Schema Extractor

A python library for extracting and analyzing schema from knowledge graphs represented in turtle representation using multi-autonomous agents.

Features

• Extract schema components from TTL files
• Provide technical and semantic analysis
• Merge insights using different strategies
• Multiple output formats (text/JSON)
• Command-line interface
• Example script for quick start

Installation

Install this package via :

pip install SchemaExtractor

Or get the newest development version via:

pip install git+https://github.com/sensein/SchemaExtractor.git

Important

Copy the default configuration file and update the OPEN ROUTER API key. The rest can be used default.

# Copy and customize config
cp src/configuration/config.yaml my_config.yaml 

Configuration

The library uses a YAML configuration file to manage settings for all agents. You have two options:

Use Configuration:

• Create your own YAML file with the same structure
• Pass it using the --config option
• Allows customization of all agent settings

Example Configuration Structure:

openrouter:
  base_url: "https://openrouter.ai/api/v1"
  api_key: "your-api-key"
  headers:
    HTTP-Referer: "https://brainkb.org"
    X-Title: "Schema Extractor MultiAutonomous Agents"

base_config:
  temperature: 0.7
  top_p: 0.9
  presence_penalty: 0.0
  frequency_penalty: 0.0

agent_configs:
  executor:
    name: "SPARQLExecutorAgent"
    system_message: "..."
    model: "openai/gpt-4o-2024-11-20"
    temperature: 0.5
  prompt:
    name: "PromptAgent"
    system_message: "..."
    model: "openai/gpt-4o-2024-11-20"
    temperature: 0.7
  merger:
    name: "MergerAgent"
    system_message: "..."
    model: "openai/gpt-4o-2024-11-20"
    temperature: 0.6

Usage

1. Example Script (Quick Start)

python experiment/example.py --config my_config.yaml


python experiment/example.py \
    --ttl-file experiment/schema.ttl \
    --config my_config.yaml \
    --strategy comprehensive \
    --format json \
    --output analysis.json

script Options:

• --ttl-file: Path to TTL file (default: tests/sample.ttl)
• --config: Path to config YAML file (optional)
• --strategy: Analysis strategy (comprehensive, selective, conflict_resolution)
• --format: Output format (text, json)
• --output: Output file path (optional)

2. Command Line Interface

# Basic usage with default config
schema-extract extract data/schema.ttl

# With custom config
schema-extract extract data/schema.ttl --config my_config.yaml

# Full example with all options
schema-extract extract data/schema.ttl \
    --config my_config.yaml \
    --strategy comprehensive \
    --format json \
    --output results.json

# Show version
schema-extract version

# Show help
schema-extract --help
schema-extract extract --help

3. Python Library

from SchemaExtractor.app import SchemaExtractorApp

app = SchemaExtractorApp()

app = SchemaExtractorApp(config_file="my_config.yaml")

# Process  
result = app.extract_schema(
    ttl_file_path="data/schema.ttl",
    merge_strategy="comprehensive",
    output_format="text"
)


if result["status"] == "success":
    if result.get("output"):  # text format
        print(result["output"])
    else:  # json format
        print(f"Technical Analysis: {result['technical_analysis']}")
        print(f"Semantic Analysis: {result['semantic_analysis']}")
        print(f"Merged Analysis: {result['merged_analysis']}")
        print(f"Insights: {result['insights']}")

Output Sample for experiment/sample.ttl data:

Contributing

Please read CONTRIBUTING.md for details on our code of conduct and the process for submitting pull requests.

License

This project is licensed under the Apache License 2.0 - see the LICENSE.txt file for details.

Authors

• Tek Raj Chhetri tekraj@mit.edu

Acknowledgments

• This work is part of the BrainKB project.