mch-dd commited on
Commit
b489399
1 Parent(s): 2f3f0ba

Update README.md

Browse files
Files changed (1) hide show
  1. README.md +62 -15
README.md CHANGED
@@ -5,32 +5,79 @@ task_categories:
5
  pretty_name: GATT Library
6
  ---
7
 
8
- # GATT Library
9
 
10
- **GATT Library**
11
 
12
- Total number of words: 133196586
13
 
14
- Average number of words: 2213.23
 
15
 
16
- Total number of characters: 941277211
 
 
 
 
 
 
 
17
 
18
- Average number of characters: 15640.51
19
 
20
- Date range: 1946-01-01 to 1996-09-06
21
 
22
- **Languages on a 10000 entries sample:**
23
 
24
- en 7878
25
 
26
- fr 1357
27
 
28
- Unknown 486
29
 
30
- es 261
31
 
32
- ca 12
 
 
 
 
 
 
 
 
 
 
 
 
 
33
 
34
- pt 4
35
 
36
- de 2
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
5
  pretty_name: GATT Library
6
  ---
7
 
8
+ # GATT Library Dataset Card
9
 
10
+ ## Dataset Overview
11
 
12
+ **Dataset Name:** GATT Library
13
 
14
+ **Description:**
15
+ The GATT Library dataset comprises a comprehensive collection of documents related to the General Agreement on Tariffs and Trade (GATT), spanning from January 1, 1946, to September 6, 1996. This dataset is organized into a single Parquet file, which contains detailed information about the documents, including metadata extracted from the original files. The original files are stored in a tar.gz archive, organized into folders.
16
 
17
+ **Languages Covered (based on a sample of 10,000 entries):**
18
+ - English (en): 7,878 documents
19
+ - French (fr): 1,357 documents
20
+ - Unknown (Unknown): 486 documents
21
+ - Spanish (es): 261 documents
22
+ - Catalan (ca): 12 documents
23
+ - Portuguese (pt): 4 documents
24
+ - German (de): 2 documents
25
 
26
+ **Total Number of Words:** 133,196,586
27
 
28
+ **Average Number of Words per Document:** 2,213.23
29
 
30
+ **Total Number of Characters:** 941,277,211
31
 
32
+ **Average Number of Characters per Document:** 15,640.51
33
 
34
+ **Date Range:** January 1, 1946 to September 6, 1996
35
 
36
+ ## Dataset Structure
37
 
38
+ The dataset is stored in a single Parquet file, which contains the following metadata for each document:
39
 
40
+ - `source`: The source of the document.
41
+ - `document_id`: A unique identifier for the document.
42
+ - `title`: The title of the document.
43
+ - `short_title`: A shorter version of the document title.
44
+ - `author`: The author of the document.
45
+ - `date`: The publication date of the document.
46
+ - `type_of_document`: The type of document.
47
+ - `identifier`: An identifier for the document.
48
+ - `link`: A link to the document.
49
+ - `file`: The file name.
50
+ - `folder`: The folder where the document is stored.
51
+ - `word_count`: Total number of words in the document.
52
+ - `character_count`: Total number of characters in the document.
53
+ - `text`: Full text content of the document.
54
 
55
+ Each Parquet file is accompanied by the original XML files collected into tar.gz files, one for each Parquet file.
56
 
57
+ ## Usage
58
+
59
+ The dataset can be used for various purposes including, but not limited to:
60
+
61
+ - Analyzing historical trade documents and agreements.
62
+ - Studying the evolution and impact of GATT on international trade.
63
+ - Developing natural language processing models to analyze the content of trade documents.
64
+ - Extracting insights related to trade policies and their implementation over time.
65
+
66
+ ## Source
67
+
68
+ The dataset is collected from the General Agreement on Tariffs and Trade (GATT) archive. The original files are processed and converted into Parquet format for efficient storage and analysis.
69
+
70
+ ## Dataset Citation
71
+
72
+ If you use this dataset in your research, please cite it as follows:
73
+
74
+ ```
75
+ @dataset{GATT_Library_2024,
76
+ title={GATT Library Dataset},
77
+ author={Pleias},
78
+ year={2024},
79
+ description={Collection of GATT-related documents, organized by date and language, with detailed metadata extracted from the original files.}
80
+ }
81
+ ```
82
+
83
+ **Note:** The dataset is presented and maintained by Pleias. All rights reserved.