packyak 0.1.3

Infrastructure for AI applications and machine learning pipelines

PackYak

Packyak makes it easy to build Lakehouses, Data Pipelines and and AI applications on AWS.

Roadmap

• StreamlitSite - deploy a Streamlit application to ECS with VPC and Load Balancing
• Infer least privilege IAM Policies for Streamlit scripts (home.py, pages/*.py)
• @function - host an Lambda Function
• Infer least privilege IAM Policies for functions
• Bucket - work with files in S3, attach event handlers
• Queues - send messages to, attach event handlers
• Stream - send and consume records through AWS Kinesis
• Table - store structured data (Parquet, Orc, etc.) in a Glue Catalog. Model data using pydantic
• @asset - build data pipelines with dependency graphs
• @train - capture the inputs and outputs of a function for ML training and human feedback
• Generate audit reports for HIPAA and GDPR compliance policies

Installation

Pre-requisites

1. Docker (for bundling Python applications for the target runtime, e.g. in an Amazon Linux Lambda Function)
2. Python Poetry

curl -sSL https://install.python-poetry.org | python3 -

3. poetry-plugin-export - see https://python-poetry.org/docs/plugins/#using-plugins

poetry self add poetry-plugin-export

How To: Deploy Streamlit

Custom Domain

1. Create a Hosted Zone
2. Transfer the DNS nameservers from your DNS provider to the Hosted Zone
3. Create a Certificate

HTTPS

1. Create a Certificate via the AWS Console

Example

🔧 Note: Packyak is in active development. Not all features are implemented. Check back to see the following example grow.

Below is the most simple Packyak application: a Bucket with a Function that writes to it.

Your application's infrastructure is declared in code. The Packyak compiler analyzes it to auto-provision cloud resources (in this case AWS S3 Bucket and Lambda Function) with least privilege IAM Policy inference.

from packyak import Bucket, function

videos = Bucket("videos")

@function()
async def upload_video():
    await videos.put("key", "value")

@videos.on("create")
async def on_uploaded_video(event: Bucket.ObjectCreatedEvent):
  video = await videos.get(event.key)
  transcription

@asset()
async def transcribed_videos():
  ...

Nessie Setup

TODO: should be done as part of packyak init

pip install pynessie

mkdir -p ~/.config

cat <<EOF > ~/.config/nessie
auth:
    type: aws
    timeout: 10
endpoint: http://localhost:19120/api/v1
verify: yes
EOF