Fashion

aka, python-openfaas

Work in progress!

Fashion is a library for using OpenFaaS functions in a Pythonic way. Fashion functions run inside their own Docker containers in your cluster, but act like local functions in your application's code! Cool!

For instance, let's imagine that we've installed the figlet function from the OpenFaaS function store. To call it from our application, we can simply:

from fashion import figlet
hi = figlet("Hello, world!")
#  _   _      _ _                             _     _ _
# | | | | ___| | | ___    __      _____  _ __| | __| | |
# | |_| |/ _ \ | |/ _ \   \ \ /\ / / _ \| '__| |/ _` | |
# |  _  |  __/ | | (_) |   \ V  V / (_) | |  | | (_| |_|
# |_| |_|\___|_|_|\___( )   \_/\_/ \___/|_|  |_|\__,_(_)
#                    |/

Fashion doesn't need any prior knowledge of the OpenFaas function, you can simply call it as if it were a local function!

Table of Contents

• Installation
• Usage
  • Ordinary Usage
  • Async
    • Callbacks
    • Instance Type Limiting / Automatic Cost Optimization
  • Other OpenFaaS utilities
  • Settings
• Related Projects
• TODO
• License

Installation

(Coming soon!)

pip install fashion

Usage

Ordinary Usage

Fashion is very easy to use, just import your OpenFaaS function as if it were native!

For instance:

from fashion import leftpad
leftpad("Hello!") # "      Hello!"

Alternately, you can use Fashion's trigger function directly, like so:

import fashion
fashion.trigger("leftpad", "Hello!") #"      Hello!"

Functions can be chained together without friction:

from fashion import leftpad, figlet
figlet(leftpad("Hello!"))
#  _                _   _      _ _       _ _
# ( )              | | | | ___| | | ___ | ( )
# |/               | |_| |/ _ \ | |/ _ \| |/
#                  |  _  |  __/ | | (_) |_|
#                  |_| |_|\___|_|_|\___/(_)

Async

(Still working on the interface for this, hang on a bit!) Similarly, you can use OpenFaaS functions asyncronously. To invoke an async function, import it with the prefix async_ like so:

from fashion import async_leftpad
async_leftpad("Hello!") # "6a3ae7fb-a8ee-4988-b7de-e3b81d1aed65"

or the old fashioned way:

import fashion
fashion.async_trigger("leftpad", "Hello!") # "6a3ae7fb-a8ee-4988-b7de-e3b81d1aed65"

Callbacks

Async functions can have their results sent to callback URLs and other OpenFaaS functions. For example, you can send to a pastebin like this:

from fashion import async_leftpad
async_leftpad("Hello", callback_url="https://postb.in/b/1269724059086-0568930923473")

or to another OpenFaaS function like this:

from fashion import async_leftpad
async_leftpad("Hello", callback_function="figlet")

Instance Type Limiting / Automatic Cost Optimization

An interesting use of this is limiting execution of functions to certain instance types. For instance, if this is included in an update_model function definition:

#update_model.yml
  [..]
   constraints:
     - "node.labels.instance_type == gpu"

Then you can write code which is automatically cost-optimized when it executes:

# Runs on expensive GPU instance
from fashion import update_model

# Runs on cheap CPU instance
from fashion import send_result_email

result = update_model(input)
send_email(result)
# Email sent!

Other OpenFaaS utilities

You can use Fashion to progromatically get information about all your OpenFaaS functions like so:

import fashion
fashion.system_functions()

Settings

You can configure your endpoints and other variables like so:

import fashion

fashion.GATEWAY_URL = "http://internal.network.location"
fashion.GATEWAY_PORT = 8081
fashion.TIMEOUT_SECONDS = 30

Related Projects

• Zappa - Python's Serverless Framework for AWS Lambda

TODO

• Tests
• Async
• Auth/HMAC settings
• Turn left__pad into left-pad invocation

License

(C) Rich Jones 2019, MIT License.