Package com.tnsai.prompt.optimizer

Class PromptOptimizer

java.lang.Object
com.tnsai.prompt.optimizer.PromptOptimizer
public class PromptOptimizer extends Object
Automated prompt optimization framework.

PromptOptimizer automatically tunes prompts for better performance through iterative refinement, evaluation, and strategy selection.

Features

  • Automatic strategy selection based on task type
  • Iterative prompt refinement
  • A/B testing for prompt variants
  • Performance metrics tracking

Usage


 PromptOptimizer optimizer = PromptOptimizer.builder()
     .llmClient(llmClient)
     .maxIterations(5)
     .targetScore(0.9f)
     .build();

 OptimizationResult result = optimizer.optimize(
     "Summarize the following article",
     testCases
 );

 System.out.println("Best prompt: " + result.getBestPrompt());
 System.out.println("Score: " + result.getScore());

Reference

Based on Salesforce AI Research Promptomatix pattern.

  • Method Details

    • optimize

      public OptimizationResult optimize(String basePrompt, List<TestCase> testCases)
      Optimizes a prompt using test cases.
      Parameters:
      basePrompt - the initial prompt to optimize
      testCases - test cases for evaluation
      Returns:
      optimization result with best prompt and metrics

    • suggestStrategies

      public List<com.tnsai.prompt.strategy.PromptStrategy> suggestStrategies(String taskDescription)
      Suggests strategies based on task analysis.
      Parameters:
      taskDescription - description of the task
      Returns:
      recommended strategies

    • builder

      public static PromptOptimizer.Builder builder()