Documentation

flowine.js is an open source engine for flow based programming. It is written in JavaScript.

Installation

npm i flowine

Usage

import flowine from "flowine"

(This is a ES module. You can also use it with Node.js, but have to add the following line to your package.json: "type": "module")

Introduction

Available Methods:

flowine.init() // Initializes flowine 
flowine.terminate() // Terminates flowine
flowine.status() // Get the status of flowine
flowine.createCanvas(canvasName) // Create a blank new canvas
flowine.deleteCanvas() // Delete the canvas
flowine.createNode(nodeType) // Create a new node
flowine.setPortData(id, data) // Set the port data of a node
flowine.deleteNode(id) // Delete a node
flowine.createEdge(sourcePortId, targetPortId) // Create a edge
flowine.updateEdge(id, sourcePortId, targetPortId) // Update a edge
flowine.deleteEdge(id) // Delete a edge
flowine.saveCanvasAsJSON(fileName) // Save the canvas as a JSON file
flowine.openCanvasFromJSON(fileName) // Open a canvas from a JSON file
flowine.runNode(id) // Run (Execute) a node
flowine.solveGraph() // Solve (Execute) the whole graph

Execution of Code:

Every node has a actionFunction. This can be JavaScript- or Python-Code. At the moment both can only be executed in a Node.js serverside environment. However in the future it will be possible toexecute JavaScript Code on the client-side.

Example of a JavaScript actionFunction:

"actionFunction": {
  "language": "JavaScript",
  "code": "(function uppercaseNode() { opdArray_[0] = ipdArray_[0].toUpperCase(); return opdArray_; })();"
}

Example of a Python actionFunction:

⚠️ Running Pythoncode is still in very early development. Use it with caution.

"UPPERCASE_PYTHON": {
  "nodeType": "PRIMITIVE",
  "caption": "UppercasePython",
  "inputPorts": [
    {
      "type": "TEXT",
      "visible": true
    }
  ],
  "outputPorts": [
    {
      "type": "TEXT",
      "visible": true
    }
  ],
  "actionFunction": {
    "language": "Python",
    "code": "import sys\nimport json\nimport os\nstdinArray_ = sys.stdin.readline().split(\",\")\nipdArray_ = []\nfor e in stdinArray_:\n\tipdArray_.append(e.strip().strip(\"[\").strip(\"]]\").strip(\"\\\\\"\"))\nopdArray_ = []\nopdArray_.append(ipdArray[0].upper())\nprint(json.dumps(opdArray_))"
  }
}

The structure of a Canvas

This is an empty Canvas with a Name, a creation Date and no Nodes and no Edges:

{
  "id": 1,
  "header": {
    "canvasName": "MyCanvas",
    "creationDate": "YYYY-MM-DD"
  },
  "body": {
    "nodes": [],
    "edges": []
  }
}

The structure of a Node

This Node reads TEXT from the Input-Port and transforms it to uppercase-TEXT that will be send to the Output-Port:

"UPPERCASE": {  
  "nodeType": "PRIMITIVE",  
  "caption": "Uppercase",  
  "inputPorts": [  
    {  
      "type": "TEXT",  
      "visible": true  
    }  
  ],  
  "outputPorts": [  
    {  
      "type": "TEXT",  
      "visible": true  
    }  
  ],  
  "actionFunction": {  
    "language": "JavaScript",  
    "code": "(function uppercaseNode() { opdArray_[0] = ipdArray_[0].toUpperCase(); return opdArray_; })();"  
  }
}

The Nodes are stored in the file nodelibrary.js. You can add your own custom nodes to this file.

Development

The core library consists of just 3 files:

• flowine.js This is the main library width all classes and functions
• config.js This is the configuration file that is used by flowine.js
• nodelibrary.js This is the standard node-library that is loaded by default

Run the following commands from the root folder and the use the devfolder to develop:

docker build . -t dev_env -f dev_env/Dockerfile
docker run -it --rm --name dev_env -p 127.0.0.1:3000:3000/tcp -p 127.0.0.1:8000:8000/tcp -v ${PWD}:/dev_env dev_env
cd /dev_env/src && npm i
cd /dev_env/dev_env && npm i && npm run dev

Example /example:

docker build . -t example -f example/Dockerfile
docker run -it --rm --name example -p 127.0.0.1:8000:8000/tcp -v ${PWD}:/example example
cd /example/src && npm i
cd /example/example && npm i && npm run dev

Support or Contact

flowine is maintained by Sebastian Kulle (github.com/sbstnkll)