How to Use Django App Server to consume Split Impressions Webhook

This article shows step by step how to use Python Django App Server to consume Split Impressions Webhook and store each impression in a Postgresql database.

Environment used:

• Python 2.7.15
• requests 2.18.1
• Django 1.8.11
• psycopg2 2.8.4
• Postgresql 11.1

Here are the steps to setup the integration:

1. Login to your Postgresql database using psql command and create the impressions table.

psql [Your Database]

-# create table impressions (key varchar(500),  treatment varchar(1000), bucketingKey varchar(100), label varchar(500), mtime varchar(500), split varchar(500), environmentName varchar(500), environmentId varchar(500), splitVersionNumber  varchar(500)); 

2. Assuming you already have a site created within Django, add a new app called webhook_consumer, from within your site folder, run the following command.

python manage.py startapp webhook_consumer

3. Inside the webhook_consumer folder, open the views.py file and add the following imports and variable declarations.

import gzip
import StringIO
import Postgresql
import Constants

###############################################################
constants = Constants.Constants()
userName=constants.POSTGRESUSER
password=constants.POSTGRESPASS
serverName=constants.POSTGRESSERVER
dbSchema=constants.POSTGRESDB
###############################################################

4. The Post handler function is next, we will need to use gzip library to unzip the post body coming from Split.io, then will iterate through all the impressions JSON array, then will add each impression to the Postgresql database. Paste the code below in the file.

@csrf_exempt
@require_POST
def webhook(request):
    f = StringIO.StringIO(request.body)
    f.seek(0)
    gzf = gzip.GzipFile(fileobj=f)
    content = gzf.read()
    data = json.loads(content)
    mySQL = Postgresql.Postgresql(userName, password, serverName, dbSchema)

    for item in data:
        key=""
        split=""
        environmentId=""
        environmentName=""
        treatment=""
        mtime=""
        bucketingKey=""
        label=""
        splitVersionNumber=""
        for k in item:
            if k=="key":
                key=str(item[k])
            if k=="split":
                split=str(item[k])
            if k=="environmentId":
                environmentId=str(item[k])
            if k=="environmentName":
                environmentName=str(item[k])
            if k=="treatment":
                treatment=str(item[k])
            if k=="time":
                mtime=str(item[k])
            if k=="bucketingKey":
                bucketingKey=str(item[k])
            if k=="label":
                label=str(item[k])
            if k=="splitVersionNumber":
                splitVersionNumber=str(item[k])
        mySQL.addImpression(key, split, environmentId, environmentName, treatment, bucketingKey, mtime, label, splitVersionNumber)

    return HttpResponse(status=200)

5. Next add the following code in the same app folder with new file name Constants.py, make sure to update the connection information for Postgresql Database.

class Constants:
    POSTGRESUSER="[postgres user name]"
    POSTGRESPASS="[postgres user password]"
    POSTGRESSERVER="[postgres server name or ip]"
    POSTGRESDB="[database name]"

6. The Code use the following code as a database layer for Postgresql, add the code below under file name Postgresql.py under the app folder.

import psycopg2

class Postgresql:
    def __init__(self, userName, password, serverName, dbSchema):
        try:
            self.cnx = psycopg2.connect(host=serverName,database=dbSchema, user=userName, password=password)
        except Exception, e:
            print "Failed to connect to PostgreSQL, Error "+str(e.args)+", "+str(e)
    def addImpression(self, key, split, environmentId, environmentName, treatment, bucketingKey, mtime, label, splitVersionNumber):
        try:
            addData = ("INSERT INTO impressions (key, split, environmentId, environmentName, treatment, bucketingKey, mtime, label, splitVersionNumber) VALUES (%s, %s, %s, %s, %s, %s, %s, %s, %s)")
            params = [key, split, environmentId, environmentName, treatment, bucketingKey, mtime, label, splitVersionNumber]
            cursor = self.cnx.cursor()
            cursor.execute(addData, params)
            self.cnx.commit()
            cursor.close()
        except Exception, e:
            print "Failed to Run SQL, Error "+str(e.args)+", "+str(e)

7. The code is ready now to deploy, don't forget to enable the webhook_consumer URL by updating the urls.py file for your site.

from django.conf.urls import include, url
from django.contrib import admin

urlpatterns = [
#### Append the line below to the array #### 
               url(r'^webhook_consumer/', include('webhook_consumer.urls')),
               ]

8. Restart your Django server.

 python manage.py runserver 0.0.0.0:8000

9. Your webhook service is ready now, login to Split user interface, go to Admin Settings page, click on Integrations, and click Add button for Outgoing Webhook (Impressions). 

10. Select which environment's impressions to export, and paste the URL below, and click Save.

http://[Django Server:port]/webhook_consumer/