Datasets:
Tasks:
Text Classification
Modalities:
Text
Formats:
json
Sub-tasks:
multi-label-classification
Languages:
English
Size:
1K - 10K
DOI:
Create README.md
Browse files
README.md
ADDED
@@ -0,0 +1,110 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
1 |
+
# Dataset Card for English quotes
|
2 |
+
#### Dataset Summary
|
3 |
+
english_quotes is a dataset of English quotes from [goodreads quotes](https://www.goodreads.com/quotes). This dataset can be used for multi-label text classification and text generation. The content of each quote is in English and concerns the domain of datasets for NLP and beyond.
|
4 |
+
|
5 |
+
#### Supported Tasks and Leaderboards
|
6 |
+
- Multi-label text classification : The dataset can be used to train a model for text-classification, which consists of classifying quotes by author as well as by topic (using tags). Success on this task is typically measured by achieving a high or low accuracy.
|
7 |
+
- Text-generation : The dataset can be used to train a model to generate quotes by fine-tuning an existing pretrained model on the corpus composed of all quotes (or quotes by author).
|
8 |
+
|
9 |
+
#### Languages
|
10 |
+
The texts in the dataset are in English (en).
|
11 |
+
|
12 |
+
# Dataset Structure
|
13 |
+
#### Data Instances
|
14 |
+
A JSON-formatted example of a typical instance in the dataset:
|
15 |
+
```python
|
16 |
+
{'author': 'Ralph Waldo Emerson',
|
17 |
+
'quote': '“To be yourself in a world that is constantly trying to make you something else is the greatest accomplishment.”',
|
18 |
+
'tags': ['accomplishment', 'be-yourself', 'conformity', 'individuality']}
|
19 |
+
```
|
20 |
+
#### Data Fields
|
21 |
+
- **author** : The author of the quote.
|
22 |
+
- **quote** : The text of the quote.
|
23 |
+
- **tags**: The tags could be characterized as topics around the quote.
|
24 |
+
|
25 |
+
#### Data Splits
|
26 |
+
I kept the dataset as one block (train), so it can be shuffled and split by users later using methods of the hugging face dataset library like the (.train_test_split()) method.
|
27 |
+
|
28 |
+
# Dataset Creation
|
29 |
+
#### Curation Rationale
|
30 |
+
I want to share my datasets (created by web scraping and additional cleaning treatments) with the HuggingFace community so that they can use them in NLP tasks to advance artificial intelligence.
|
31 |
+
|
32 |
+
#### Source Data
|
33 |
+
The source of Data is [goodreads](https://www.goodreads.com/?ref=nav_home) site: from [goodreads quotes](https://www.goodreads.com/quotes)
|
34 |
+
|
35 |
+
#### Initial Data Collection and Normalization
|
36 |
+
|
37 |
+
The data collection process is web scraping using BeautifulSoup and Requests libraries.
|
38 |
+
The data is slightly modified after the web scraping: removing all quotes with "None" tags, and the tag "attributed-no-source" is removed from all tags, because it has not added value to the topic of the quote.
|
39 |
+
|
40 |
+
#### Who are the source Data producers ?
|
41 |
+
The data is machine-generated (using web scraping) and subjected to human additional treatment.
|
42 |
+
|
43 |
+
below, I provide the script I created to scrape the data (as well as my additional treatment):
|
44 |
+
```python
|
45 |
+
import requests
|
46 |
+
from bs4 import BeautifulSoup
|
47 |
+
import pandas as pd
|
48 |
+
import json
|
49 |
+
from collections import OrderedDict
|
50 |
+
|
51 |
+
page = requests.get('https://www.goodreads.com/quotes')
|
52 |
+
if page.status_code == 200:
|
53 |
+
pageParsed = BeautifulSoup(page.content, 'html5lib')
|
54 |
+
|
55 |
+
# Define a function that retrieves information about each HTML quote code in a dictionary form.
|
56 |
+
def extract_data_quote(quote_html):
|
57 |
+
quote = quote_html.find('div',{'class':'quoteText'}).get_text().strip().split('\n')[0]
|
58 |
+
author = quote_html.find('span',{'class':'authorOrTitle'}).get_text().strip()
|
59 |
+
if quote_html.find('div',{'class':'greyText smallText left'}) is not None:
|
60 |
+
tags_list = [tag.get_text() for tag in quote_html.find('div',{'class':'greyText smallText left'}).find_all('a')]
|
61 |
+
tags = list(OrderedDict.fromkeys(tags_list))
|
62 |
+
if 'attributed-no-source' in tags:
|
63 |
+
tags.remove('attributed-no-source')
|
64 |
+
else:
|
65 |
+
tags = None
|
66 |
+
data = {'quote':quote, 'author':author, 'tags':tags}
|
67 |
+
return data
|
68 |
+
|
69 |
+
# Define a function that retrieves all the quotes on a single page.
|
70 |
+
def get_quotes_data(page_url):
|
71 |
+
page = requests.get(page_url)
|
72 |
+
if page.status_code == 200:
|
73 |
+
pageParsed = BeautifulSoup(page.content, 'html5lib')
|
74 |
+
quotes_html_page = pageParsed.find_all('div',{'class':'quoteDetails'})
|
75 |
+
return [extract_data_quote(quote_html) for quote_html in quotes_html_page]
|
76 |
+
|
77 |
+
# Retrieve data from the first page.
|
78 |
+
data = get_quotes_data('https://www.goodreads.com/quotes')
|
79 |
+
|
80 |
+
# Retrieve data from all pages.
|
81 |
+
for i in range(2,101):
|
82 |
+
print(i)
|
83 |
+
url = f'https://www.goodreads.com/quotes?page={i}'
|
84 |
+
data_current_page = get_quotes_data(url)
|
85 |
+
if data_current_page is None:
|
86 |
+
continue
|
87 |
+
data = data + data_current_page
|
88 |
+
|
89 |
+
data_df = pd.DataFrame.from_dict(data)
|
90 |
+
for i, row in data_df.iterrows():
|
91 |
+
if row['tags'] is None:
|
92 |
+
data_df = data_df.drop(i)
|
93 |
+
# Produce the data in a JSON format.
|
94 |
+
data_df.to_json('C:/Users/Abir/Desktop/quotes.jsonl',orient="records", lines =True,force_ascii=False)
|
95 |
+
# Then I used the familiar process to push it to the Hugging Face hub.
|
96 |
+
|
97 |
+
```
|
98 |
+
#### Annotations
|
99 |
+
Annotations are part of the initial data collection (see the script above).
|
100 |
+
|
101 |
+
# Additional Information
|
102 |
+
#### Dataset Curators
|
103 |
+
Abir ELTAIEF :
|
104 |
+
https://tn.linkedin.com/in/abir-eltaief-pmp%C2%AE-469048115
|
105 |
+
|
106 |
+
#### Licensing Information
|
107 |
+
This work is licensed under a Creative Commons Attribution 4.0 International License (all software and libraries used for web scraping are made available under this Creative Commons Attribution license).
|
108 |
+
|
109 |
+
#### Contributions
|
110 |
+
Thanks to @Abirate for adding this dataset.
|