import os, grpc, logging
from sklearn.cluster import DBSCAN
from common.rpc_method_wrapper.Decorator import create_metrics, safe_and_metered_rpc_method
from dbscanserving.proto.dbscanserving_pb2 import DetectionRequest, DetectionResponse
from dbscanserving.proto.dbscanserving_pb2_grpc import DetectorServicer

LOGGER = logging.getLogger(__name__)

SERVICE_NAME = 'DbscanServing'
METHOD_NAMES = ['Detect']
METRICS = create_metrics(SERVICE_NAME, METHOD_NAMES)


class DbscanServiceServicerImpl(DetectorServicer):

    def __init__(self):
        LOGGER.debug('Creating Servicer...')
        LOGGER.debug('Servicer Created')

    @safe_and_metered_rpc_method(METRICS, LOGGER)
    def Detect(self, request : DetectionRequest, context : grpc.ServicerContext) -> DetectionResponse:
        if request.num_samples != len(request.samples):
            context.set_details("The sample dimension declared does not match with the number of samples received.")
            LOGGER.debug(f"The sample dimension declared does not match with the number of samples received. Declared: {request.num_samples} - Received: {len(request.samples)}")
            context.set_code(grpc.StatusCode.INVALID_ARGUMENT)
            return DetectionResponse()
        # TODO: implement the validation of the features dimension
        clusters = DBSCAN(eps=request.eps, min_samples=request.min_samples).fit_predict([[x for x in sample.features] for sample in request.samples])
        response = DetectionResponse()
        for cluster in clusters:
            response.cluster_indices.append(cluster)
        return response