Synonllm

Synonllm is a Ruby module that provides intelligent method calling through AI-powered synonym matching. When you call a method that doesn't exist, it uses Large Language Model analysis to suggest and automatically call semantically similar existing methods.

How It Works

When method_missing is triggered, Synonllm:

Captures the called method name, arguments, and available methods on the class
Analyzes the source code context using Gemini AI
Suggests the most likely intended method based on semantic similarity
Automatically calls the suggested method if confidence is high enough

Installation

Add this to your Ruby project:

require_relative "synonllm"

Usage

Include the Synonllm module in any class where you want intelligent method suggestion:

class MyClass
  include Synonllm
  
  def fetch_data
    "some data"
  end
end

obj = MyClass.new
obj.get_data  # Automatically calls fetch_data if AI determines it's a good match

Real Examples

Date Enhancement

require_relative "synonllm"

date = Date.new(1995, 12, 21)
Date.include(Synonllm)

# Calling day_of_week (doesn't exist) might suggest wday (does exist)
day_number = date.day_of_week  # AI suggests calling date.wday

File Handler

class FileHandler
  include Synonllm

  def open_the_file_at_the_given_path_and_return_a_file_object(filepath)
    File.open(filepath)
  end
end

handler = FileHandler.new
# Short method names are intelligently mapped to verbose ones
handler.get_file_at("./file.txt")  # AI suggests the long method name

Configuration

AI Provider

Synonllm uses Gemini 2.0 Flash by default. The AI provider is configured automatically when the module is used. You can change the AI provider by modifying the Sublayer configuration if needed or by building your own custom Sublayer provider.

Debug Mode

Enable debug output to see the AI's analysis:

DEBUG=1 ruby your_script.rb

This will show:

Detailed analysis of what the user was trying to do
The suggested method name
Confidence score (0-100)

How the AI Makes Decisions

The AI considers multiple factors when suggesting methods:

Semantic similarity: "remove" vs "delete", "get" vs "fetch"
Common abbreviations: "desc" vs "description", "auth" vs "authenticate"
Pluralization differences: "user" vs "users"
Argument compatibility: Whether the arguments match expected parameters
Ruby naming conventions: Standard patterns in Ruby method naming

Confidence Scoring

90-100: Almost certain match (very similar names, same functionality)
70-89: Likely match (similar purpose, compatible arguments)
50-69: Possible match (related functionality but some differences)
Below 50: Weak match or no suitable alternative

Dependencies

sublayer gem for LLM integration
Access to Gemini AI (configured automatically)

Examples

Run the included examples:

# Date example - shows method name suggestions for Date class
ruby date/enhanced_date.rb

# File handling example - shows verbose method name mapping
ruby file_handling/file_example.rb

# Enable debug output to see AI analysis
DEBUG=1 ruby date/enhanced_date.rb

How It Works Internally

Method Missing Hook: When a non-existent method is called, method_missing is triggered
Context Capture: The system captures:
- The called method name
- Arguments passed
- Available methods on the class
- Source code of the class
AI Analysis: A Sublayer-based generator analyzes the context using Gemini AI
Smart Suggestion: The AI provides a method suggestion with confidence score
Automatic Execution: If confidence is sufficient, the suggested method is called

License

See LICENSE file for details.

sublayerapp/synonllm_demo