glassgen 0.1.4

A flexible synthetic data generation service

GlassGen

GlassGen is a flexible synthetic data generation service that can generate data based on user-defined schemas and send it to various destinations.

Features

• Generate synthetic data based on custom schemas
• Multiple output formats (CSV, Kafka, Webhook)
• Configurable generation rate
• Extensible sink architecture
• CLI and Python SDK interfaces

Installation

pip install glassgen

Local Development Installation

1. Clone the repository:

git clone https://github.com/glassflow/glassgen.git
cd glassgen

2. Create and activate a virtual environment:

python -m venv venv
source venv/bin/activate  # On Windows use: venv\Scripts\activate

3. Install the package in development mode:

pip install -e .

4. Install development dependencies:

pip install -r requirements-dev.txt

5. Run tests to verify installation:

pytest

Usage

Basic Usage

import glassgen
import json

# Load configuration from file
with open("config.json") as f:
    config = json.load(f)

# Start the generator
glassgen.generate(config=config)

Configuration File Format

{
    "schema": {
        "field1": "$generator_type",
        "field2": "$generator_type(param1, param2)"
    },
    "sink": {
        "type": "csv|kafka|webhook",
        "params": {
            // sink-specific parameters
        }
    },
    "generator": {
        "rps": 1000,  // records per second
        "num_records": 5000  // total number of records to generate
    }
}

Supported Sinks

CSV Sink

{
    "sink": {
        "type": "csv",
        "params": {
            "path": "output.csv"
        }
    }
}

WebHook Sink

{
    "sink": {
        "type": "webhook",
        "params": {
            "url": "https://your-webhook-url.com",
            "headers": {
                "Authorization": "Bearer your-token",
                "Custom-Header": "value"
            },
            "timeout": 30  // optional, defaults to 30 seconds
        }
    }
}

Kafka Sink

GlassGen supports multiple Kafka sink types:

1. Confluent Cloud

{
    "sink": {
        "type": "kafka.confluent",
        "params": {
            "bootstrap_servers": "your-confluent-bootstrap-server",
            "topic": "topic_name",
            "security_protocol": "SASL_SSL",
            "sasl_mechanism": "PLAIN",
            "sasl_plain_username": "your-api-key",
            "sasl_plain_password": "your-api-secret"
        }
    }
}

2. Aiven Kafka

{
    "sink": {
        "type": "kafka.aiven",
        "params": {
            "bootstrap_servers": "your-aiven-bootstrap-server",
            "topic": "topic_name",
            "security_protocol": "SASL_SSL",
            "sasl.mechanisms": "SCRAM-SHA-256",
            "ssl_cafile": "path/to/ca.pem"
        }
    }
}

Custom Sink

You can create your own sink by extending the BaseSink class:

from glassgen import generate
from glassgen.sinks import BaseSink
from typing import List

class PrintSink(BaseSink):
    def publish(self, data: str):
        print(data)
    
    def publish_bulk(self, data: List[str]):
        for d in data:
            self.publish(d)
    
    def close(self):
        pass

# Use your custom sink
config = {
    "schema": {
        "name": "$name",
        "email": "$email",
        "country": "$country",
        "id": "$uuid",        
    },    
    "generator": {
        "rps": 10,
        "num_records": 1000        
    }
}
generate(config, sink=PrintSink())

Supported Schema Generators

Basic Types

• $string: Random string
• $int: Random integer
• $intrange(min,max): Random integer within specified range (e.g., $intrange(1,100) for numbers between 1 and 100)
• $choice(value1,value2,...): Randomly picks one value from the provided list (e.g., $choice(red,blue,green) or $choice(1,2,3,4,5))
• $datetime: Current timestamp in ISO format (e.g., "2024-03-15T14:30:45.123456")
• $timestamp: Current Unix timestamp in seconds since epoch (e.g., 1710503445)
• $boolean: Random boolean value
• $uuid: Random UUID
• $uuid4: Random UUID4

Personal Information

• $name: Random full name
• $email: Random email address
• $company_email: Random company email
• $user_name: Random username
• $password: Random password
• $phone_number: Random phone number
• $ssn: Random Social Security Number

Location

• $country: Random country name
• $city: Random city name
• $address: Random street address
• $zipcode: Random zip code

Business

• $company: Random company name
• $job: Random job title
• $url: Random URL

Other

• $text: Random text paragraph
• $ipv4: Random IPv4 address
• $currency_name: Random currency name
• $color_name: Random color name

Pre Defined Schema

You can use of of the pre-defined schema:

import glassgen
from glassgen.schema.user_schema import UserSchema

config = {
    "sink": {
        "type": "csv",
        "params": {
            "path": "output.csv"
        }
    },
    "generator": {
        "rps": 50,
        "num_records": 100
    }
}
# use the pre-defined UserSchema
glassgen.generate(config=config, schema=UserSchema())

Example Configuration

{
    "schema": {
        "name": "$name",
        "email": "$email",
        "country": "$country",
        "id": "$uuid",
        "address": "$address",
        "phone": "$phone_number",
        "job": "$job",
        "company": "$company"
    },
    "sink": {
        "type": "webhook",
        "params": {
            "url": "https://api.example.com/webhook",
            "headers": {
                "Authorization": "Bearer your-token"
            }
        }
    },
    "generator": {
        "rps": 1500,
        "num_records": 5000,
        "event_options": {
            "duplication": {
                "enabled": true,
                "ratio": 0.1,
                "key_field": "email",
                "time_window": "1h"
            }
        }
    }
}

Event Options

Duplication

GlassGen supports controlled event duplication to simulate real-world scenarios where the same event might be processed multiple times.

"event_options": {
    "duplication": {
        "enabled": true,        // Enable/disable duplication
        "ratio": 0.1,          // Target ratio of duplicates (0.0 to 1.0)
        "key_field": "email",  // Field to use for duplicate detection
        "time_window": "1h"    // Time window for duplicate detection
    }
}

• enabled: Boolean to turn duplication on/off
• ratio: Decimal value (0.0 to 1.0) representing the percentage of events that should be duplicates
• key_field: Field name from the schema to use for identifying duplicates
• time_window: String representing the time window for duplicate detection (e.g., "1h" for 1 hour, "30m" for 30 minutes)

The duplication feature:

• Maintains the specified ratio across all generated events
• Only considers events within the configured time window for duplication
• Uses the specified key_field to identify potential duplicates
• Ensures memory efficiency by automatically cleaning up old events

Creating a New Release

To create a new release:

1. Make sure you have the release script installed:

pip install -e .

2. Run the release script with the new version:

./scripts/release.py release 0.1.1

This will:

• Update the version in pyproject.toml
• Create a git tag
• Push the changes
• Trigger the GitHub Actions workflow to:
  • Build the package
  • Publish to PyPI
  • Create a GitHub release

The version must follow semantic versioning (X.Y.Z format).