Skip to main content

Add your description here

Project description

FastServe

FastServe is a Python library that lets you build gRPC services quickly and easily without hand-writing protobuf files.

It automatically generates protobuf files that match the method signature of the Python object provided when starting the server.

Then, it uses the grpcio-tool to generate modules include server and client stub classes from the created protobuf files.

By dynamically generating a subclass of the server stub class and associating its methods with the original Python object's methods, it builds the gRPC server.

Installation

Install it using your favorite package manager. Below is the method for installation using pip.

pip install fastserve

Usage

Here's an example of creating a dead simple greeter service.

greeting.py

from fastserve.core import Server, Message


class HelloRequest(Message):
    name: str


class HelloReply(Message):
    message: str


class Greeter:
    def say_hello(self, request: HelloRequest) -> HelloReply:
        return HelloReply(message=f"Hello, {request.name}!")


if __name__ == "__main__":
    s = Server()
    s.run(Greeter())

To start the server, run:

$ python greeting.py
gRPC server is running...

In another terminal, you can call the gRPC service:

$ grpcurl -plaintext -d '{"name": "World"}' localhost:50051 greeter.v1.Greeter/SayHello
{
  "message": "Hello, World!"
}

To list the services, use:

$ grpcurl -plaintext localhost:50051 list
greeter.v1.Greeter
grpc.health.v1.Health
grpc.reflection.v1alpha.ServerReflection

By the way, running this will create greeter.proto, greeter_pb2.py, and greeter_pb2_grpc.py in the current directory. The generated greeter.proto will look like this:

syntax = "proto3";

package greeter.v1;

service Greeter {
    rpc SayHello (HelloRequest) returns (HelloReply);
}

message HelloReply {
    string message = 1;
}

message HelloRequest {
    string name = 1;
}

You can share the generated greeter.proto with other services/clients written in other programming languages or Python if you want or need.

If You Want to Use Existing Stub Modules

If you already have greeter_pb2.py and greeter_pb2_grpc.py and want to use them without generating, use mount_using_pb2_modules as shown below:

from fastserve.core import Server, Message

import greeter_pb2_grpc, greeter_pb2


class HelloRequest(Message):
    name: str


class HelloReply(Message):
    message: str


class Greeter:
    def say_hello(self, request: HelloRequest) -> HelloReply:
        return HelloReply(message=f"Hello, {request.name}!")


if __name__ == "__main__":
    s = Server()
    s.mount_using_pb2_modules(greeter_pb2_grpc, greeter_pb2, Greeter())
    s.run()

asyncio Version

import asyncio

from fastserve.core import AsyncIOServer, Message


class HelloRequest(Message):
    name: str


class HelloReply(Message):
    message: str


class Greeter:
    async def say_hello(self, request: HelloRequest) -> HelloReply:
        return HelloReply(message=f"Hello, {request.name}!")


if __name__ == "__main__":
    s = AsyncIOServer()
    loop = asyncio.get_event_loop()
    loop.run_until_complete(s.run(Greeter()))

If You Only Want to Generate protobuf Files and Stub Modules

You can execute the fastserve.core module to only generate the protobuf files and stub modules as follows:

$ python -m fastserve.core greeting.py Greeter

Doing so will create greeter.proto, greeter_pb2.py, and greeter_pb2_grpc.py in the current directory.

Note

  • Yuo can serve multiple services by passing multiple objects to Server.run or calling Server.mount multiple times.
  • You can specify the port number with Server.set_port method before calling run.
  • You can set interceptors with Server constructor's interceptors argument.

Mapping of Python Types to Protobuf Types

The following table shows the mapping of Python types to protobuf types.

Python Type Protobuf Type
int int32
float float
str string
bytes bytes
bool bool
List[T] repeated T
Dict[str, T] map<string, T>
Dict[int, T] map<int32, T>
fastserve.core.Message message

TODO

  • Support for mapping of types.UnionType to oneof
  • Support for mapping of typing.Optional or typing.Union[T, None] to oneof
  • Support for mapping of enum.Enum to enum
  • Support for mapping of datetime.datetime to google.protobuf.Timestamp
  • Support for mapping of datetime.timedelta to google.protobuf.Duration

License

MIT License

Author

Yasushi Itoh (i2y)

Project details


Download files

Download the file for your platform. If you're not sure which to choose, learn more about installing packages.

Source Distribution

fastserve-0.1.0.tar.gz (10.4 kB view hashes)

Uploaded Source

Built Distribution

fastserve-0.1.0-py3-none-any.whl (8.5 kB view hashes)

Uploaded Python 3

Supported by

AWS AWS Cloud computing and Security Sponsor Datadog Datadog Monitoring Fastly Fastly CDN Google Google Download Analytics Microsoft Microsoft PSF Sponsor Pingdom Pingdom Monitoring Sentry Sentry Error logging StatusPage StatusPage Status page