supabase-models 0.0.2

Generate type-safe Pydantic models that simplify working with supabase-py

supabase-models

Generate type-safe Pydantic models from your Supabase database schema ready to use with supabase-py.

Key features

• Schema introspection: Automatically extracts table structures, constraints, and relationships
• Type-safe models: Creates Pydantic models with type hints and validation
• JSON serialization: Generated models include dump() and load() methods for sending/receiving data via supabase-py
• Constraint validation: Translates database column constraints to Pydantic validators
• Customizable output: Uses a built-in template by default, or modify the Jinja2 template to match your needs

Prerequisites

This package is designed to work with:

• supabase-py - the recommended Supabase Python client
• PostgreSQL databases (including Supabase projects)

Installation

pip install supabase-models
# or with uv (recommended)
uv add supabase-models

Basic Usage

1. Generate Models

supabase-models --database-url postgresql://user:password@localhost:5432/database

Default behavior:

• Output file: models.py in current directory
• Schema: public
• Template: Built-in Jinja2 template

[!TIP] See CLI Reference below for all available options and configuration methods

2. Use with supabase-py

The generated models provide dump() and load() methods to simplify working with supabase-py:

• dump() - Use when sending data to Supabase
• load() - Use when loading received data from Supabase responses

from supabase import create_client, Client
from models import Product, ProductStatusEnum  # Noqa # Your generated models

# Initialize Supabase client
supabase_client: Client = ... # Noqa

# INSERT: Create and insert a new product
product = Product(name="Wireless Mouse", sku="WM-2024", price=29.99, status=ProductStatusEnum.DRAFT)
insert_response = supabase_client.table(Product.table_name).insert(product.dump()).execute()

# SELECT: Query and parse products back to typed models
select_response = supabase_client.table(Product.table_name).select("*").execute()
products: list[Product] = Product.load(select_response)

See the section below for details on how these models are generated.

Generated Output

Given this database schema:

-- Create enum types
CREATE TYPE product_status AS ENUM ('draft', 'active', 'archived');

-- Create tables
CREATE TABLE categories (
    id BIGSERIAL PRIMARY KEY,
    name VARCHAR(100) NOT NULL
);

CREATE TABLE products (
    id BIGSERIAL PRIMARY KEY,
    name VARCHAR(100) NOT NULL,
    sku VARCHAR(20) UNIQUE NOT NULL CHECK (sku ~ '^[A-Z]{2,3}-[0-9]{3,4}$'),
    price DECIMAL(10,2) CHECK (price > 1),
    category_id BIGINT REFERENCES categories(id),
    status product_status DEFAULT 'draft',
    created_at TIMESTAMPTZ DEFAULT NOW()
);

The tool generates the following models:

Note: Simplified example showing key capabilities. Some code sections abbreviated for clarity.

from datetime import datetime
from decimal import Decimal
from enum import Enum
from typing import Any, ClassVar
from pydantic import BaseModel, Field


class ProductStatusEnum(str, Enum):
    """Enum for product_status values."""
    DRAFT = "draft"
    ACTIVE = "active"
    ARCHIVED = "archived"


class SupabaseBaseModel(BaseModel):
    """Base model with supabase-py integration helpers."""
    
    ...
    # All helper logic is automatically generated
    # Main methods: load() for parsing responses, dump() for preparing data
    # Additional utilities for validation and type conversion are included


class Product(SupabaseBaseModel):
    """Model for 'products' table.

    Attributes:
        id (int | None): Primary key column; Auto-increment.
        name (str | None): Required column.
        sku (str | None): Required column; Unique.
        price (Decimal | float | None): Optional column.
        category_id (int | None): Optional column; Foreign key to 'categories'.
        status (ProductStatusEnum | None): Optional column; Default: 'draft'.
        created_at (datetime | None): Optional column; Default: now().
        categories (Category | None): Related table Category (requires categories(*) in query)
    """
    table_name: ClassVar[str] = "products"
    _required_columns: ClassVar[list[str]] = ["name", "sku"]

    # Primary key columns:
    id: int | None = Field(default=None, description="Auto-increment")

    # Required columns:
    name: str | None = Field(default=None, max_length=100)
    sku: str | None = Field(default=None, description="Unique", max_length=20, pattern=r"^[A-Z]{2,3}-[0-9]{3,4}$")

    # Optional columns:
    price: Decimal | float | None = Field(default=None, gt=1)
    category_id: int | None = Field(default=None, description="Foreign key to 'categories'")
    status: ProductStatusEnum | None = Field(default=None, description="Default: 'draft'")
    created_at: datetime | None = Field(default=None, description="Default: now()")

    # Relations:
    categories: "Category | None" = Field(default=None, description="Related table Category. Include categories(*) in query to populate.")


class Category(SupabaseBaseModel):
    """Model for 'categories' table.

    Attributes:
        id (int | None): Primary key column; Auto-increment; Default: nextval('categories_id_seq').
        name (str | None): Required column.
    """
    table_name: ClassVar[str] = "categories"
    _required_columns: ClassVar[list[str]] = ["name"]

    # Primary key columns:
    id: int | None = Field(default=None, description="Auto-increment; Default: nextval('categories_id_seq')")

    # Required columns:
    name: str | None = Field(default=None, max_length=100)

Supported Features

PostgreSQL Feature	Pydantic Output
Basic Features
PRIMARY KEY	Field descriptions
FOREIGN KEY	Relationship information
NOT NULL	Required field detection
DEFAULT value	Field descriptions
UNIQUE	Field descriptions
AUTOINCREMENT	Field descriptions
Data Types
VARCHAR(n), CHAR(n), TEXT	str with Field(max_length=n)
INTEGER, BIGINT, SMALLINT	int
DECIMAL(p,s), NUMERIC(p,s)	Decimal | float with precision bounds
REAL, DOUBLE PRECISION	float
BOOLEAN	bool
DATE, TIMESTAMP, TIMESTAMPTZ	datetime
TIME, TIMETZ	time | str (for timezone types)
JSON, JSONB	dict[str, Any]
UUID	UUID
BYTEA	bytes
SERIAL, BIGSERIAL	int with auto-increment
ENUM types	CustomEnum(str, Enum)
Constraints
VARCHAR(n)	Field(max_length=n)
CHECK (x > 0)	Field(gt=0)
CHECK (x >= 10)	Field(ge=10)
CHECK (x <= 1000)	Field(le=1000)
CHECK (x < 100)	Field(lt=100)
CHECK (x BETWEEN 0 AND 1000)	Field(ge=0, le=1000)
CHECK (char_length(x) >= 5)	Field(min_length=5)
CHECK (char_length(x) <= 50)	Field(max_length=50)
CHECK (x ~ '^.+@.+$')	Field(pattern=r"^.+@.+$")
CHECK (x ~* '^.+@.+$')	Field(pattern=r"^.+@.+$")

[!NOTE]
Constraint parsing is continuously improving. Please open an issue with examples of constraints you'd like to see supported!

CLI Reference

supabase-models [OPTIONS]

Options:
  --database-url TEXT    PostgreSQL connection string
  -o, --output TEXT      Output file (default: models.py)
  -s, --schema TEXT      Database schema (default: public)
  -t, --template TEXT    Custom Jinja2 template file
  -v, --verbose          Enable verbose logging
  --version              Show version
  --help                 Show help

CLI Examples

# Basic usage with environment variable
export DATABASE_URL="postgresql://user:pass@host:port/db"
supabase-models

# Direct database URL
supabase-models --database-url "postgresql://user:pass@host:port/db"

# Custom output and schema
supabase-models --output app/models.py --schema public --verbose

# Multiple schemas
supabase-models --schema auth --output auth_models.py
supabase-models --schema public --output public_models.py

Development

# Install development dependencies
uv sync

# Format and lint
uv run ruff format .
uv run ruff check .

# Run tests
uv run pytest

Contributing

For issues and contributions, visit the GitHub repository.