gRPC

Abstract

---

• An RPC framework that allows applications to communicate across networks as if they were calling local functions, hiding the complexity of network communication
• Built on top of. HTTP 2.0, using protobuf for efficient data transfer in binary format

Network protocol VS Framework

Network protocol is the low-level rules for how data moves across networks (like HTTP/2, TCP).

Framework is the higher-level tools and conventions for building applications (like gRPC).

Passing gRPC Connector Instance to Stub

---

import grpc 
from your_service_pb2_grpc import DataServiceStub
 
# Step 1: Pass grpc connector instance to the stub 
channel = grpc.insecure_channel('localhost:50051') 
stub = DataServiceStub(channel)

• Purpose: Establishes the network connection and creates a stub instance
• The connector handles the underlying HTTP/2 connection, authentication, and network details
• The stub acts as a local object that represents the remote service

Craft Request Instance with Parameters

---

from your_service_pb2 import DataRequest
from google.protobuf.field_mask_pb2 import FieldMask
from google.protobuf import timestamp_pb2
import filter_pb2 
 
# Step 2: Craft a request instance with parameters
filter_obj = filter_pb2.Filter()
filter_obj.column = "sensor_id"
filter_obj.op = filter_pb2.EQUALS
filter_obj.str = "123"
 
start_time = timestamp_pb2.Timestamp()
start_time.FromDatetime(datetime(2025, 1, 1))
 
request = DataRequest(
    start_datetime=start_time,
    end_datetime=timestamp_pb2.Timestamp(),
    field_mask=FieldMask(paths=["temperature", "humidity"]),
    filters=[filter_obj]
)

• Purpose: Creates a strongly-typed message object containing your query parameters
• Structures your data (start_datetime, end_datetime, filters) according to the service’s protobuf schema
• Ensures type safety and proper serialization before sending over the network

Filter building

filter_pb2.Filter() is constructor for filter criteria.

• .column specifies which field to filter on
• .op specifies operation enum (EQUALS, GREATER_THAN, etc.)
• .str specifies a single string value
• .str_list_list specifies complex list-of-lists for IN .op clauses

I like to have a function for each filter type, like

• create_not_null_filter(column: str) -> filter_pb2.Filter
• create_equals_filter(column: str, value: str) -> filter_pb2.Filter
• create_in_list_filter(column: str, values: list[int]) -> filter_pb2.Filter

Then for a particular data, I can just put a combination of the these functions in a list [] and assign it to the filters of the DataRequest.

Common Types

FieldMask specifies which response fields to return. timestamp_pb2.Timestamp() is standard timestamp handling.

Pass Request Instance to Lib Function for Response

---

# Step 3: Pass request to lib function for response 
response = stub.GetData(request)

• Purpose: Executes the actual remote procedure call
• The library serializes your request, sends it over the network, waits for the response, and deserializes the result back into a typed object
• Handles the low-level gRPC communication protocol transparently