#WEBEX TEAMS BOTS HOW TO#
#WEBEX TEAMS BOTS CODE#

On the left hand side select API Gateway as the trigger. Also create a new role with basic permissions for the execution role.

Give the function a name and select Python 2.7 as the runtime. Within the lambda section of the AWS console click on Create function and Author from scratch. Now its time to create the lambda function that will perform the computations for our bot.

Then Review Policy and give it a name such as mywebexfunctionpolicy.

Back at the Service prompt type CloudWatch Logs and select CreateLogGroup,Д¬reateLogStream and PutLogEvents as the Actions.

Under Actions select Read and DetectLabels and DetectModerationLabels.

Click on Create policy and then type Rekognition as the service.

Select Lambda as the service that will use the role and then Next: Permissions.

Within the IAM section of the AWS console, click on Roles and then Create a Role.

Make a note of the access token as you will need it later. The first step is to head over to the WebEx Developer site and the Create a New App page. The labels consist of an attribute and the confidence that it is correct in percentages

A number of a number of labels are returned from the function and posted to the WebEx Teams space.

API gateway triggers the invocation of the Rekognition Lambda function which makes use of the Rekognition API to perform object and scene detection on the image.

Webex Teams sends a webhook request to API gateway address associated with the Rekognition Lambda function.

A user sends a message to the Webex Teams Rekognition bot which includes an attached image.

The Lambda function is coded in Python.ДЄlthough this example uses the Rekognition Label Detection API to recognise object and scene labels, it could easily be modified to identify faces (including faces from custom collections), perform sentiment analysis or recognise celebrities.ДЄt a high level, the following steps occur:

#WEBEX TEAMS BOTS HOW TO#

This post will show you how to build a serverless scene and object detecting WebEx teams bot utilising AWS API Gateway, Lambda and Rekognition. There is no need to gather and prepare training data, choose a model, train the mode, evaluate it, tune the hyper parameters before getting any predictions.ДЄlthough if we did want to build our own customer machine learning model then AWS also offers that capability through the Sage Maker managed service or even direct access to various Frameworks and Insfrascture such as TensorFlow, Keras, Caffe, PyTorch and more running GPU or CPU instances. Rekognition makes high level APIs available which we can utilise to perform image analysis simply and easily.

#WEBEX TEAMS BOTS CODE#

You pay only for the compute time you consume and there is no charge when the code is not running.ДЄnd for the machine learning elements we have AWS Rekognition which is part of the Application Services section of the AWS ML Stack. Services such as AWS Lambda let you run the bot code without provisioning or managing servers. So I thought I would take this opportunity to combine that with my passion for serverless applications and machine learning to create my own Webex Teams Machine Learning bot.

There were a couple of examples I came across that made use of Lambda however they were not immediately function, had poor documentation or made use of outdated APIs (ciscosparkapi instead of webexteamssdk). I had a look to see what examples I could find online and whilst there are multiple examples of Python based bots, most of them assume the use of a server which I wanted to avoid. I have been wanting to create a Webex Teams bot for a while now.

Become familiar with CloudWatch whilst debugging.

Create a Lambda function for the bot - An API should be implicitly created for the Lambda function.Create IAM role for our Lambda function with permissions to Rekognition and Cloudwatch logs.Create a bot in Webex teams and get the access token.#AWS #Serverless #Webex Teams #Bot #Lambda #Rekognition #API Gateway WebEx teams bot scene and object detection

Webex teams bots