protoc-gen-fake: Protocol Buffer Fake Data Generator

protoc-gen-fake is a protoc plugin that generates fake data based on your Protocol Buffer schema definitions and custom annotations.

The fake data it generates is binary data in exactly the same format as defined by that schema. It helps developers quickly create realistic mock data at scale for testing, development, and demonstrations.

Features

• Schema-driven Data Generation: Generates fake data directly from your .proto file definitions.
• Customizable Fake Data: Use (gen_fake.fake_data) options to provide specific data types (e.g., names, addresses, emails), and minimum and maximum count for repeated or optional fields.
• Internationalization (i18n): Generate fake data in various languages and locales.
• Flexible Output: Choose between binary or Base64 encoded output, and control output paths.

Table of Contents

• Installation

• Usage

  • Direct protoc Calls
  • gen_fake.sh Wrapper Script
  • Generating Language-Specific Protobuf Files
  • Utilizing Generated Fake Data

• Configuration

  • Message-Level Options
  • Field-Level Options

• Internationalization (i18n)

• Contributing

Installation

protoc-gen-fake can be installed in several ways, depending on your preferences and development setup.

From GitHub Releases (Recommended for most users)

The easiest way to get protoc-gen-fake is to download the pre-compiled binary for your operating system from the GitHub Releases page.

1. Download: Go to the Releases page and download the appropriate .zip or .tar.gz file for your system (e.g., protoc-gen-fake-darwin-arm64 for macOS Apple Silicon).

2. Extract: Extract the downloaded archive.

3. Add to PATH: Move the protoc-gen-fake executable to a directory in your system's $PATH (e.g., /usr/local/bin or ~/.cargo/bin).

   # Example for macOS/Linux
   mv /path/to/downloaded/protoc-gen-fake /usr/local/bin/
   

   Ensure the chosen directory is in your $PATH. If not, you might need to add it to your shell's configuration file (e.g., .bashrc, .zshrc, or .profile):

   export PATH="/usr/local/bin:$PATH"
   

Via Cargo (For Rust developers)

If you have Rust and Cargo installed, you can install protoc-gen-fake directly from crates.io:

cargo install protoc-gen-fake

This will compile the plugin from source and place the executable in ~/.cargo/bin/, which should already be in your $PATH.

From Source

If you prefer to build the plugin yourself:

1. Clone the repository:

   git clone https://github.com/mike-williamson/protoc-gen-fake.git
   cd protoc-gen-fake
   

2. Build in release mode:

   cargo build --release
   

   This will create an executable named protoc-gen-fake in the target/release/ directory.

3. Copy the executable to your $PATH:

   You can copy the binary to a directory that is already in your $PATH, such as ~/.cargo/bin/ (or /usr/local/bin):

   cp target/release/protoc-gen-fake ~/.cargo/bin/
   

Via Homebrew (For macOS & Linux users - Coming Soon)

A Homebrew formula will be available soon for easy installation on macOS and Linux:

brew install protoc-gen-fake # Placeholder command

Once installed, you should be able to invoke protoc-gen-fake directly via protoc.

Usage

Direct protoc Calls

When using protoc-gen-fake directly with protoc, you need to specify the output directory for the generated fake data. For binary output, protoc does not pass the output directory to the plugin, so you must provide it explicitly using the --fake_opt output_path parameter.

protoc --proto_path=proto \
       --fake_out=. \
       --fake_opt output_path=mike_data/full_customer.bin \
       proto/examples/full_customer.proto

• --proto_path=proto: Specifies the directory where your .proto files are located.
• --fake_out=.: This tells protoc to invoke protoc-gen-fake. The value here (.) is largely ignored by protoc-gen-fake for binary output, but it's required by protoc.
• --fake_opt output_path=mike_data/full_customer.bin: This is the crucial part for binary output. It tells protoc-gen-fake where to write the generated binary fake data file.
• proto/examples/full_customer.proto: The .proto file for which to generate fake data.

To generate Base64 encoded output directly in the CodeGeneratorResponse (which protoc then writes to stdout or a file if --fake_out points to a file), you can specify encoding=Base64:

protoc --proto_path=proto \
       --fake_out=base64_output.txt \
       --fake_opt encoding=Base64 \
       proto/examples/full_customer.proto

gen_fake.sh Wrapper Script

To simplify the command-line usage and avoid specifying the output path twice for binary output, you can use the provided gen_fake.sh wrapper script. This script handles the internal translation of a single --out_path argument into the necessary protoc and --fake_opt parameters.

./gen_fake.sh --out_path=mike_data/full_customer.bin proto/examples/full_customer.proto

• --out_path=mike_data/full_customer.bin: Specifies the desired output path for the generated binary fake data.
• proto/examples/full_customer.proto: The .proto file for which to generate fake data.

Generating Language-Specific Protobuf Files

Before you can utilize the fake data in your application, you often need to generate language-specific protobuf classes from your .proto definitions. Here are examples for Python and Go.

Python

protoc --proto_path=proto \
       --python_out=examples \
       proto/examples/full_customer.proto \
       proto/gen_fake/fake_field.proto

• --python_out=examples: Specifies the output directory for the generated Python protobuf files.

Go

(Note: This requires protoc-gen-go to be installed and in your $PATH)

protoc --proto_path=proto \
       --go_out=examples \
       --go_opt=paths=source_relative \
       proto/examples/full_customer.proto \
       proto/gen_fake/fake_field.proto

• --go_out=examples: Specifies the output directory for the generated Go protobuf files.
• --go_opt=paths=source_relative: Ensures that generated Go files have correct import paths.

Utilizing Generated Fake Data

Once you have generated both the fake data (binary) and the language-specific protobuf classes, you can load and use the fake data in your application.

Python

Assuming you have generated full_customer_pb2.py into the examples directory and full_customer.bin into mike_data/:

import sys
sys.path.append('examples') # Add the directory where protobuf files are generated

from full_customer_pb2 import FullCustomer

def load_fake_customer(file_path):
    with open(file_path, 'rb') as f:
        data = f.read()
    customer = FullCustomer()
    customer.ParseFromString(data)
    return customer

if __name__ == '__main__':
    fake_customer = load_fake_customer('mike_data/full_customer.bin')
    print(fake_customer)

Go

Assuming you have generated full_customer.pb.go into the examples directory and full_customer.bin into mike_data/:

package main

import (
	"fmt"
	"io/ioutil"
	"log"
	"path/filepath"

	"google.golang.org/protobuf/proto"

	examples "your_module_path/examples" // Replace with your actual module path
)

func main() {
	filePath := filepath.Join("mike_data", "full_customer.bin")
	data, err := ioutil.ReadFile(filePath)
	if err != nil {
		log.Fatalf("Failed to read fake data file: %v", err)
	}

	customer := &examples.FullCustomer{} // Assuming FullCustomer is in your examples package
	if err := proto.Unmarshal(data, customer); err != nil {
		log.Fatalf("Failed to unmarshal fake data: %v", err)
	}

	fmt.Printf("Loaded Fake Customer: %+v\n", customer)
}

Configuration

protoc-gen-fake uses custom protobuf options to configure fake data generation for messages and fields.

Message-Level Options

• option (gen_fake.fake_msg).include = true;

  • Purpose: By default, protoc-gen-fake will only generate data for the first top-level message defined in a .proto file. To explicitly include a specific message for fake data generation (especially useful if you have multiple top-level messages and want to target one, or if it's a nested message you want to ensure is generated as a root object), set this option to true. This effectively marks the message as an "entry point" for generation.

  message MyMessage {
    option (gen_fake.fake_msg).include = true;
    string id = 1 [(gen_fake.fake_data).uuid = true];
  }
  

Field-Level Options ((gen_fake.fake_data) = {...})

These options are applied to individual fields within a message to control the type and characteristics of the fake data generated.

• Enabling Default Fake Data:

  • [(gen_fake.fake_data) = {}]
    • Purpose: To opt-in a field for default fake data generation. If no specific fake data type (like email, uuid, etc.) is provided, the plugin will attempt to generate a sensible default based on the field's protobuf type (e.g., random string for string, random number for int32).

  message User {
    string name = 1 [(gen_fake.fake_data) = {}]; // Generates a default fake string
  }
  

• Specifying Fake Data Types:

  • You can specify a wide range of fake data types using the options within (gen_fake.fake_data).
  • Example: Email and UUID

  message User {
    string email = 1 [(gen_fake.fake_data).email = true];
    string id = 2 [(gen_fake.fake_data).uuid = true];
  }
  

• Controlling Optional Field Generation:

  • [(gen_fake.fake_data).min_count = 1]
    • Purpose: For optional fields, by default, protoc-gen-fake might or might not generate a value. Setting min_count = 1 ensures that the optional field will always be generated.

  message Product {
    optional string description = 1 [(gen_fake.fake_data).min_count = 1]; // Always generate a description
  }
  

• Nested Messages:

  • protoc-gen-fake automatically recurses into nested messages. You apply field-level options to fields within the nested messages as usual.

  message Address {
    string street = 1 [(gen_fake.fake_data).street_name = true];
    string city = 2 [(gen_fake.fake_data).city = true];
  }
  
  message Customer {
    string name = 1 [(gen_fake.fake_data).full_name = true];
    Address home_address = 2 [(gen_fake.fake_data) = {}]; // Generates a fake Address
  }
  

  In the example above, home_address will have a fake Address generated for it, with its street and city fields populated according to their respective fake_data options.

Internationalization (i18n)

protoc-gen-fake supports generating localized fake data by specifying a language for specific fields using the language option.

message UserProfile {
  string given_name = 1 [(gen_fake.fake_data).given_name = true, (gen_fake.fake_data).language = "ZH_TW"];
  string address_street = 2 [(gen_fake.fake_data).street_name = true, (gen_fake.fake_data).language = "AR_SA"];
}

• language = "ZH_TW": Generates a given name in Traditional Chinese.
• language = "AR_SA": Generates a street name in Arabic (Saudi Arabia).

Refer to the fake crate's documentation for a full list of supported locales and data providers.

Contributing

Contributions are welcome! Please feel free to open issues or submit pull requests.