How to create a service from a function?

Question

I want to create a service using Django Rest API. I have a function. The result of this function should return 2 values and I should return these values in JSON API format. The function will work like this. I will receive the features_list as a parameter and I will use it to create a result and display it as a service in json format in def prediction function.

I created a sample API (I guess) it is class PredictionSet in my views but I actually want to make service the def prediction function in my views.

I cannot understand how to apply it. I am so confused. Any help would be appreciated.

models.py

class Liquidity(models.Model):
    pred_y = models.CharField(max_length=600)
    score = models.FloatField()

views.py

class PredictionSet(viewsets.ModelViewSet):
    queryset = Liquidity.objects.all()
    serializer_class = LiquiditySerializer


def prediction(request, features_list):
    filename = config.FINAL_MODEL_PATH
    classifier = pickle.load(open(filename, 'rb'))
    scale_file = config.SCALER_PATH
    scaler = pickle.load(open(scale_file, 'rb'))

    sample = np.array(features_list).reshape(1, -1)
    sample_scaled = scaler.transform(sample)

    pred_y = classifier.predict(sample_scaled)
    prob_y = classifier.predict_proba(sample_scaled)
    if prob_y[0][1] < 0.5:
        score = 0
    elif prob_y[0][1] <= 0.69:
        score = 1
    else:
        score = 2
    pred_y = pred_y[0]

    prediction_obj = Liquidity.objects.get_or_create(pred_y=pred_y, score=score)
    prediction_result = prediction_obj.pred_y
    prediction_score = prediction_obj.score

    context = {
        'prediction_result ': prediction_result,
        'prediction_score ': prediction_score,
    }

    return context

serializer.py

class LiquiditySerializer(serializers.HyperlinkedModelSerializer):
    class Meta:
        model = Liquidity
        fields = '__all__'

Do you want to return the result of the function `prediction` by the viewset ? — Rvector, Oct 19 '21 at 12:22

score 1 · Accepted Answer · answered Oct 19 '21 at 12:42

If you want to return custom JSON from a ModelViewset in DRF, you can override .list() and/or .retrieve() like this :

from rest_framework import status
from rest_framework.response import Response

class PredictionSet(viewsets.ModelViewSet):
    queryset = Liquidity.objects.all()
    serializer_class = LiquiditySerializer
    
    # Your custom function definition
    def prediction(self, request, features_list):
        # The content
    
    def retrieve(self, request, *args, **kwargs):
        result = prediction(...)  # Call your custom service and got result
        # Return the result as JSON (url = /api/v1/predictions/1) an object
        return Response({'data': result}, status=status.HTTP_200_OK)
    
    def list(self, request, *args, **kwargs):
        result = prediction(...)  # Call your custom service and got result
        # Return the result as JSON (url = /api/v1/predictions) a list of objects
        return Response({'data': result}, status=status.HTTP_200_OK)

For more details follow this link

````def retrieve(self, request, *args, **kwargs): result = self.prediction(features_list) ```` Should I take parameter like this? — edche, Oct 19 '21 at 12:54
If `prediction is a static method` then yes you can remove the `self` and `request` parameter. — Rvector, Oct 19 '21 at 13:23

How to create a service from a function?

1 Answers1