Examples / Specifications / Fuzzy Matching

Fuzzy Matching

Has SpecsRequires API Keys

Focused demonstration of fuzzy matching for AI output validation. This example shows: - High-threshold fuzzy matching (0.95) for near-exact matches - Medium-threshold fuzzy matching (0.45) for semantic similarity - Default threshold behavior when not specified - Using fuzzy matching to handle natural language variation

Source Code

World = Agent {
  model = "openai/gpt-4o-mini",
  system_prompt = "Your name is World."
}

Specification([[
Feature: Fuzzy Matching

  Scenario: High-threshold fuzzy match (nearly exact)
    Given the procedure has started
    And the message is "Hello, World!"
    And the agent "World" responds with "Hello! I'm World, nice to meet you!"
    When the procedure runs
    Then the output should fuzzy match "Hello! I'm World, nice to meet you!" with threshold 0.95

  Scenario: Medium-threshold fuzzy match (allows variation)
    Given the procedure has started
    And the message is "Hello, World!"
    And the agent "World" responds with "Hello! How can I assist you today?"
    When the procedure runs
    Then the output should fuzzy match "Hello! I'm World, nice to meet you!" with threshold 0.45

  Scenario: Default fuzzy match threshold
    Given the procedure has started
    And the message is "Hello, World!"
    And the agent "World" responds with "Hello! I'm World, nice to meet you!"
    When the procedure runs
    Then the output should fuzzy match "Hello! I'm World, nice to meet you!"
]])

local result = World({message = "Hello, World!"})
return (result and (result.output or result)) or ""

Quick Start

Run the example:

$tactus run 04-specifications/03-fuzzy-matching.tac

Test with mocks:

$tactus test 04-specifications/03-fuzzy-matching.tac --mock

Note

This example requires API keys. Set your OPENAI_API_KEY environment variable before running.

View source on GitHub →

Explore more examples

Learn Tactus through practical, runnable examples organized by topic.

View All Examples