vivek-dodia/mikrotik-docs

MikroTik Technical Documentation Dataset

Overview

A structured dataset containing MikroTik's technical documentation, prepared for LLM fine-tuning. The dataset preserves the hierarchical structure of the original documentation while maintaining technical accuracy and formatting.

Dataset Statistics

• —Total documents: 285
• —Maximum sections per document: 91
• —Average sections per document: 12.3
• —Format: Parquet

Data Structure

Each row represents a complete documentation page with the following fields:

Metadata

• —doc_title: Document title
• —doc_section: Documentation category (e.g., RouterOS, Hardware)
• —source_url: Original documentation URL
• —section_count: Number of sections in document

Content Fields (per section)

For each section (1 to n):

• —section_{n}_title: Section heading
• —section_{n}_content: Section content including code blocks
• —section_{n}_level: Heading level (1 = main heading, 2 = subheading, etc.)

Content Characteristics

• —Preserved code blocks with syntax
• —Maintained command examples and outputs
• —Intact technical formatting
• —Complete command parameters and descriptions
• —Original markdown structure

Typical Document Structure

1. 1.Summary/Overview
2. 2.Property/Parameter Descriptions
3. 3.Configuration Examples
4. 4.Troubleshooting/Notes

Training Considerations

• —Clean, consistently formatted technical content
• —Hierarchical structure preserved through level indicators
• —Complete context within each document
• —Code blocks and command examples ready for model training
• —Technical accuracy maintained in parameter descriptions

Usage Example

import pandas as pd

# Load the dataset
docs = pd.read_parquet('documentation.parquet')

# Access a document with all its sections
doc = docs.iloc[0]
print(f"Title: {doc['doc_title']}")

# Access specific sections
for i in range(int(doc['section_count'])):
    print(f"\nSection {i+1}: {doc[f'section_{i+1}_title']}")
    print(f"Content: {doc[f'section_{i+1}_content'][:200]}...")

Data Quality

• —Full preservation of technical content
• —Consistent section numbering
• —Maintained code formatting
• —Preserved command syntax
• —Complete parameter descriptions
• —Original document hierarchy

Source

Derived from official MikroTik documentation. Original formatting and technical accuracy preserved for training purposes.

License

This dataset is derived from MikroTik's official documentation. Use should comply with MikroTik's terms of service and documentation licensing terms.