubernetes
docker
infrastructure-as-code
License:
Dataset card Files
xet
 Community
Skilln commited on
Commit
040417f
·
verified ·
1 Parent(s): 1109a18

Upload README.md with huggingface_hub

Browse files
Files changed (1) hide show
  1. README.md +44 -46
README.md CHANGED
@@ -13,53 +13,51 @@ size_categories:
13
  - 1K<n<10K
14
  ---
15
 
16
- # DevOps Q&A Dataset Summary
17
-
18
- **Generated on:** 2026-01-17
19
- **Version:** 2.1 Final
20
-
21
- ## 📊 Overview
22
- Based on user requirements, we have collected, validated, and deduplicated a high-quality dataset of DevOps Questions & Answers.
23
-
24
- | Metric | Value |
25
- |:--- |:--- |
26
- | **Total Unique Items** | **7,219** |
27
- | **Primary Source** | GitHub Repositories (Advanced Search) |
28
- | **Secondary Source** | Official Documentation (Docker, K8s, etc.) |
29
- | **Average Quality Score** | ~0.81 |
30
- | **Deduplication** | Exact (MD5) + Semantic (BGE-Small) |
31
-
32
- ## 📁 File Manifest
33
- All files are located in `/opt/Data set/output/`:
34
-
35
- * `devops_dataset.jsonl` - Main dataset (JSON Lines).
36
- * `devops_dataset.parquet` - Optimized Parquet format for pandas/polars.
37
- * `dataset_metadata.json` - Technical metadata (last run stats).
38
-
39
- ## 🧩 Data Composition
40
- The dataset covers a wide range of DevOps topics, including but not limited to:
41
- * **Kubernetes:** Troubleshooting, Configs, Best Practices.
42
- * **Docker:** Composition, Optimization, Security.
43
- * **IaC:** Terraform, Ansible, Pulumi.
44
- * **CI/CD:** GitHub Actions, GitLab CI, Jenkins.
45
- * **Monitoring:** Prometheus, Grafana, ELK.
46
 
47
- ## 🚀 Usage Example (Python)
48
  ```python
49
- import pandas as pd
50
-
51
- # Load the dataset
52
- df = pd.read_parquet('/opt/Data set/output/devops_dataset.parquet')
53
-
54
- # Inspect
55
- print(df.head())
56
- print(df['source'].value_counts())
57
-
58
- # Filter high quality
59
- high_quality = df[df['quality_score'] > 0.85]
60
  ```
61
 
62
- ## 🛠 Collection Details
63
- * **Tools:** `devops_collector.py` (Custom Python Collector)
64
- * **Validation:** Strict quality rules (length, spam, content safety).
65
- * **Dedup:** Semantic embedding filtering enabled for new data.
 
 
 
 
 
 
 
 
 
13
  - 1K<n<10K
14
  ---
15
 
16
+ # DevOps Q&A Dataset v1.0
17
+
18
+ ## Overview
19
+ High-quality dataset of **7,219** DevOps technical examples collected from GitHub repositories and official documentation.
20
+
21
+ ## Statistics
22
+ - **Total examples:** 7,219
23
+ - **Average quality score:** 0.81
24
+ - **Unique (semantic dedup):** 96%
25
+ - **Categories:** Docker, Kubernetes, CI/CD, Cloud, Linux, Terraform, Ansible
26
+ - **Sources:** GitHub Repositories (96.6%), Official Documentation (3.4%)
27
+
28
+ ## Use Cases
29
+ - Fine-tuning LLMs for DevOps automation
30
+ - Training specialized models for technical documentation
31
+ - DevOps support chatbot development
32
+ - Infrastructure-as-Code assistant
33
+
34
+ ## Dataset Schema
35
+ ```json
36
+ {
37
+ "question": "How to fix OOMKilled errors in Kubernetes?",
38
+ "answer": "1. Check pod memory limits...",
39
+ "category": "kubernetes",
40
+ "difficulty": "intermediate",
41
+ "quality_score": 0.89,
42
+ "source": "github"
43
+ }
44
+ ```
 
45
 
46
+ ## Download
47
  ```python
48
+ from datasets import load_dataset
49
+ dataset = load_dataset("Skilln/devops-qa-dataset")
 
 
 
 
 
 
 
 
 
50
  ```
51
 
52
+ ## License
53
+ CC-BY-SA 4.0 (compatible with source licenses)
54
+
55
+ ## Citation
56
+ ```text
57
+ @dataset{devops_qa_2026,
58
+ title={DevOps Q&A Dataset},
59
+ author={Skilln},
60
+ year={2026},
61
+ url={https://huggingface.co/datasets/Skilln/devops-qa-dataset}
62
+ }
63
+ ```