Datasets:

BassemE
/

mulesoft-documentation-embeddings

license: mit
task_categories:
  - question-answering
  - text-retrieval
language:
  - en
tags:
  - mulesoft
  - documentation
  - embeddings
  - rag
  - api-integration
size_categories: 10K<n<100K
configs:
  - config_name: default
    data_files: dataset.parquet
    default: true

mulesoft-documentation-embeddings

MuleSoft Documentation Embeddings for RAG Applications

Dataset Information

Version: 1.0.0
Created: 2025-09-16T02:41:16.352809
Source: Vector Database
License: MIT
Language: en

Task Categories

question-answering, retrieval, knowledge-base

Dataset Statistics

SkillPilotDataSet_v11

Total Objects: 6430
Unique Properties: 13
Knowledge Sources: mulesoft, user_defined_docs
Average Content Length: 5079 characters

Dataset Schema

The dataset contains the original document properties with the following fields:

author: Document author
chunk_id: Unique chunk identifier
chunk_index: Position of chunk in document
document_id: Source document identifier
entities: Extracted entities
keywords: Document keywords
knowledge_source: Source of knowledge (mulesoft, user_defined_docs)
page_content: Main text content
relationships: Entity relationships
source_url: Original document URL
tags: Document tags
title: Document title
total_chunks: Total number of chunks in document
vector: Embedding vector (3072 dimensions from OpenAI text-embedding-3-large)

RAG Configuration

This dataset was created using the following RAG (Retrieval-Augmented Generation) configuration:

Embedding Model

Model Type: OpenAI
Model ID: text-embedding-3-large
Embedding Dimensions: 3072
Provider: OpenAI

Chunking Configuration

Chunk Size: 2048 characters
Chunk Overlap: 256 characters

Technical Details

Vector Database: Vector Database
Collection: SkillPilotDataSet_v11
Total Vectors: 6430
Vector Distance: Cosine similarity

Usage

This dataset can be used for:

Question answering systems
Retrieval-augmented generation (RAG)
Knowledge base construction
Technical interview preparation
AI assistant training

Files

dataset.parquet: Dataset in Parquet format (recommended for large datasets)
README.md: This documentation file

Citation

If you use this dataset, please cite:

@dataset{{mulesoft_documentation_embeddings,
  title={{MuleSoft Documentation Embeddings}},
  author={{Bassem Elsodany}},
  year={{2025}},
  url={{https://huggingface.co/datasets/BassemE/mulesoft-documentation-embeddings}}
}}