Skip to main content

DeepPavlov (Sentiment analysis)

In this article I'll show you how to setup sentiment analysis using https://deeppavlov.ai/ and my prepared docker image

Installing DeepPavlov​

git clone https://github.com/LiveHelperChat/sentiment && cd sentiment
docker-compose -f docker-compose.yml pull
wget https://livehelperchat.com/var/deep.tgz
tar zxfv deep.tgz
rm -f deep.tgz

Run one time

docker-compose -f docker-compose.yml up

Run as a service

docker-compose -f docker-compose.yml up -d

Testing

curl -X POST "http://localhost:5000/model" -H "accept: application/json" -H "Content-Type: application/json" -d "{\"x\":[\"all went no so good, but could have been better\"]}"

If you did everything right you should see output like this

[["neutral"]]

Configuring Live Helper Chat​

Our requirements are

  • We should send set sentiment on chat close event
  • Only for chats where operator participated
  • Set relevant chat subject

For that we will be using webhooks.

Configuring Rest API call​

In this case we are sending to bot only messages (operator and visitor) which happened after chat become pending. For that there is placeholder {{msg_all_since_transfer_content}}. You can also use other if you want.

In output combination we expect 200 header and very first element should contain our evaluation.

Configuring bot​

In the bot for simplicity we will have

  • chat.close This trigger will be the one who get's executed on chat close event.
  • Set sentiment This trigger will be executed if Rest API returns sentiment.
  • very_negative, negative, neutral, positive, very_positive just trigger which set's a subject.

chat.close trigger screenshot

Set sentiment trigger sceenshot

In this trigger at the moment only two evaluations are defined very_positive,neutral same way you can add the missing ones.

very_positive,neutral trigger screenshot

First text message is just for debug purposes. Second one response we just set a subject with ID 1.

Configuring webhook​

In webhook we define what event we want to listen. Also we want to listen to chats which had an operator.

That's it :)