CLDataset (Extracted on 20-09-2020)

Russian dataset for natural language processing problems Each file contains a single JSON array containing articles information Object structure is as follow cyberleninka.txt: Full information crawled from CyberLeninka (2098 papers)

[
	{
		'keywords': [...],
		'fulltext': [...],
		'title': '...',
		'abstract': '...',
		'views': '...',
		'downloads': '...'
	},
	...
]

cyberleninka_ds.txt: Reduced information, for most keyphrase extraction tasks (2098 papers)

[
	{
		'keywords': [...],
		'fulltext': [...]
	},
	...
]

cyberleninka_long.txt: Reduced information, for most keyphrase extraction tasks. Only longer papers (longer than 10000 characters). (900 papers)

[
	{
		'keywords': [...],
		'fulltext': [...]
	},
	...
]

cyberleninka_short.txt: Reduced information, for most keyphrase extraction tasks. Only shorter papers (shorter than 10000 characters). (900 papers)

[
	{
		'keywords': [...],
		'fulltext': [...]
	},
	...
]