Cookbook / Examples

[rjmk]

User Stories

• As a potential user, I want to understand how to adapt existing workflows to nstack
• As a potential user, I want to understand how to create new nstack workflows
• As a potential user, I want an idea of when I should be reaching for nstack

Notes from Epic

Examples / 'Cookbook'

• Introduction: What can you build with NStack?
• Productionise a Python model as an HTTP endpoint
• Build a workflow from two Postgres databases

Other notes

When we productionise, what are the features we want to show off? Might be cool if we start with a workflow with a subtle bug, productionise and then catch the bug

• What are the minimal features we want to show how to integrate (e.g. calling to an external service, DB read, DB write, system dependencies ...)?

What is the minimum version of this and a sketch of an iteration plan?

[rjmk]

re: #1 (comment)

Yeah, maybe the MVP is the nstack-examples repo

[lanthias]

This will partly be solved by #17, which covers various scenarios, but it would be helpful to have some common examples too.

[lanthias]

Anyone got any cool ideas here? I was thinking things like:

Create a reusable HTTP source that accepts four Doubles

def resusableEndpoint = sources.http : (Double, Double, Double, Double) { http_path = "/endpoint" };

Create a reusable sink which writes to Postgres

def customerSink = sinks.postgres : (Double, Double, Double, Double) { postgres_host = "localhost", pg_database = "db", pg_table = "tbl" };

Connect sources and sinks together directly, without any functions in between

def workflow = reusableEndpoint | customerSink

Create a reusable, partially constructed Postgres source, which has connection parameters specified, but can be reused with multiple queries

def pgPartialSrc = sources.postgres : Text {
  pg_host = "localhost", 
  pg_port = "5432",
  pg_user = "user", 
  pg_password = "123456",
  pg_database = "db"
};

def workflow = pgPartialSrc { pg_query = "SELECT * FROM tbl;" } | myModule.Workflow | sinks.log : Text;

Create a reusable HTTP source which runs all queries through a function

def safeCustomerSource = sources.http : { customerId : Text } { http_path = "/customers" } |  Customers.EnsureCustomerExists;