Datasets:

Azzindani
/

ID_REG_DB_2511

language:
  - id
license: cc-by-4.0
task_categories:
  - text-retrieval
  - question-answering
tags:
  - legal
  - indonesian-law
  - knowledge-graph
  - sqlite
  - database
size_categories:
  - 100K<n<1M

Indonesian Legal Regulations - SQLite Database

Dataset Description

This is a SQLite database version of the Indonesian Legal Regulations Knowledge Graph dataset, optimized for easy querying and integration.

Database Statistics

Total Regulations: 748,558
Total Embeddings: 748,558
Total TF-IDF Vectors: 748,558
Database Size: 10498.50 MB

Database Schema

Main Tables

regulations - Core regulation data with KG features
embeddings - 1024D semantic embeddings (BLOB storage)
tfidf_vectors - 20000D TF-IDF vectors (BLOB storage)
kg_json_data - Knowledge Graph JSON data
regulations_fts - Full-Text Search virtual table

Features

Optimized Indexing: Fast queries on common fields
Full-Text Search: FTS5 for content search
Normalized Storage: Separate tables for vectors
BLOB Storage: Efficient storage for large vectors
Foreign Keys: Referential integrity

Usage Example

import sqlite3
import numpy as np
from huggingface_hub import hf_hub_download

# Download database
db_path = hf_hub_download(
    repo_id="Azzindani/ID_REG_DB_2510",
    filename="id_regulations.db",
    repo_type="dataset"
)

# Connect
conn = sqlite3.connect(db_path)
cursor = conn.cursor()

# Query regulations
cursor.execute("""
    SELECT global_id, content, kg_authority_score 
    FROM regulations 
    WHERE regulation_type = 'UU' 
    ORDER BY year DESC 
    LIMIT 10
""")

for row in cursor.fetchall():
    print(row)

# Get embedding
cursor.execute("""
    SELECT embedding, dimension 
    FROM embeddings 
    WHERE global_id = ?
""", ('some_id',))

row = cursor.fetchone()
if row:
    embedding_blob, dim = row
    embedding = np.frombuffer(embedding_blob, dtype=np.float32)
    print(f"Embedding shape: {embedding.shape}")

# Full-text search
cursor.execute("""
    SELECT global_id, content 
    FROM regulations_fts 
    WHERE regulations_fts MATCH 'pajak OR perpajakan'
    LIMIT 10
""")

# Close
conn.close()

Query Examples

Find regulations by domain

SELECT * FROM regulations 
WHERE kg_primary_domain = 'tax' 
ORDER BY kg_authority_score DESC;

Search content with FTS

SELECT * FROM regulations_fts 
WHERE regulations_fts MATCH 'undang-undang pajak';

Get regulation with embedding

SELECT r.*, e.embedding 
FROM regulations r
JOIN embeddings e ON r.global_id = e.global_id
WHERE r.regulation_type = 'UU' AND r.year >= 2020;

Find highly cited regulations

SELECT * FROM regulations 
WHERE kg_pagerank > 0.001 
ORDER BY kg_pagerank DESC 
LIMIT 20;

Source Dataset

This database was created from: Azzindani/ID_REG_KG_2511

License

CC-BY-4.0

Citation

@dataset{indonesian_legal_db_2024,
  author = {Azzindani},
  title = {Indonesian Legal Regulations - SQLite Database},
  year = {2024},
  publisher = {HuggingFace},
  url = {https://huggingface.co/datasets/Azzindani/ID_REG_DB_2511}
}