Skip to content

A pythonic interface to Amazon's DynamoDB for single-table design

License

Notifications You must be signed in to change notification settings

3mcloud/falcano

Repository files navigation

Falcano

A Pythonic interface for Amazon's DynamoDB that supports Python 3 and single-table design based on PynamoDB.

Installation

pip install falcano

Basic Usage

Basic usage is nearly identical to PynamoDB. Meta must inherit from Model.Meta and Type must be defined for every model.

Create a model that describes a model in your table.

from falcano.model import Model
from falcano.attributes import UnicodeAttribute

class User(Model):
    '''
    A DynamoDB User
    '''
    class Meta(Model.Meta):
        table_name = 'dynamodb-user'
        billing_mode = 'PAY_PER_REQUEST'
    email = UnicodeAttribute(null=True)
    first_name = UnicodeAttribute(range_key=True)
    last_name = UnicodeAttribute(hash_key=True)
    Type = UnicodeAttribute(default='user')

Create the table if needed:

User.create_table(billing_mode='PAY_PER_REQUEST')

Create a new user:

user = User('John', 'Denver')
user.email = 'djohn@company.org'
user.save()

Now, search your table for all users with a last name of 'Denver' and whose first name begins with 'J':

for user in User.query('Denver', User.first_name.startswith('J')):
    print(user.first_name)

Examples of ways to query your table with filter conditions:

for user in User.query('Denver', User.email.eq('djohn@company.org')):
    print(user.first_name)
for user in User.query('Denver', User.email=='djohn@company.org'):
    print(user.first_name)

Retrieve an existing user:

try:
    user = User.get('John', 'Denver')
    print(user)
except User.DoesNotExist:
    print('User does not exist')

Advanced Usage

Indexes? No problem:

from falcano.model import Model
from falcano.indexes import GlobalSecondaryIndex, AllProjection
from falcano.attributes import NumberAttribute, UnicodeAttribute

class ViewIndex(GlobalSecondaryIndex):
    class Meta:
        billing_mode = 'PAY_PER_REQUEST'
        projection = AllProjection()
    view = NumberAttribute(default=0, hash_key=True)

class TestModel(Model):
    class Meta(Model.Meta):
        table_name = 'TestModel'
    forum = UnicodeAttribute(hash_key=True)
    thread = UnicodeAttribute(range_key=True)
    view = NumberAttribute(default=0)
    Type = UnicodeAttribute(default='test')
    view_index = ViewIndex()

Now query the index for all items with 0 views:

for item in TestModel.view_index.query(0):
    print(f'Item queried from index: {item}')

It's simple!

Want to use DynamoDB local? Add a host name attribute and specify your local server.

from falcano.models import Model
from falcano.attributes import UnicodeAttribute

class User(Model):
    '''
    A DynamoDB User
    '''
    class Meta(Model.Meta):
        table_name = 'dynamodb-user'
        host = 'http://localhost:8000'
    email = UnicodeAttribute(null=True)
    first_name = UnicodeAttribute(range_key=True)
    last_name = UnicodeAttribute(hash_key=True)
    Type = UnicodeAttribute(default='user')

Single-Table Design Usage

Want to follow single-table design with an index and rename the Type attribute? No problem:

from falcano.model import Model
from falcano.indexes import GlobalSecondaryIndex, AllProjection
from falcano.attributes import NumberAttribute, UnicodeAttribute

class TypeIndex(GlobalSecondaryIndex):
    class Meta:
        index_name = 'Type'
        billing_mode = 'PAY_PER_REQUEST'
        projection = AllProjection()
    Kind = UnicodeAttribute(default=0, hash_key=True)

class BaseModel(Model):
    class Meta(Model.Meta):
        table_name = 'single_table'
        # use the Kind attribute in place of Type for deserialization
        model_type = 'Kind'
    PK = UnicodeAttribute(hash_key=True)
    SK = UnicodeAttribute(range_key=True)
    TypeIndex = TypeIndex()

class User(BaseModel):
    email = UnicodeAttribute(null=True)
    Kind = UnicodeAttribute(default='user')

class Team(BaseModel):
    owner = UnicodeAttribute(null=True)
    Kind = UnicodeAttribute(default='team')

Features

  • Python >= 3.8 support
  • Use of Table boto3 resource
    • DynamoDB API conditions objects
    • Auto-Typing
  • An ORM-like interface with query and scan filters
  • Compatible with DynamoDB Local
  • Support for Unicode, Binary, JSON, Number, Set, and UTC Datetime attributes
  • Support for Global and Local Secondary Indexes
  • Automatic pagination for bulk operations(?)
  • Complex queries
  • Multiple models per table and deserialization into objects

Features not yet implemented

  • Provides iterators for working with queries, scans, that are automatically - paginated
  • Iterators for working with Query and Scan operations
  • Supports the entire DynamoDB API
  • Full table backup/restore
  • Batch operations with automatic pagination