sklearn_deployment.py

import joblib
import s3fs
import sklearn

@serve.deployment(route_prefix="/sentiment", name="sentiment-deployment")
class SentimentDeployment:
    def __init__(self):
        fs = s3fs.S3FileSystem(anon=True)
        with fs.open('ray-serve-blog/unigram_vectorizer.joblib', 'rb') as f:
            self.vectorizer = joblib.load(f)
        with fs.open('ray-serve-blog/unigram_tf_idf_transformer.joblib', 'rb') as f:
            self.preprocessor = joblib.load(f)
        with fs.open('ray-serve-blog/unigram_tf_idf_classifier.joblib', 'rb') as f:
            self.classifier = joblib.load(f)
            
    async def __call__(self, request):
        data = await request.body()
        vectorized = self.vectorizer.transform([str(data)])
        transformed = self.preprocessor.transform(vectorized)
        [result] = self.classifier.predict(transformed)
        if result == 1:
            return 'POSITIVE'
        else:
            return 'NEGATIVE'