Skip to main content

Utilities for building and testing AWS applications in Python

Project description


PyPI Dev Status

CI Build codecov

High-level utilities for building and testing AWS applications in Python.




To use DynamoDB either the TEST_DYNAMODB_ENDPOINT (for integration testing) or AWS_REGION (for production) environment variable must be set.

DynamoDB is accessed through Python data models that users define to represent structured data in tables.

from awstin.dynamodb import Attr, DynamoModel, Key

class User(DynamoModel):
    # Name of the DynamoDB table (required!)
    _table_name_ = "Users"

    # Sort or hash keys are marked with Key
    user_id = Key()

    # Other attributes are marked with Attr
    favorite_color = Attr()

    # The names of attributes and keys can differ from the names on the data
    # model - the name of the attribute in DynamoDB should be passed to Attr
    account_age = Attr("ageDays")

Tables are tied to these data models. They'll be returned when items are retrieved from the table. Also, put_item takes instances of this data model class.

These data models also define projection expressions, so only those attributes are retrieved from get_item, query, and scan calls.

from awstin.dynamodb import DynamoDB

dynamodb = DynamoDB()

# List of available tables
tables = dynamodb.list_tables()

# Access a table by model
users_table = dynamodb[User]

# Put an item into the table
user = User(

# Tables that only have a partition key can be accessed directly by their
# partition key
item1 = users_table["user123"]

# Tables that have partition and sort keys can be accessed by a tuple
table2 = dynamodb[AnotherTableModel]
item2 = table2[("hashval", 123)]

# Full primary key access is also available
item3 = table2[{"hashkey_name": "hashval", "sortkey_name": 123}]

Query and scan filters can be built up using these data models as well. Results can be iterated without worrying about pagination. Table.scan and Table.query yield items, requesting another page of items lazily only when it's out of items in a page.

scan_filter = (
    (User.account_age > 30)
    & (User.favorite_color.in_(["Blue", "Green"]))

for user in users_table.scan(scan_filter):

Queries must be given a query expression and can optionally be given a filter expression. Query expressions must represent valid DynamoDB queries.

class Student(DynamoModel):
    _table_name_ = "Students"

    # Hash key
    name = Key()

    # Sort key
    year = Key()

    homeroom = Attr()

students_table = dynamodb[Student]

query_expression = ( == "John") & (Student.year >= 10)
filter_expression = Student.homeroom == "Smith"

results = students_table.query(

Indexes work identically, but must have a _index_name_ attribute on the data model. Indexes can be used for queries and scans.

class ByHomeroomIndex(DynamoModel):
    _table_name_ = "Students"
    _index_name_ = "ByHomeroom"

    # Hash key
    homeroom = Key()

    # Sort key
    name = Key()

    year = Attr()

homeroom_index = dynamodb[ByHomeroomIndex]

query_expression = (
    (ByHomeroomIndex.homeroom == "Doe")
    & ( > "B")
filter_expression = ByHomeroomIndex.year > 11

items = list(homeroom_index.query(query_expression, filter_expression))

Nested Values

Filters on nested attributes work as well:

scan_filter = (
    (MyModel.map_attr.key == "value")
    & (MyModel.list_attr[3] == 10)

results = my_table.scan(scan_filter)

Updating Items

A syntax is also available for updating items, with an optional condition expression:

update_expression = (
    MyModel.an_attr.set(5 - MyModel.another_attr)
    & MyModel.third_attr.add(100)
    & MyModel.another_attr.remove()
    & MyModel.set_attr.delete([2, 3])

condition_expression = MyModel.an_attr > 11

updated = my_table.update_item(

update_item returns None if the condition evaluates to False.

Float and Decimal

Floats should be used when working with DynamoDB through awstin. Conversions between float and Decimal is done internally.

Unset Values

Values in a data model class that are unset, either by user instantiation or by retrieval from DynamoDB, are given the value awstin.dynamodb.NOT_SET.


For integration testing, a context manager to create and then automatically tear-down a DynamoDB table is provided. The context manager waits for the table to be created/deleted before entering/exiting to avoid testing issues. Hashkey and sortkey info can be provided.

from awstin.dynamodb.testing import temporary_dynamodb_table

with temporary_dynamodb_table(User, "hashkey_name") as table:
    item = User(




Lambda handlers can be made more readable by separating event parsing from business logic. The lambda_handler decorator factory takes a parser for the triggering event and context, and returns individual values to be used in the wrapped function.

from awstin.awslambda import lambda_handler

def event_parser(event, context):
    request_id = event["requestContext"]["requestId"]
    memory_limit = context["memory_limit_in_mb"]
    return request_id, memory_limit

def handle_custom_event(request_id, memory_limit):

API Gateway

Authorization Lambdas


Authorizor lambda responses can be generated with helper functions provided by awstin.apigateway.auth. accept, reject, unauthorized, and invalid will produce properly formatted auth lambda responses.

from awstin.apigateway import auth

def auth_event_parser(event, _context):
    token = event["headers"]["AuthToken"]
    resource_arn = event["methodArn"]
    principal_id = event["requestContext"]["connectionId"]

    return token, resource_arn, principal_id

def token_auth(token, resource_arn, principal_id):
    if token == "good token":
        return auth.accept(principal_id, resource_arn)
    elif token == "bad token":
        return auth.reject(principal_id, resource_arn)
    elif token == "unauthorized token":
        return auth.unauthorized()
        return auth.invalid()


A function wrapped with lambda_handler is stored on the inner attribute of the returned function. That way, the business logic of the handler can be tested separately without having to build events.

def my_handler(a: int, b: str):

# ------

def test_parser():
    args = my_parser(test_event, test_context)
    assert ...

def test_handler():
    result = my_handler.inner(1, "abc")
    assert ...



Websocket pushes can be performed with a callback URL and message:

from awstin.apigateway.websocket import Websocket

Websocket("endpoint_url", "dev").send("callback_url", "message")




SNS topics can be retrieved by name and published to with the message directly. This requires either the TEST_SNS_ENDPOINT (for integration testing) or AWS_REGION (for production) environment variable to be set.

from awstin.sns import SNSTopic

topic = SNSTopic("topic-name")
message_id = topic.publish("a message")

Message attributes can be set from the kwargs of the publish:

    "another message",
    attrib_a="a string",
    attrib_c=["a", "b", False, None],
    attrib_d=b"bytes value",

Project details

Download files

Download the file for your platform. If you're not sure which to choose, learn more about installing packages.

Files for awstin, version 0.0.15
Filename, size File type Python version Upload date Hashes
Filename, size awstin-0.0.15-py3-none-any.whl (29.8 kB) File type Wheel Python version py3 Upload date Hashes View
Filename, size awstin-0.0.15.tar.gz (26.1 kB) File type Source Python version None Upload date Hashes View

Supported by

AWS AWS Cloud computing Datadog Datadog Monitoring DigiCert DigiCert EV certificate Facebook / Instagram Facebook / Instagram PSF Sponsor Fastly Fastly CDN Google Google Object Storage and Download Analytics Microsoft Microsoft PSF Sponsor Pingdom Pingdom Monitoring Salesforce Salesforce PSF Sponsor Sentry Sentry Error logging StatusPage StatusPage Status page