Test code generation prompts

Caution: The Pathways Language Models (PaLM) are deprecated. As of October 9, 2024, you can no longer access these models from new Google Cloud projects, and access for all projects will be removed on April 9, 2025. If you're a new user, we recommend using Gemini. If you're an existing PaLM user, see Migrate from PaLM 2 to Gemini.

To design a prompt that works well, test different versions of the prompt and experiment with prompt parameters to determine what results in the optimal response. You can test prompts programmatically with the Codey APIs and in the Google Cloud console with Vertex AI Studio.

Test code generation prompts

To test code generation prompts, choose one of the following methods.

REST

To test a code generation prompt with the Vertex AI API, send a POST request to the publisher model endpoint.

Before using any of the request data, make the following replacements:

• PROJECT_ID: Your project ID.
• PREFIX: For code models, prefix represents the beginning of a piece of meaningful programming code or a natural language prompt that describes code to be generated.
• TEMPERATURE: The temperature is used for sampling during response generation. Temperature controls the degree of randomness in token selection. Lower temperatures are good for prompts that require a less open-ended or creative response, while higher temperatures can lead to more diverse or creative results. A temperature of 0 means that the highest probability tokens are always selected. In this case, responses for a given prompt are mostly deterministic, but a small amount of variation is still possible.
• MAX_OUTPUT_TOKENS: Maximum number of tokens that can be generated in the response. A token is approximately four characters. 100 tokens correspond to roughly 60-80 words.

  Specify a lower value for shorter responses and a higher value for potentially longer responses.

• CANDIDATE_COUNT: The number of response variations to return. For each request, you're charged for the output tokens of all candidates, but are only charged once for the input tokens.

  Specifying multiple candidates is a Preview feature that works with generateContent (streamGenerateContent is not supported). The following models are supported:

  • Gemini 1.5 Flash: 1-8, default: 1
  • Gemini 1.5 Pro: 1-8, default: 1
  • Gemini 1.0 Pro: 1-8, default: 1
  The range of valid values is an int between 1 and 4.

HTTP method and URL:

POST https://us-central1-aiplatform.googleapis.com/v1/projects/PROJECT_ID/locations/us-central1/publishers/google/models/code-bison:predict

Request JSON body:

{
  "instances": [
    { "prefix": "PREFIX" }
  ],
  "parameters": {
    "temperature": TEMPERATURE,
    "maxOutputTokens": MAX_OUTPUT_TOKENS,
    "candidateCount": CANDIDATE_COUNT
  }
}

To send your request, choose one of these options:

curl

Note: The following command assumes that you have logged in to the gcloud CLI with your user account by running gcloud init or gcloud auth login , or by using Cloud Shell, which automatically logs you into the gcloud CLI . You can check the currently active account by running gcloud auth list.

Save the request body in a file named request.json, and execute the following command:

curl -X POST \
     -H "Authorization: Bearer $(gcloud auth print-access-token)" \
     -H "Content-Type: application/json; charset=utf-8" \
     -d @request.json \
     "https://us-central1-aiplatform.googleapis.com/v1/projects/PROJECT_ID/locations/us-central1/publishers/google/models/code-bison:predict"

PowerShell

Note: The following command assumes that you have logged in to the gcloud CLI with your user account by running gcloud init or gcloud auth login . You can check the currently active account by running gcloud auth list.

Save the request body in a file named request.json, and execute the following command:

$cred = gcloud auth print-access-token
$headers = @{ "Authorization" = "Bearer $cred" }

Invoke-WebRequest `
    -Method POST `
    -Headers $headers `
    -ContentType: "application/json; charset=utf-8" `
    -InFile request.json `
    -Uri "https://us-central1-aiplatform.googleapis.com/v1/projects/PROJECT_ID/locations/us-central1/publishers/google/models/code-bison:predict" | Select-Object -Expand Content

You should receive a JSON response similar to the following.

Response

{
  "predictions": [
    {
      "citationMetadata": {
        "citations": []
      },
      "content": "RESPONSE",
      "safetyAttributes": {
        "blocked": false,
        "categories": [],
        "scores": []
      },
      "score": float
    }
  ]
}

Python

To learn how to install or update the Vertex AI SDK for Python, see Install the Vertex AI SDK for Python. For more information, see the Python API reference documentation.

from vertexai.language_models import CodeGenerationModel

parameters = {
    "temperature": 0.1,  # Temperature controls the degree of randomness in token selection.
    "max_output_tokens": 256,  # Token limit determines the maximum amount of text output.
}

code_generation_model = CodeGenerationModel.from_pretrained("code-bison@001")
response = code_generation_model.predict(
    prefix="Write a function that checks if a year is a leap year.", **parameters
)

print(f"Response from Model: {response.text}")
# Example response:
# Response from Model: I will write a function to check if a year is a leap year.
# **The function will take a year as input and return a boolean value**.
# **The function will first check if the year is divisible by 4.**
# ...

return response

Node.js

Before trying this sample, follow the Node.js setup instructions in the Vertex AI quickstart using client libraries. For more information, see the Vertex AI Node.js API reference documentation.

To authenticate to Vertex AI, set up Application Default Credentials. For more information, see Set up authentication for a local development environment.

/**
 * TODO(developer): Uncomment these variables before running the sample.\
 * (Not necessary if passing values as arguments)
 */
// const project = 'YOUR_PROJECT_ID';
// const location = 'YOUR_PROJECT_LOCATION';
const aiplatform = require('@google-cloud/aiplatform');

// Imports the Google Cloud Prediction service client
const {PredictionServiceClient} = aiplatform.v1;

// Import the helper module for converting arbitrary protobuf.Value objects.
const {helpers} = aiplatform;

// Specifies the location of the api endpoint
const clientOptions = {
  apiEndpoint: 'us-central1-aiplatform.googleapis.com',
};
const publisher = 'google';
const model = 'code-bison@001';

// Instantiates a client
const predictionServiceClient = new PredictionServiceClient(clientOptions);

async function callPredict() {
  // Configure the parent resource
  const endpoint = `projects/${project}/locations/${location}/publishers/${publisher}/models/${model}`;

  const prompt = {
    prefix: 'Write a function that checks if a year is a leap year.',
  };
  const instanceValue = helpers.toValue(prompt);
  const instances = [instanceValue];

  const parameter = {
    temperature: 0.5,
    maxOutputTokens: 256,
  };
  const parameters = helpers.toValue(parameter);

  const request = {
    endpoint,
    instances,
    parameters,
  };

  // Predict request
  const [response] = await predictionServiceClient.predict(request);
  console.log('Get code generation response');
  const predictions = response.predictions;
  console.log('\tPredictions :');
  for (const prediction of predictions) {
    console.log(`\t\tPrediction : ${JSON.stringify(prediction)}`);
  }
}

callPredict();

Java

Before trying this sample, follow the Java setup instructions in the Vertex AI quickstart using client libraries. For more information, see the Vertex AI Java API reference documentation.

To authenticate to Vertex AI, set up Application Default Credentials. For more information, see Set up authentication for a local development environment.

import com.google.cloud.aiplatform.v1.EndpointName;
import com.google.cloud.aiplatform.v1.PredictResponse;
import com.google.cloud.aiplatform.v1.PredictionServiceClient;
import com.google.cloud.aiplatform.v1.PredictionServiceSettings;
import com.google.protobuf.InvalidProtocolBufferException;
import com.google.protobuf.Value;
import com.google.protobuf.util.JsonFormat;
import java.io.IOException;
import java.util.ArrayList;
import java.util.List;

public class PredictCodeGenerationFunctionSample {

  public static void main(String[] args) throws IOException {
    // TODO(developer): Replace this variable before running the sample.
    String project = "YOUR_PROJECT_ID";

    // Learn how to create prompts to work with a code model to generate code:
    // https://cloud.google.com/vertex-ai/docs/generative-ai/code/code-generation-prompts
    String instance = "{ \"prefix\": \"Write a function that checks if a year is a leap year.\"}";
    String parameters = "{\n" + "  \"temperature\": 0.5,\n" + "  \"maxOutputTokens\": 256,\n" + "}";
    String location = "us-central1";
    String publisher = "google";
    String model = "code-bison@001";

    predictFunction(instance, parameters, project, location, publisher, model);
  }

  // Use Codey for Code Generation to generate a code function
  public static void predictFunction(
      String instance,
      String parameters,
      String project,
      String location,
      String publisher,
      String model)
      throws IOException {
    final String endpoint = String.format("%s-aiplatform.googleapis.com:443", location);
    PredictionServiceSettings predictionServiceSettings =
        PredictionServiceSettings.newBuilder().setEndpoint(endpoint).build();

    // Initialize client that will be used to send requests. This client only needs to be created
    // once, and can be reused for multiple requests.
    try (PredictionServiceClient predictionServiceClient =
        PredictionServiceClient.create(predictionServiceSettings)) {
      final EndpointName endpointName =
          EndpointName.ofProjectLocationPublisherModelName(project, location, publisher, model);

      Value instanceValue = stringToValue(instance);
      List<Value> instances = new ArrayList<>();
      instances.add(instanceValue);

      Value parameterValue = stringToValue(parameters);

      PredictResponse predictResponse =
          predictionServiceClient.predict(endpointName, instances, parameterValue);
      System.out.println("Predict Response");
      System.out.println(predictResponse);
    }
  }

  // Convert a Json string to a protobuf.Value
  static Value stringToValue(String value) throws InvalidProtocolBufferException {
    Value.Builder builder = Value.newBuilder();
    JsonFormat.parser().merge(value, builder);
    return builder.build();
  }
}

Console

To test a code generation prompt using Vertex AI Studio in the Google Cloud console, do following:

1. In the Vertex AI section of the Google Cloud console, go to Vertex AI Studio.

   Go to Vertex AI Studio

2. Click Get started.
3. Click Create prompt.
4. In Model, select the model with the name that begins with code-bison. A three digit number after code-bison indicates the version number of the model. For example, code-bison@002 is the name of version one of the code generation model.
5. In Prompt, enter a code generation prompt.
6. Adjust Temperature and Token limit to experiment with how they affect the response. For more information, see Code generation model parameters.
7. Click Submit to generate a response.
8. Click Save if you want to save a prompt
9. Click View code to see the Python code or a curl command for your prompt

Example curl command

MODEL_ID="code-bison"
PROJECT_ID=PROJECT_ID

curl \
-X POST \
-H "Authorization: Bearer $(gcloud auth print-access-token)" \
-H "Content-Type: application/json" \
https://us-central1-aiplatform.googleapis.com/v1/projects/${PROJECT_ID}/locations/us-central1/publishers/google/models/${MODEL_ID}:predict -d \
$"{
  'instances': [
    { 'prefix': 'Write a function that checks if a year is a leap year.' }
  ],
  'parameters': {
    'temperature': 0.2,
    'maxOutputTokens': 1024,
    'candidateCount': 1
  }
}"

To learn more about prompt design for code generation, see Create prompts for code generation.

Stream response from code model

To view sample code requests and responses using the REST API, see Examples using the streaming REST API.

To view sample code requests and responses using the Vertex AI SDK for Python, see Examples using Vertex AI SDK for Python for streaming.

What's next

• Learn how to create code chat prompts.
• Learn how to create code completion prompts.
• Learn about responsible AI best practices and Vertex AI's safety filters.