Datasets:
Update README.md
Browse files
README.md
CHANGED
@@ -12,6 +12,8 @@ size_categories:
|
|
12 |
# Overview
|
13 |
This dataset provides a convenient and user-friendly format of data from [Aozora Bunko (青空文庫)](https://www.aozora.gr.jp/), a website that compiles public-domain books in Japan, ideal for Machine Learning applications.
|
14 |
|
|
|
|
|
15 |
# Methodology
|
16 |
|
17 |
The code to reproduce this dataset is made available on GitHub: [globis-org/aozorabunko-exctractor](https://github.com/globis-org/aozorabunko-extractor).
|
@@ -41,5 +43,41 @@ The data in the `text` field was then cleaned in the following sequence:
|
|
41 |
|
42 |
If you prefer to employ only modern Japanese, you can filter entries with: `row["meta"]["文字遣い種別"] == "新字新仮名"`.
|
43 |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
44 |
# License
|
45 |
CC BY 4.0
|
|
|
12 |
# Overview
|
13 |
This dataset provides a convenient and user-friendly format of data from [Aozora Bunko (青空文庫)](https://www.aozora.gr.jp/), a website that compiles public-domain books in Japan, ideal for Machine Learning applications.
|
14 |
|
15 |
+
[For Japanese] 日本語での概要説明を Qiita に記載しました: https://qiita.com/akeyhero/items/b53eae1c0bc4d54e321f
|
16 |
+
|
17 |
# Methodology
|
18 |
|
19 |
The code to reproduce this dataset is made available on GitHub: [globis-org/aozorabunko-exctractor](https://github.com/globis-org/aozorabunko-extractor).
|
|
|
43 |
|
44 |
If you prefer to employ only modern Japanese, you can filter entries with: `row["meta"]["文字遣い種別"] == "新字新仮名"`.
|
45 |
|
46 |
+
# Example
|
47 |
+
|
48 |
+
```py
|
49 |
+
>>> from datasets import load_dataset
|
50 |
+
>>> ds = load_dataset('globis-university/aozorabunko-clean')
|
51 |
+
>>> ds
|
52 |
+
DatasetDict({
|
53 |
+
train: Dataset({
|
54 |
+
features: ['text', 'footnote', 'meta'],
|
55 |
+
num_rows: 16951
|
56 |
+
})
|
57 |
+
})
|
58 |
+
>>> ds = ds.filter(lambda row: row['meta']['文字遣い種別'] == '新字新仮名') # only modern Japanese
|
59 |
+
>>> ds
|
60 |
+
DatasetDict({
|
61 |
+
train: Dataset({
|
62 |
+
features: ['text', 'footnote', 'meta'],
|
63 |
+
num_rows: 10246
|
64 |
+
})
|
65 |
+
})
|
66 |
+
>>> book = ds['train'][0] # one of works
|
67 |
+
>>> book['meta']['作品名']
|
68 |
+
'ウェストミンスター寺院'
|
69 |
+
>>> text = book['text'] # main content
|
70 |
+
>>> len(text)
|
71 |
+
10639
|
72 |
+
>>> print(text[:100])
|
73 |
+
深いおどろきにうたれて、
|
74 |
+
名高いウェストミンスターに
|
75 |
+
真鍮や石の記念碑となって
|
76 |
+
すべての王侯貴族が集まっているのをみれば、
|
77 |
+
今はさげすみも、ほこりも、見栄もない。
|
78 |
+
善にかえった貴人の姿、
|
79 |
+
華美と俗世の
|
80 |
+
```
|
81 |
+
|
82 |
# License
|
83 |
CC BY 4.0
|