I want people to use existing Python libraries available now (multivac) to run a function remotely on the cloud. We already have some sample code. Process data remotely.
Presentation will be in a notebook. Then we can run the code from the notebook.
Go through cloudpipe setup and put together some examples. Put together an outline that makes sense to me from our docs. Find examples I can think of. Calculating pi and addition.
Real world use cases are a little different. You have a workload that you need to run. Like Travis, when you ask for work not when you make a commit.
How am I supposed to get this done?
Using tmpnb because Kyle has the certs for that. developer.rackspace.com jupyter.com Cloudpipe, domain is part of Rackspace
Background:
- Straight to hands on.
- What's Pickle? How does it work?
- Code snippet. API submit a job.
Demo it. Give high-level explanation of what is happening. Then get people doing it hands 5.
Objectives:
- People can consume a service to run jobs. Three layers: a discrete thing I want to run one function, a service exists for it;
Everyone should write a simple function that they run remotely.
Logistics To get started, they need a Rackspace account. Have them sign up for Dev+. Get list of people who signed up for workshop (from Ruben) and send them prequsites. Prereqs: Dev+ account; Python 2+ installed. 2.7 Install pip. Anaconda. Printouts with links and relevant documentation.
Pip install multyvac. The command you run to configure is three lines. Five lines of setup. The first two of which are Python packaging related (could be a little tricky for some workshop attendees) People know how to get started.
Custom rack.to URL Onboard customers. Talk to onboarding about making the sign-up process easy.