saantiaguilera/go-pipeline

Build, execute and represent pipelines (aka workflows / templates) in Go

Pipeline

Pipeline is a GPL3-licensed Go package for building, executing and representing pipelines (aka workflows / templates).

Getting started

• API documentation and examples are available via godoc.
• The examples directory contains more elaborate example applications.
• No specific mocks are needed for testing, every element is completely decoupled and atomic. You can create your own ones however you deem fit.

API stability

Pipeline follows semantic versioning and provides API stability via the gopkg.in service. You can import a version with a guaranteed stable API via http://gopkg.in/saantiaguilera/go-pipeline.v1

go-pipeline's Issues

How to handle errors?

Hi, nice library but I have a question, short of wrapping stages with lifecycle then kicking off another flow is there a cleaner way to handle error branches so they can also be included in the generated svg? As it stands my UML is incomplete because it does not show error paths. Using tags and context is awkward for error handling as well because it requires a lot of extra logic and wrapping of conditional groups at different steps to check for errors.

Playing with it

I have several questions.

1. recreate svg
   For what I understood, the program also makes the svg. But when I rename an example one, it doesn't recreate that. Did I misunderstood?

2. read form config (like toml)
   What if I could set up a pipeline via some config file. I would be able to build all the possibilities that I want, but have that config-file that will use only the parts needed.
   Let's say, I build a CRM but my customers have different wishes after a new user registers. I could then easily have them pick some predefined config (or even have them by themselves) That would mean that it's going to behave more like BPML (Perhaps even read that to drive the pipeline). I see a lot of possibilities here.