Data Generation

Generate realistic test data using AI-powered synthetic data generation, ensuring your tests have the data they need.

info

Coming Soon This feature is currently in development.

Generation Methods

AI Synthetic Data

Let AI create contextually appropriate data:

Feature	Description
Context-Aware	Understands data relationships
Realistic	Mimics real-world patterns
Diverse	Varied data distribution
Consistent	Maintains data integrity

Faker Library

Built-in fake data generators:

Category	Examples
Person	Name, email, phone, address
Commerce	Product, price, company
Internet	URL, IP, username
Date	Past, future, range
Finance	Credit card, IBAN, currency
Location	Country, city, coordinates

Custom Rules

Define your own generation logic:

Rules:
  order_total:
    type: calculated
    formula: 'sum(items.price * items.quantity)'

  discount:
    type: conditional
    when: 'order_total > 100'
    then: 'order_total * 0.1'
    else: 0

  status:
    type: weighted
    values:
      completed: 70
      pending: 20
      cancelled: 10

Generation Profiles

Quick Generation

Simple data creation:

Profile: Basic User
Generate:
  - name: faker.name()
  - email: faker.email()
  - phone: faker.phone()

Advanced Generation

Complex data with relationships:

Profile: E-commerce Order
Generate:
  User:
    count: 100
    fields:
      name: faker.name()
      email: faker.email(unique=true)

  Product:
    count: 500
    fields:
      name: faker.commerce.productName()
      price: faker.price(10, 1000)

  Order:
    count: 1000
    fields:
      user_id: random(User.id)
      status: weighted(completed:70, pending:30)
      created_at: faker.date.past(30)

  OrderItem:
    count: 3000
    fields:
      order_id: sequential(Order.id, 1-5)
      product_id: random(Product.id)
      quantity: random(1, 10)

Data Patterns

Distribution Patterns

Pattern	Description	Use Case
Uniform	Equal probability	Random selection
Normal	Bell curve	Natural variation
Weighted	Custom probabilities	Status distribution
Sequential	Ordered values	IDs, dates

Realistic Patterns

# Age distribution matching demographics
age:
  type: normal
  mean: 35
  std: 15
  min: 18
  max: 80

# Purchase amount with realistic skew
amount:
  type: lognormal
  mean: 50
  std: 30

# Time-based patterns
created_at:
  type: time_series
  pattern: business_hours
  timezone: UTC

Constraints & Validation

Uniqueness

email:
  type: string
  generator: faker.email()
  unique: true
  retry: 10

Referential Integrity

order:
  user_id:
    reference: users.id
    on_missing: create # create, skip, error

Custom Validation

age:
  type: integer
  validate:
    - 'value >= 18'
    - 'value <= 120'

email:
  type: string
  validate:
    - regex: "^[a-z]+@[a-z]+\\.[a-z]+$"

Bulk Generation

Large Datasets

Generate millions of records efficiently:

Size	Strategy
< 10K	In-memory
10K - 1M	Batched
> 1M	Streaming

Performance Options

Generation:
  records: 1000000
  batch_size: 10000
  parallel: true
  workers: 4
  output: streaming

Progress Tracking

• Real-time progress
• Estimated completion
• Error reporting
• Pause/resume capability

Templates

Built-in Templates

Template	Description
User Account	Standard user profile
E-commerce Order	Order with items
Financial Transaction	Payment records
Healthcare Patient	Medical records
Inventory Item	Product inventory

Custom Templates

Create reusable generation templates:

Template: Banking Customer
Version: 1.0
Fields:
  account_number:
    type: string
    pattern: '[A-Z]{2}[0-9]{18}'

  balance:
    type: decimal
    min: 0
    max: 1000000

  account_type:
    type: enum
    values: [checking, savings, investment]

Export Options

Formats

Format	Best For
JSON	API testing
CSV	Database import
SQL	Direct database insert
Excel	Manual review
Parquet	Big data

Streaming Export

For large datasets:

• Direct database insert
• File streaming
• API endpoint delivery

Best Practices

Data Quality

• Validate generated data
• Check distributions
• Verify relationships

Performance

• Use appropriate batch sizes
• Enable parallel generation
• Stream large datasets

Reproducibility

• Use seed values
• Save generation profiles
• Document configurations