Sentiment Classifier Deployment with FastAPI and Docker

0. Clone this repository

And change the directory:

git clone https://github.com/Extremesarova/sentiment_classification_deployment.git
cd sentiment_classification_deployment

1. Download the model

Using this link (two files: vectorizer and classifier) and place it inside /app/model directory.

2. Create Docker container and run it

docker build -t sentiment_classification-app:0.4.0 .
docker run -p 80:80 sentiment_classification-app:0.4.0

3. Open browser

Using this link to check that everything is okay:

{
    "health_check":"OK", 
    "model_version":"0.4.0"
}

To get the prediction:

1. Open webpage using this link.

2. Press "Try it out".

3. Replace string inside the form with your text containing sentiment in Russian:

   {
       "text": "string"
   }

4. Press execute.

5. Look at the result below:

   {
       "sentiment": {
           "negative": 0.6240637133186361,
           "positive": 0.3759362866813639
       }
   }

Or just send a POST request using CURL like the one below in your terminal, replacing string with your text containing sentiment in Russian:

curl -X 'POST' \
  'http://0.0.0.0/predict' \
  -H 'accept: application/json' \
  -H 'Content-Type: application/json' \
  -d '{
  "text": "string"
}'