README.md

metadata

annotations_creators:
  - machine-generated
language_creators:
  - crowdsourced
  - expert-generated
language:
  - code
  - ru
  - en
license: other
multilinguality:
  - multilingual
pretty_name: GitVerse Code Dataset
size_categories:
  - 1M<n<10M
source_datasets:
  - original
task_categories:
  - text-generation
tags:
  - code
  - russian
configs:
  - config_name: default
    data_files:
      - split: train
        path: Data/*.parquet
    default: true
dataset_info:
  features:
    - name: file_text
      dtype: string
    - name: language
      dtype: string
    - name: file_name
      dtype: string

GitVerse Code Dataset

A comprehensive code dataset compiled from GitVerse, Russia's premier code hosting platform. This dataset is specifically designed to support training code models with strong Russian language understanding and authentic Russian coding practices.

---

Overview

The GitVerse Code Dataset represents a significant code corpus from a Russian-native platform, capturing both open-source and enterprise projects across 416 programming languages. It serves as a valuable resource for developing multilingual code understanding models tailored to Russian developers and organizations.

Key Statistics

Metric	Value
Total Files	2,802,994
Total Repositories	9,014
Compressed Size	2 GB (Parquet)
Programming Languages	416
File Format	Single Parquet file

---

Dataset Characteristics

Scope and Coverage

This dataset captures code from over 9,000 repositories hosted on GitVerse, including:

• Russian-centric content: Extensive coverage of code written by Russian developers, featuring Russian comments, documentation, and variable naming conventions
• Diverse language ecosystem: Support for 416 distinct programming languages
• Developer and enterprise projects: A mix of individual developer projects and Russian enterprise codebases
• Quality-assured: Curated to provide meaningful representation of Russian coding practices

Programming Languages

The dataset encompasses 416 languages. The 30 most represented languages by file count are:

Rank	Language	File Count
1	C	580,713
2	JavaScript	275,744
3	C++	197,896
4	Shell	166,527
5	Python	116,065
6	Markdown	112,811
7	TypeScript	107,867
8	Java	88,429
9	PHP	80,341
10	Makefile	77,619
11	XML	75,320
12	Go	69,155
13	C#	68,185
14	Text	65,677
15	JSON	64,253
16	SVG	58,107
17	HTML	43,261
18	YAML	40,178
19	Unity3D Asset	33,917
20	Rust	32,872
21	LLVM	29,819
22	Unix Assembly	27,672
23	Roff	25,884
24	CSS	21,809
25	TSX	21,637
26	reStructuredText	19,683
27	Perl	18,576
28	Gettext Catalog	17,071
29	Diff	14,225
30	CMake	14,132

---

Dataset Structure

Data Fields

Each record contains three fields providing content and metadata:

Field	Type	Description
file_text	string	Complete file content in UTF-8 encoding
language	string	Programming language identified using github-linguist
file_name	string	Unique file identifier within the dataset

Sample Record

{
    "file_text": "Процедура ОбработкаПроведения(Отказ, Режим)\n\t// Нерабочий вариант без ошибок\n...",
    "language": "1C Enterprise",
    "file_name": "004_work.code.bsl"
}

File Format

• Format: Apache Parquet
• Structure: Single consolidated file (data.parquet)
• Encoding: UTF-8
• Split: All examples are included in a single training split (no validation or test splits)

---

Data Creation Process

Language Detection Methodology

Programming languages are identified using github-linguist, GitHub's robust library for language detection and syntax highlighting. This ensures consistent and reliable language classification across all files in the dataset.

Source Data

All data originates from public repositories hosted on GitVerse, Russia's leading code hosting platform.

---

Usage Considerations

Data Privacy and Security

The dataset may contain sensitive information that requires careful handling:

• Email Addresses: Present in code comments, documentation, or configuration files
• Credentials: Accidentally committed API keys or authentication tokens
• Personal Information: Names, phone numbers, and other identifiable data in comments or documentation

Users should implement appropriate filtering and anonymization when preparing data for model training.

Licensing and Attribution

This dataset aggregates source code from repositories with diverse licenses. Any use of code or data derived from this dataset must comply with the original repository licenses, including attribution requirements where applicable.

Users are responsible for:

• Reviewing applicable license terms for code they utilize
• Providing proper attribution when required
• Ensuring compliance with license restrictions
• Respecting the rights of original authors

Responsible Use

This dataset should be used responsibly with consideration for:

• Original developer contributions and intellectual property rights
• License compliance for any derivative works
• Transparency about the source and composition of training data
• Ethical considerations in model training and deployment

---

Technical Details

Source: Public repositories hosted on GitVerse

Annotations: Machine-generated (language detection)

Multilingual Support: Includes multilingual code and documentation with emphasis on Russian content

Task Categories: Text generation, code modeling, language understanding

Tags: Code, Russian language, multilingual

---