ts-proto

What is ts-proto?

The ts-proto npm package is a TypeScript code generator for Protocol Buffers (protobufs). It allows you to generate TypeScript types and client code from .proto files, making it easier to work with protobufs in TypeScript projects.

What are ts-proto's main functionalities?

Generate TypeScript Types

This feature allows you to generate TypeScript types from your .proto files. The generated types can be used in your TypeScript code to ensure type safety when working with protobuf messages.

protoc --plugin=protoc-gen-ts_proto=./node_modules/.bin/protoc-gen-ts_proto --ts_proto_out=. your_proto_file.proto

Generate gRPC Client Code

This feature generates gRPC client code for your .proto files. The generated client code can be used to make gRPC calls from your TypeScript code.

protoc --plugin=protoc-gen-ts_proto=./node_modules/.bin/protoc-gen-ts_proto --ts_proto_out=. --ts_proto_opt=outputServices=grpc-js your_proto_file.proto

Generate JSON Serialization Code

This feature generates JSON serialization and deserialization methods for your protobuf messages. This is useful for converting protobuf messages to and from JSON format.

protoc --plugin=protoc-gen-ts_proto=./node_modules/.bin/protoc-gen-ts_proto --ts_proto_out=. --ts_proto_opt=outputJsonMethods=true your_proto_file.proto

Other packages similar to ts-proto

protobufjs

protobufjs is a popular library for working with Protocol Buffers in JavaScript and TypeScript. It provides a runtime library for parsing and serializing protobuf messages, as well as a code generator for generating TypeScript definitions. Unlike ts-proto, protobufjs does not generate gRPC client code.

grpc-tools

grpc-tools is a package that provides the protoc compiler and gRPC plugins for generating gRPC client and server code. It supports multiple languages, including JavaScript and TypeScript. However, grpc-tools does not generate TypeScript types directly from .proto files, which is a key feature of ts-proto.

grpc-web

grpc-web is a JavaScript implementation of gRPC for browser clients. It allows you to use gRPC in web applications. While grpc-web focuses on enabling gRPC in the browser, ts-proto focuses on generating TypeScript types and client code for Node.js and browser environments.

README

Goals

• Pure/idiomatic TypeScript/ES6 modules
• Only supports codegen *.proto-to-*.ts workflow, currently no runtime reflection/loading of dynamic .proto files
• Currently ambivalent about browser support, current focus is on Node/server-side use cases

Highlights

• Wrapper types, i.e. google.protobuf.StringValue, are mapped as optional values, i.e. string | undefined
• Timestamp is mapped as Date
• fromJSON/toJSON support the canonical Protobuf JS format (i.e. timestamps are ISO strings)

Assumptions

• TS/ES6 module name is the proto package

Todo

• Better Long support; currently any values greater than Number.MAX_SAFE_INTEGER blow up at runtime
• Model OneOfs as an ADT
• Support the string-based encoding of duration in fromJSON/toJSON
• Support bytes as base64 encoded strings in fromJSON/toJSON
• Support the json_name annotation

Typing Approach

• Missing fields on read
  • When decoding from binary, we setup a prototype for our returned object, which has default values.
    • This assumes missing keys trigger the default value, e.g. storing key=undefined would subvert the approach
  • When decoding from JSON, we may have missing keys.
    • We could convert them to our prototype.
  • When using an instantiated object, our types enforce all keys to be set.

OneOf Handling

Currently fields that are modeled with oneof either_field { string field_a; string field_b } are generated as field_a: string | undefined; field_b: string | undefined.

This means you'll have to check if object.field_a and if object.field_b, and if you set one, you'll have to remember to unset the other.

It would be nice/preferable to model this as an ADT, so it would be:

object.either_field = { kind: 'field_a', value: 'name' };

However this differs sufficiently from the wire-level format that there might be wrinkles.

An original design notion of ts-proto was that ideally we could get JSON off the wire and immediately cast it to the generated ts-proto types, but features like oneof ADTs require walking the JSON looking for things to massage.

Similarily, writing a ts-proto object as protobuf-compliant JSON would not be a straight JSON.stringify(tsProtoObject).

(Idea: maybe either_field exists in the prototype, and wraps/manages the underlying primitive values.)

Primitive Types

Protobuf has the somewhat annoying behavior that primitives types cannot differentiate between set-to-defalut-value and unset.

I.e. if you have a string name = 1, and set object.name = '', Protobuf will skip sending the tagged name field over the wire, because its understood that readers on the other end will, when they see name is not included in the payload, return empty string.

ts-proto models this behavior, of "unset" values being the primitive's default. (Technically by setting up an object prototype that knows the default values of the message's primitive fields.)

If you want fields where you can model set/unset, see Wrapper Types.

Wrapper Types

In core Protobuf, while unset primitives are read as default values, unset messages are returned as null.

This allows a cute hack where you can model a logical string | null by creating a field that is a message (can be null) and the message has a single string value (for when the value is not null).

Protobuf has several built-in types for this pattern, i.e. google.protobuf.StringValue.

ts-proto understands these wrapper types and will generate google.protobuf.StringValue name = 1 as a name: string | undefined.

This hides some of the StringValue mess and gives a more idiomatic way of access them.

Granted, it's unfortunate this is not as simple as marking the string as optional.

Current Status of Optional Values

• Required primitives: use as-is, i.e. string name = 1.
• Optional primitives: use wrapper types, i.e. StringValue name = 1.
• Required messages: not available
• Optional primitives: use as-is, i.e. SubMessage message = 1.