Tunaflow

Tunaflow is a backend-as-config tool. It can generate a data transformation micro-service ready to deploy, all with YAML. With power of TunaSchool(WIP), easyjson and gin, your data flow can scale infinity and light-weight.

Overview

How many hours do you spend on simple data transformation tasks/data mapping services? Say hi to Tunaflow.

With Tunaflow you can write a yaml and generate a deploy-ready package instantly.

All you need to do is define schema, and use flow to link everything together.

With TunaSchool you can balance flows across different packages.

Installation

go get -u github.com/borgmon/tunaflow

Usage

After create a new empty folder, run

tunaflow init {package path}

This will generate an example config file for you to start with.

After writing your YAML config, run

tunaflow apply

to generate your project.

YAML Config Example

version: 1 # YAML Config version
name: example # name of the project
app-version: v1.0 # version of the project
package-path: example # package path of the project, this is for go.mod

schemas:
  - name: input # name of the schema
    payload: # json schema with [key](data type)
      stringField: string
      boolField: boolean
      nested:
        value: int

  - name: output
    payload:
      name: string
      isFish: boolean
      value: int

flows:
  - name: myflow # name of the schema
    upstream: input # schema name for the incoming data
    downstream: output # schema name for the outgoing data
    mapping: # actual transformation. [downstream field name](upstream field name aka where to get it)
      name: stringField
      isFish: boolField
      value: nested.value

Road Map

• json example file as input
• subflow to handle one to many mappings
• code injection in YAML to handle complex mappings (time format etc)
• openapi doc generation
• finish up go doc
• validation for config
• more CLI options(clean, build)
• exclusive files
• tests
• refactor using interfaces to allow extension
• pass file path in yaml
• TunaSchool
• allow custom middleware
• async handler option
• fuzzy match or use ML