hannah-eee commited on
Commit
5790e09
·
verified ·
1 Parent(s): 65f8f14

Update README.md

Browse files
Files changed (1) hide show
  1. README.md +128 -1
README.md CHANGED
@@ -1,3 +1,130 @@
1
  ---
 
2
  license: cc-by-sa-3.0
3
- ---
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
  ---
2
+
3
  license: cc-by-sa-3.0
4
+ language:
5
+ - en
6
+ - de
7
+ - fr
8
+ - es
9
+ - it
10
+ - ja
11
+ - ru
12
+ - zh
13
+ - ... (all languages present in the Arch Wiki)
14
+ pretty_name: Arch Wiki - Complete Offline Dump
15
+ size_categories:
16
+ - 100K-1M
17
+ task_categories:
18
+ - document-retrieval
19
+ - question-answering
20
+ - text-generation
21
+ tags:
22
+ - arch-linux
23
+ - wiki
24
+ - documentation
25
+ - rag
26
+ - offline
27
+
28
+ # Arch Wiki - JSONL Dataset
29
+
30
+ This dataset provides the complete Arch Linux Wiki in a clean, machine‑readable JSON Lines format (.jsonl).
31
+ It is derived from the official arch-wiki-docs package, which contains static HTML exports of all language versions.
32
+
33
+ Each line is one wiki page, with extracted plain text, headings, language code, and a relative URL path – ready for RAG (Retrieval-Augmented Generation), embedding, full‑text search, or offline browsing.
34
+
35
+ ## Dataset Structure
36
+
37
+ One JSON object per line. Example:
38
+
39
+ {
40
+ "lang": "en",
41
+ "title": "Arch Linux",
42
+ "url_path": "en/Arch_Linux.html",
43
+ "content": "Arch Linux is an independently developed, x86-64 general-purpose Linux distribution that strives to provide the latest stable versions of most software...",
44
+ "headings": ["Arch Linux", "History", "Installation", "Package management"]
45
+ }
46
+
47
+ | Field | Type | Description |
48
+ |------------|--------------|-------------|
49
+ | lang | string | Two‑letter language code (e.g., en, fr, de). Chinese variants are zh-hans/zh-hant. |
50
+ | title | string | Page title as displayed in the wiki (without “ - ArchWiki” suffix). |
51
+ | url_path | string | Path relative to the root of the offline dump (e.g., en/Arch_Linux.html). |
52
+ | content | string | Clean plain‑text version of the article. All HTML tags, edit links, tables of contents, and navigation elements have been removed. |
53
+ | headings | list[string] | All h1, h2, h3 headings found on the page, in order of appearance. |
54
+
55
+ ## Statistics
56
+
57
+ - Total pages: 5,713 (may vary slightly with updates)
58
+ - Languages: All languages served by the official Arch Wiki (determined by the arch-wiki-docs package)
59
+ - Average content length: ~2,500 characters per page
60
+ - File size (compressed): ~50‑80 MB (.7z or .gz)
61
+ - File size (uncompressed JSONL): ~200‑300 MB
62
+
63
+ ## Usage
64
+
65
+ Load with Hugging Face datasets:
66
+
67
+ from datasets import load_dataset
68
+
69
+ dataset = load_dataset("your-username/arch-wiki-jsonl", split="train")
70
+
71
+ # Examine a page
72
+ print(dataset[0]["title"])
73
+ print(dataset[0]["content"][:500])
74
+
75
+ Load from raw JSONL (pure Python):
76
+
77
+ import json
78
+
79
+ pages = []
80
+ with open("arch_wiki.jsonl", "r", encoding="utf-8") as f:
81
+ for line in f:
82
+ pages.append(json.loads(line))
83
+
84
+ # Filter by language
85
+ en_pages = [p for p in pages if p["lang"] == "en"]
86
+
87
+ Build a simple RAG pipeline (example with LangChain):
88
+
89
+ from langchain.document_loaders import JSONLoader
90
+ from langchain.text_splitter import RecursiveCharacterTextSplitter
91
+
92
+ loader = JSONLoader(
93
+ file_path="arch_wiki.jsonl",
94
+ jq_schema=".content",
95
+ text_content=False
96
+ )
97
+ documents = loader.load()
98
+ text_splitter = RecursiveCharacterTextSplitter(chunk_size=500, chunk_overlap=50)
99
+ chunks = text_splitter.split_documents(documents)
100
+ # → now embed and index
101
+
102
+ ## Licenses and Attribution
103
+
104
+ The content of the Arch Wiki is licensed under the GNU Free Documentation License 1.3 and Creative Commons Attribution-ShareAlike 3.0 Unported (CC BY-SA 3.0).
105
+
106
+ This dataset is a derivative work. Full attribution to the Arch Linux community and all wiki contributors is required. When using this dataset, please include:
107
+
108
+ > Based on the Arch Linux Wiki (https://wiki.archlinux.org/), available under CC BY‑SA 3.0 and GFDL.
109
+
110
+ The conversion script is provided under the MIT License.
111
+
112
+ ## Generation Process
113
+
114
+ 1. Installed the official arch-wiki-docs package:
115
+ sudo pacman -S arch-wiki-docs
116
+ 2. Wrote a Python script using BeautifulSoup4 to parse all HTML files.
117
+ 3. Extracted: title (from title tag), clean text (from #mw-content-text), headings (from h1-h3).
118
+ 4. Output as JSON Lines (.jsonl) with one page per line.
119
+
120
+ You can find the conversion script in the scripts/ folder of this repository (or linked in the dataset card).
121
+
122
+ ## Changelog
123
+
124
+ - 2026-05-27: Initial release based on arch-wiki-docs version as of May 27 2026.
125
+
126
+ ## Notes
127
+
128
+ - This dataset is static. For the very latest wiki content, regenerate from an updated arch-wiki-docs package or pull from the live wiki.
129
+ - If you need the original HTML (e.g., for full styling), install arch-wiki-docs directly on an Arch Linux system.
130
+ - The headings field is useful for document structure‑aware chunking (e.g., split by heading boundaries).