DAGit

Currently being developed by Anubhav Jana along with Prof Puru IITB

This serverless FaaS platform supports individual function registrations, DAG registrations, Trigger registrations associated with DAGs/functons. This platform also supports various DAG primitives which is provided in this document for reference.

Guide: Register a Function

This section will guide you how to register a function. The following pre-requites are to be fulfilled before you register a function

• DockerFile - based on which the image will be build to run your function

• Python file - application logic to run the action/function (Here, in this example this is "test.py")

• requirements.txt - add all dependant pip packages in this file. In case you dont have any library dependancies,submit a blank requirements.txt

You must have the above 3 files before you register the function

Following is the sample code register_function.py to register a function. This will create a new function named "testaction" and register it onto our function store handled by us. The url endpoint is: /regster/function/function_name

register_function.py

import requests
import sys
import json


def server():
    
    url = "http://10.129.28.219:5001/register/function/testaction"
    files = [
    ('pythonfile', open(sys.argv[1],'rb')),
    ('dockerfile', open(sys.argv[2],'rb')),
    ('requirements.txt', open(sys.argv[3],'rb'))
   ]

   reply = requests.post(url = url,files = files,verify=False)
   print(reply.json())
   
def main():
    server()

if __name__=="__main__":
    main()

• Usage: python3 register_function.py test.py Dockerfile requirements.txt

Guide: Register a DAG

This section will guide you how to register a DAG. The following pre-requites are to be fulfilled before you register a DAG

• dag.json - a JSON specification file to define the DAG. Accepted DAG Format and a sample example is provided in this readme file itself.

Following is the sample code dag_register.py to register a DAG. This will register a new DAG onto our DAG store handled by us. The url endpoint is: /regster/dag

dag_register.py

import requests
import sys
import json

def server():
    url = "http://10.129.28.219:5001/register/dag"
    input_json_file = open(sys.argv[1])
    params = json.load(input_json_file)
    reply = requests.post(url = url,json = params,verify=False)
    print(reply.json())


def main():
    server()

if __name__=="__main__":
    main()

• Usage: python3 dag_register.py dag.json

Guide: Register a Trigger

This section will guide you how to register a trigger. The following pre-requites are to be fulfilled before you register a trigger

• trigger.json - a JSON specification file to define the trigger. Accepted DAG Format and a sample example is provided in this readme file itself.

Accepted Trigger Format

DAG specification includes both control dependancy as well as the control dependancy

Trigger Fields

• "trigger_name" : Name of the trigger. Type accepted is string

• "type": Type specifies whether the trigger is for function or dag. Accepted values are "dag" and "function"

• "trigger": Specifies the endpoint route

• "dags": If "type" field is specified as "dag","dags" will accept a list of dags to trigger (type = list). Else keep it as ""

• "functions": If "type" field is specified as "function","functions" will accept a list of functions to trigger (type = list). Else keep it as ""

Example format of trigger.json

{
    "trigger_name": "mydagtrigger",
    "type":"dag",
    "dags": ["odd-even-test","dummy-dag"],
    "functions":""
}

{
    "trigger_name": "myfunctiontrigger",
    "type":"function",
    "dags":"",
    "functions": ["odd-even-action"]
}

Following is the sample code trigger_register.py to register a trigger. This will register a new trigger onto our Trigger store handled by us. The url endpoint is: /regster/trigger

trigger_register.py

import requests
import sys
import json

def server():
    url = "http://10.129.28.219:5001/register/trigger/"
    input_json_file = open(sys.argv[1])
    params = json.load(input_json_file)
    reply = requests.post(url = url,json = params,verify=False)
    print(reply.json())
   

def main():
    server()

if __name__=="__main__":
    main()

• Usage: python3 trigger_register.py trigger.json

List of triggers

• http://10.129.28.219:5001/register/function/<function_name>

• http://10.129.28.219:5001/run/<trigger_name>

• http://10.129.28.219:5001/view/<dag_id>

• http://10.129.28.219:5001/view/activation/<function_activation_id>

• http://10.129.28.219:5001/view/dags

• http://10.129.28.219:5001/view/dag/<dag_name>

• http://10.129.28.219:5001/register/dag

• http://10.129.28.219:5001/view/functions

• http://10.129.28.219:5001/

Supported DAG Primitive

Accepted DAG Format

DAG specification includes both control dependancy as well as the control dependancy

DAG Fields

• "name" : Name of the DAG

• "node_id": Name of the function/action

• "node_id_label": Name you want to give to the node

• "primitive": Type of primitive the action supports - condition,parallel,serial(sequential)

• "condition": If primitive type is "condition", then you should provide the following fields "source", "operator" and "target", else you should leave it as ""

• "source": Specify any one of the response keys of the current node_id. For e.g. if one of the keys in response json is "result", and you want to provide a condition that if result=="even", then specify "source" as "result" and "target" as "even"

• "operator": Mathematical operations like "equals", "greater_than" , "less_than", "greater_than_equals", "less_than_equals" are accepted.

• "target": Specify the target value. It can accept both integer and string.

• "next": Specify the name of next node_id to be executed. If primitive = "parallel", "next" will take list of node_ids, else it will accept a single node_id in "" format. If this is the last node_id(ending node of the workflow), keep it as "".

• "branch_1": Specify node_id if primitive == condition else keep "". This is the target branch which will execute if condition is true

• "branch_2": Specify node_id if primitive == condition else keep "". This is the alternate branch which will execute if condition is false

• "arguments": Keep it blank for each node_id. It will get populated with json when the DAG is instantiated with the trigger

• "outputs_from": Specify the list of node_id/node_ids whose output current node_id needs to consume. This is for data dependancy.

{

"name":<string>

"dag":[
    {
        "node_id": "<string>",
        "properties":
        {
            "node_id_label": "<string>"
             
            "primitive":"<condition | parallel | serial>",
            "condition":
            {
                "source":"<key obtained from node)id result json>",
                "operator":"<equals || greater_than || less_than || ..>",
                "target":"<string|integer>"

            },
            "next": "<next node_id to be executed : if primitive=parallel, "next" will take list of node_ids, if primitive: serial then specify a single node_id >",
            "branch_1": "<node_id if type==condition else keep "">",
            "branch_2": "<node_id if type==condition else keep "">",
            "arguments": {} ---> Keep it blank for all action. It will get updated when the DAG is run
            "outputs_from": "<list>"

        }
    },
    {
       
        
    },

    .
    .
    .

    {

    }
    
]

}

Sample Example Usage

{

"name": "odd-even-test",

"dag": [

    {
        
        "node_id": "odd-even-action",
        "properties":
        {
            "label": "Odd Even Action",
            "primitive": "condition",
            "condition":
            {
                "source":"result",
                "operator":"equals",
                "target":"even"
            },

            "next": "",
            "branch_1": "even-print-action",
            "branch_2": "odd-print-action",
            "arguments": {},
            "outputs_from":[]

        }
    },
    {
        "node_id": "even-print-action",
        "properties":
        {

            "label": "Even Print Action",
            "primitive": "parallel",
            "condition": {},
            "next": ["increment-action","multiply-action"],
            "branch_1": "",
            "branch_2": "",
            "arguments":{},
            "outputs_from":["odd-even-action"]

        }
        
    },
    {
        "node_id": "increment-action",
        "properties":
        {

            "label": "INCREMENT ACTION",
            "primitive": "serial",
            "condition": {},
            "next": "",
            "branch_1": "",
            "branch_2": "",
            "arguments":{},
            "outputs_from":["even-print-action"]
            

        }
        
    },
    {
        "node_id": "multiply-action",
        "properties":
        {

            "label": "MULTIPLY ACTION",
            "primitive": "serial",
            "condition": {},
            "next": "",
            "branch_1": "",
            "branch_2": "",
            "arguments":{},
            "outputs_from":["even-print-action"]

        }
        
    },
    {
        "node_id": "odd-print-action",
        "properties":
        {
            "label": "Odd Print Action",
            "primitive": "serial",
            "condition":{},
            "next": "prime-check-action",
            "branch_1": "",
            "branch_2": "",
            "arguments":{},
            "outputs_from":["odd-even-action"]
        }
        
    },
    {
        "node_id": "prime-check-action",
        "properties":
        {
            "label": "Prime Check Action",
            "primitive": "serial",
            "condition":{},
            "next": "",
            "branch_1": "",
            "branch_2": "",
            "arguments":{},
            "outputs_from":["odd-print-action"]


        }
        
    }
    
]

}

Handle output from multiple actions

Suppose you want to merge outputs from two actions action_1 and action_2 in your action_3, then you must include the following lines in your action_3 to process incoming inputs from action_1 and action_2

. This is applicable for merging primitive as well as handling output from multiple actions.

• "key_action_1" refers to a key from action_1 response which you want to use in action_3
• "key_action_2" refers to a key from action_2 response which you want to use in action_3

params = json.loads(sys.argv[1])

op_1 = params["__ow_body"][0]["key_action_1"]

op_2 = params["__ow_body"][1]["key_action_2"]

Use these op_1 and op_2 to process