Wire Protocol Buffers

“A man got to have a code!” - Omar Little

As our teams and programs grow, the variety and volume of data also grows. Success will turn your simple data models into complex ones! Whether your application is storing data to disk or transmitting it over a network, the structure and interpretation of that data should be clear. Consumers work best with data they understand!

Schemas describe and document data models. If you have data, you should have a schema.

Protocol Buffers

Google's Protocol Buffers are built around a great schema language:

• It's cross platform and language independent. Whatever programming language you use, you'll be able to use proto schemas with your application.

• Proto schemas are backwards-compatible and future-proof. You can evolve your schema as your application loses old features and gains new ones.

• It's focused. Proto schemas describe your data models. That's it.

Here's a sample message definition:

syntax = "proto2";

package squareup.dinosaurs;

option java_package = "com.squareup.dinosaurs";

import "squareup/geology/period.proto";

message Dinosaur {
  // Common name of this dinosaur, like "Stegosaurus".
  optional string name = 1;

  // URLs with images of this dinosaur.
  repeated string picture_urls = 2;

  optional squareup.geology.Period period = 5;
}

And here's an enum definition:

syntax = "proto2";

package squareup.geology;

option java_package = "com.squareup.geology";

enum Period {
  // 145.5 million years ago — 66.0 million years ago.
  CRETACEOUS = 1;

  // 201.3 million years ago — 145.0 million years ago.
  JURASSIC = 2;

  // 252.17 million years ago — 201.3 million years ago.
  TRIASSIC = 3;
}

This schema language is Protocol Buffers' best feature. You might even use it purely for documentation purposes, such as to describe a JSON API.

Protocol Buffers also defines a compact binary encoding of messages that conform to the schema. This encoding is fast to encode, fast to decode, small to transmit, and small to store. The binary encoding uses numeric tags from the schema, like the 5 for period above.

For example, let's encode this dinosaur:

{
  name: "Stegosaurus",
  period: JURASSIC
}

The encoded value is just 15 bytes:

Hex  Description
 0a  tag: name(1), field encoding: LENGTH_DELIMITED(2). 1 << 3 | 2
 0b  "Stegosaurus".length()
 53  'S'
 74  't'
 65  'e'
 67  'g'
 6f  'o'
 73  's'
 61  'a'
 75  'u'
 72  'r'
 75  'u'
 73  's'
 28  tag: period(5), field encoding: VARINT(0). 5 << 3 | 0
 02  JURASSIC(2)

Why Wire?

The Protocol Buffers schema language and binary encoding are both defined by Google. Wire is an independent implementation from Square that's specifically designed for Android and Java.

For each message type defined in the schema, Wire generates an immutable model class and its builder. The generated code looks like code you'd write by hand: it's documented, formatted, and simple. Wire's APIs should feel at home to programmers who like Effective Java.

That said, there are some interesting design decisions in Wire:

• Wire messages declare public final fields instead of the usual getter methods. This cuts down on both code generated and code executed. Less code is particularly beneficial for Android programs.

• Wire avoids case mapping. A field declared as picture_urls in a schema yields a Java field picture_urls and not the conventional pictureUrls camel case. Though the name feels awkward at first, it's fantastic whenever you use grep or more sophisticated search tools. No more mapping when navigating between schema, Java source code, and data. It also provides a gentle reminder to calling code that proto messages are a bit special.

• Primitive types are always boxed. If a field is absent, its value is null. This is used for naturally optional fields, such as a dinosaur whose period is unknown. A field may also be null due to schema evolution: if tomorrow we add a carnivore boolean to our message definition, today's data won’t have a value for that field.

Here's the compact generated code for the Dinosaur message defined above:

// Code generated by Wire protocol buffer compiler, do not edit.
// Source file: squareup/dinosaurs/dinosaur.proto at 9:1
package com.squareup.dinosaurs;

import com.squareup.geology.Period;
import com.squareup.wire.Message;
import com.squareup.wire.ProtoAdapter;
import com.squareup.wire.WireField;
import java.util.List;
import okio.ByteString;

public final class Dinosaur extends Message<Dinosaur, Dinosaur.Builder> {
  public static final ProtoAdapter<Dinosaur> ADAPTER = ProtoAdapter.newMessageAdapter(Dinosaur.class);

  private static final long serialVersionUID = 0L;

  public static final String DEFAULT_NAME = "";

  public static final Period DEFAULT_PERIOD = Period.CRETACEOUS;

  /**
   * Common name of this dinosaur, like "Stegosaurus".
   */
  @WireField(
      tag = 1,
      adapter = "com.squareup.wire.ProtoAdapter#STRING"
  )
  public final String name;

  /**
   * URLs with images of this dinosaur.
   */
  @WireField(
      tag = 2,
      adapter = "com.squareup.wire.ProtoAdapter#STRING",
      label = WireField.Label.REPEATED
  )
  public final List<String> picture_urls;

  @WireField(
      tag = 5,
      adapter = "com.squareup.geology.Period#ADAPTER"
  )
  public final Period period;

  public Dinosaur(String name, List<String> picture_urls, Period period) {
    this(name, picture_urls, period, ByteString.EMPTY);
  }

  public Dinosaur(String name, List<String> picture_urls, Period period, ByteString unknownFields) {
    super(unknownFields);
    this.name = name;
    this.picture_urls = immutableCopyOf("picture_urls", picture_urls);
    this.period = period;
  }

  @Override
  public Builder newBuilder() {
    Builder builder = new Builder();
    builder.name = name;
    builder.picture_urls = copyOf("picture_urls", picture_urls);
    builder.period = period;
    builder.addUnknownFields(unknownFields());
    return builder;
  }

  @Override
  public boolean equals(Object other) {
    if (other == this) return true;
    if (!(other instanceof Dinosaur)) return false;
    Dinosaur o = (Dinosaur) other;
    return equals(unknownFields(), o.unknownFields())
        && equals(name, o.name)
        && equals(picture_urls, o.picture_urls)
        && equals(period, o.period);
  }

  @Override
  public int hashCode() {
    int result = super.hashCode;
    if (result == 0) {
      result = unknownFields().hashCode();
      result = result * 37 + (name != null ? name.hashCode() : 0);
      result = result * 37 + (picture_urls != null ? picture_urls.hashCode() : 1);
      result = result * 37 + (period != null ? period.hashCode() : 0);
      super.hashCode = result;
    }
    return result;
  }

  public static final class Builder extends com.squareup.wire.Message.Builder<Dinosaur, Builder> {
    public String name;

    public List<String> picture_urls;

    public Period period;

    public Builder() {
      picture_urls = newMutableList();
    }

    /**
     * Common name of this dinosaur, like "Stegosaurus".
     */
    public Builder name(String name) {
      this.name = name;
      return this;
    }

    /**
     * URLs with images of this dinosaur.
     */
    public Builder picture_urls(List<String> picture_urls) {
      checkElementsNotNull(picture_urls);
      this.picture_urls = picture_urls;
      return this;
    }

    public Builder period(Period period) {
      this.period = period;
      return this;
    }

    @Override
    public Dinosaur build() {
      return new Dinosaur(name, picture_urls, period, buildUnknownFields());
    }
  }
}

The Java code to create and access proto models is compact and readable:

Dinosaur stegosaurus = new Dinosaur.Builder()
    .name("Stegosaurus")
    .period(Period.JURASSIC)
    .build();

System.out.println("My favorite dinosaur existed in the " + stegosaurus.period + " period.");

Each type has a corresponding ProtoAdapter that can encode a message to bytes and decode bytes back into a message.

Dinosaur stegosaurus = ...
byte[] stegosaurusBytes = Dinosaur.ADAPTER.encode(stegosaurus);

byte[] tyrannosaurusBytes = ...
Dinosaur tyrannosaurus = Dinosaur.ADAPTER.decode(tyrannosaurusBytes);

When accessing a field, use Wire.get() to replace null values with the corresponding default:

Period period = Wire.get(stegosaurus.period, Dinosaur.DEFAULT_PERIOD);

This is equivalent to the following:

Period period = stegosaurus.period != null ? stegosaurus.period : Dinosaur.DEFAULT_PERIOD;

Generating Code With Wire

Wire's compiler is available via a Maven plugin. Put .proto sources in your project's src/main/proto directory, then use the plugin to generate .java files. The plugin will automatically add the generated Java code to your project's source roots.

<build>
  <plugins>
    <plugin>
      <groupId>com.squareup.wire</groupId>
      <artifactId>wire-maven-plugin</artifactId>
      <version>${project.version}</version>
      <executions>
        <execution>
          <phase>generate-sources</phase>
          <goals>
            <goal>generate-sources</goal>
          </goals>
          <configuration>
            <includes>
              <!-- proto package names to generate code for -->
              <include>squareup.dinosaurs.*</include>
              <include>squareup.geology.*</include>
            </includes>
          </configuration>
        </execution>
      </executions>
    </plugin>
  </plugins>
</build>

Wire can read .proto files from the local file system and from within .jar files.

The compiler can optionally prune your schema to a subset of root types and their transitive dependencies. This is useful when sharing a schema between projects: a Java service and Android app may each use a subset of a larger shared schema.

If you don't use Maven, the compiler also has a command line interface. Just substitute wire-compiler-VERSION-jar-with-dependencies.jar with the path to your jar.

% java -jar wire-compiler-VERSION-jar-with-dependencies.jar \
    --proto_path=src/main/proto \
    --java_out=out \
    squareup/dinosaurs/dinosaur.proto \
    squareup/geology/period.proto
Writing com.squareup.dinosaurs.Dinosaur to out
Writing com.squareup.geology.Period to out

If you use Proguard, then you need to add keep rules. The simplest option is to tell Proguard not to touch the Wire runtime library and your generated protocol buffers (of course these simple rules will miss opportunities to shrink and optimize the code):

-keep class com.squareup.wire.** { *; }
-keep class com.yourcompany.yourgeneratedcode.** { *; }

Get Wire

The wire-runtime package contains runtime support libraries that must be included in applications that use Wire-generated code.

With Maven:

<dependency>
  <groupId>com.squareup.wire</groupId>
  <artifactId>wire-runtime</artifactId>
  <version>2.2.0</version>
</dependency>

With Gradle:

compile 'com.squareup.wire:wire-runtime:2.2.0'

Snapshots of the development version are available in Sonatype's snapshots repository.

Unsupported

Wire does not support:

• Groups - they are skipped when parsing binary input data

Wire supports custom options on messages and fields. Other custom options are ignored. Pass --excludes=google.protobuf.* to the compiler to omit options from the generated code.

Further Documentation

See Google's excellent documentation on the structure and syntax of proto schemas.