irlspbru/RusLawOD
The Russian Legislative Corpus, 1991–2026 Russian primary and secondary legislation corpus covering laws of Russian Federation, decrees by the President of RF, regulations by the government published as of July, 2026. The corpus collects all 308,056 texts (198,777,737 tokens) of non-secret federal regulations and acts, along with their metadata. The corpus has two versions: the original text with minimal preprocessing and a version prepared for linguistic analysis with… See the full description on the dataset page: https://huggingface.co/datasets/irlspbru/RusLawOD.
The Russian Legislative Corpus, 1991–2026
Russian primary and secondary legislation corpus covering laws of Russian Federation, decrees by the President of RF, regulations by the government published as of July, 2026. The corpus collects all 308,056 texts (198,777,737 tokens) of non-secret federal regulations and acts, along with their metadata. The corpus has two versions: the original text with minimal preprocessing and a version prepared for linguistic analysis with morphosyntactic markup.
For lemmatization, POS tagging, and dependency parsing we use the Ru-syntax tool developed by the Computational Linguistics team at the Higher School of Economics (Russia). This tool gathers the results from the morphological analyzer MyStem, part-of-speech tagging (Schmid, 2013) TreeTagger, and dependency grammar analyzer MaltParser. Finally, the result is stored in the CONLL-U format.
Structure
The corpus is stored in XML files each for one document. Every field appears if only the document has that kind of information. Fields descriptions are in the example below.
<act> <!-- Legal act as the type of a document -->
<meta> <!-- major sections are metadata and text -->
<identification> <!-- see the Limitations on information about legal act identification in Russia -->
<pravogovruNd val="000000000" /> <!-- document internal number at the IPS Zakonodatelstvo
website at the moment of download. It may be subject to change -->
<issuedByIPS val="Entity that issued the act according to the IPS Zakonodatelstvo" />
<docdateIPS val="00.00.0000" /> <!-- document day of signature according to the IPS
Zakonodatelstvo, date format is dd.mm.yyyy-->
<docNumberIPS val="000" /> <!-- document number at signature according to the IPS Zakonodatelstvo -->
<headingIPS>title of the document in the IPS Zakonodatelstvo.</headingIPS>
<doc_typeIPS val="Document type as was in the source"/>
<doc_author_normal_formIPS val="State organ that adopted the act, in normal language form"/>
<signedIPS val="______"/> <!-- Person name who signed this legal act as provided in the source -->
<statusIPS val="Утратил силу"/> <!-- In force, Not in force, In force with amendments: Acting status at the date of scrapping and as it was provided by the source -->
<actual_datetimeIPS val="1710792705.7460072"/> <!-- Date and time when this data was scrapped from the original website -->
<actual_datetime_humanIPS val="Mon Mar 18 23:11:45 2024"/> <!-- Date and time when this data was scrapped from the original website, in human readable format -->
<is_widely_used val="1"/></identification> <!-- 1 if yes, 0 if no: is the document normative and in wide use (see article preprint for the details) -->
</identification>
<references>
<classifierByIPS val="000.000.000.000.000" /> <!-- classification code according to the IPS Zakonodatelstvo -->
</references>
<keywords>
<keywordByIPS val="KEYWORD" />
</keywords>
</meta>
<body>
<textIPS><-- Text parsed from the IPS Zakonodatelstvo -->
<!-- It can include hyperlinks to other acts, mostly amendments,
like this: --> text <ref>linked text</ref> text
</textIPS>
<taggedTextIPS> <-- CONLL_U morphosyntactic tagged text, cleaned -->
</taggedTextIPS>
</body>
</act>
</act>Metadata
Collection
We queried the online Legislation of Russia service (pravo.gov.ru) base called "IPS Zakonodatelstvo Rossii" maintained by the Special Communications Service for the texts legislation and their metadata. Our web scraping took place in the period spanning 2017—2026 with the last data collection taking part in August, 2026.
Usage
This dataset could be loaded in Python with HF's Datasets library:
from datasets import load_dataset
dataset = load_dataset(<path-to-parquet-files>)
dataset['train'][0]['textIPS']. # first chunk of main texts variableFor more information please see Datasets' documentation.
The dataset is published also at GitHub
Citation
If you use this corpus in your scientific work, please refer to: Saveliev, Denis, and Ruslan Kuchakov. "The Russian Legislative Corpus." arXiv preprint arXiv:2406.04855 (2024).
@ruslaw{saveliev2024russian,
title={The {R}ussian {L}egislative {C}orpus},
author={{S}aveliev, {D}enis and {K}uchakov, {R}uslan},
journal={arXiv preprint arXiv:2406.04855},
year={2024}
}