k2bd / awstin

High-level utilities for building and testing AWS Serverless Applications in Python

My Stack

I primarily use Python for backend, scientific, and general-purpose work, and use React/Typescript for building web apps.

For deployment, I am happy with both GCP and AWS. For hosting, I use GitHub Pages for open source projects.

I'm Currently Working On

DBMS Development

I'm currently going through the CMU Intro to Database Systems course, learning more about database systems and following the coursework in Rust. Following this I will be going through the advanced course as well as starting fresh on the long path of building out a system designed for production use.

You can follow my work on the coursework here:

Libraries I Maintain

Templates I Maintain

awstin's Issues

Project migration to github flow & poetry, streamline release

Acceptance Criteria:

• Project uses github flow, master & develop branches deleted
• CI migrated to github flow
• Project migrated to Poetry
• Release process streamlined (poe command to push a new tag from remote main, release action triggered on new tag)

Fix up standard logging of event and context

The lambda_handler CM's logging of the event and context doesn't seem to actually log on AWS. The logger just needs to be configured properly.

At the same time, the logger setup could be made available to all of awstin or even to users of awstin as a quick way to set up logging.

Updates on nested properties don't work

Maps and lists and stuff

Move to poetry

Batch writer

Better update item interface

More methods can be added to BaseAttribute to create functionality for update_item

General DynamoDB code tidy

DynamoDB service that maps tables to python data models

Basic table definitions parsed from serverless.yml

Enable static type checking of model attributes through type hinting

e.g. if I have

class MyModel(DynamoModel):
    my_key: str = Key()

    an_attr: Optional[int] = Attr()

mypy should complain if I try to do item = MyModel(my_key=123)

Ability to unit test the wrapped Lambda function easily

If I have:

@lambda_handler(my_parser)
def handler(a: int, b: str):
    ...

I should be able to do:

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

This way I can unit test my parser and my business logic separately.

Context manager for catching SNS topic publications

An integration test context manager to catch all messages published to an SNS topic and catch them in a list.

DynamoDB stream integration testing tools

In addition to an interface to DynamoDB streams in integration testing, it'd be cool to get a context manager that just captures and returns stream entries that occur inside the context (exclude those that were in the streams beforehand)

Support for if_not_exists and list_append in UpdateItem SET

Support for size in dynamodb condition expressions

Add doc building test to CI

Should fail and prevent merging if docs fail to build

Nested queries on map and list attributes

e.g. MyTable.a_list[3] == 5 or MyTable.a_map.prop.exists() and stuff

PynamoDB support

Support PynamoDB models in the awstin way of doing things, so PynamoDB projects can make use of the integration testing tools etc.

CI tests that examples pass

Structured dynamodb user docs

Better exceptions

Inbuilt exceptions for things like missing items on get

Support for list_append

Support for DynamoDB indexes

Allow nested map attributes

(via Attr or DynamoModel subclass?)

DynamoDB Streams

Replace set_env with mock.dict

DynamoDB table and stream integration testing tools from serverless.yml or cloudformation template

Create a chatroom example

The current chat example in a separate repo is out-of-date and doesn't work on the latest version of awstin

Bonus: Make it a room-based chatroom to show of ddb indexes
(main table's pkey is user ID or callback, global secondary index has composite key (room code, user ID))

Bonus bonus: Write a small frontend app that gets served via s3

Type annotations across DynamoDB

Interpret custom, env, and opt from serverless.yml

Extend create_serverless_tables to parse {env:...} from the environment, ${self:custom....} from the sls file, and ${opt:} from an argument (or **kwargs) to create_serverless_tables

Data model attribute default values