strickvl commited on
Commit
4201f53
1 Parent(s): 040699d

markdown tags

Browse files
Files changed (1) hide show
  1. README.md +26 -24
README.md CHANGED
@@ -48,6 +48,7 @@ The dataset is only in English, but the original source articles were in Dari an
48
  ### Data Instances
49
  Here's a JSON example from the main "articles" table:
50
 
 
51
  {
52
  "id": "97",
53
  "paper": "142",
@@ -74,36 +75,37 @@ Here's a JSON example from the main "articles" table:
74
  "datetranslation": "",
75
  "scan": ""
76
  }
 
77
 
78
  The dataset consists of several supporting tables that are referenced in the main "articles" table, such as papers, article_tags, bib_books, cities, current_events, ethnics, glossary, historical_events, issue, languages, organisations, people, provinces, region_tree, renderbackgrounder, top_topcs, topic, and types.
79
 
80
  The dataset as a whole consists of 7990 articles that were translated during the period Afghanwire was open as an organisation.
81
 
82
  ### Data Fields
83
- - id - basic id for the article
84
- - paper - id/number for a paper mentioned in papers.parquet
85
- - title - article title
86
- - date
87
- - author (if present)
88
- - translator (who translated the article)
89
- - topic (associated with topic table)
90
- - abstract (sometimes mentions the translator)
91
- - comment (sometimes also mentions the translator)
92
- - rating (not always used. was a measure for interest level)
93
- - language (associated with the separate table)
94
- - type (associated with the article types table)
95
- - article - the full translation
96
- - ok - whether the translation has been edited
97
- - no_newsletter - a metatag to represent whether the article should be sent out as part of our newsletter or not
98
- - eingegeben - a unix timestamp for when the article was uploaded to the database
99
- - newsletter - whether to include the article in our newsletter
100
- - free - whether to make the article available for free or not
101
- - url - if available
102
- - top_topic - what high-level topic the article was associated with
103
- - words - word count (not always present)
104
- - translatorcomment - not always present
105
- - datetranslation - not always present
106
- - scan - whether there's a scan for the article or not
107
 
108
  ### Data Splits
109
  There are no predefined splits. The dataset is provided as a single large collection.
 
48
  ### Data Instances
49
  Here's a JSON example from the main "articles" table:
50
 
51
+ ```json
52
  {
53
  "id": "97",
54
  "paper": "142",
 
75
  "datetranslation": "",
76
  "scan": ""
77
  }
78
+ ```
79
 
80
  The dataset consists of several supporting tables that are referenced in the main "articles" table, such as papers, article_tags, bib_books, cities, current_events, ethnics, glossary, historical_events, issue, languages, organisations, people, provinces, region_tree, renderbackgrounder, top_topcs, topic, and types.
81
 
82
  The dataset as a whole consists of 7990 articles that were translated during the period Afghanwire was open as an organisation.
83
 
84
  ### Data Fields
85
+ - `id` - basic id for the article
86
+ - `paper` - id/number for a paper mentioned in papers.parquet
87
+ - `title` - article title
88
+ - `date`
89
+ - `author` (if present)
90
+ - `translator` (who translated the article)
91
+ - `topic` (associated with topic table)
92
+ - `abstract` (sometimes mentions the translator)
93
+ - `comment` (sometimes also mentions the translator)
94
+ - `rating` (not always used. was a measure for interest level)
95
+ - `language` (associated with the separate table)
96
+ - `type` (associated with the article types table)
97
+ - `article` - the full translation
98
+ - `ok` - whether the translation has been edited
99
+ - `no_newsletter` - a metatag to represent whether the article should be sent out as part of our newsletter or not
100
+ - `eingegeben` - a unix timestamp for when the article was uploaded to the database
101
+ - `newsletter` - whether to include the article in our newsletter
102
+ - `free` - whether to make the article available for free or not
103
+ - `url` - if available
104
+ - `top_topic` - what high-level topic the article was associated with
105
+ - `words` - word count (not always present)
106
+ - `translatorcomment` - not always present
107
+ - `datetranslation` - not always present
108
+ - `scan` - whether there's a scan for the article or not
109
 
110
  ### Data Splits
111
  There are no predefined splits. The dataset is provided as a single large collection.