Add EchoAgent and Evaluation Pipeline with Custom Relevancy Scorer

[praneetha0909]

This PR includes:

• EchoAgent (my_agents/agents/echo_agent.py): returns "You said: "
• prompts.json (my_agents/examples/prompts.json): test inputs and expected outputs
• agent_battle.py (my_agents/agent_battle.py): runs the agent, evaluates results
• Custom scorer (answer_relevancy_scorer.py) in src/judgeval/scorers/
• Readme documentation (readme_agent.md)

This demonstrates how to benchmark a custom agent using the judgeval framework.

[gemini-code-assist]

Summary of Changes

Hello @praneetha0909, I'm Gemini Code Assist¹! I'm currently reviewing this pull request and will post my feedback shortly. In the meantime, here's a summary to help you and other reviewers quickly get up to speed!

This pull request establishes a complete demonstration project for evaluating custom AI agents using the judgeval framework. It introduces a simple 'EchoAgent,' defines a set of test prompts, implements a custom relevancy scoring mechanism, and provides a runnable script to execute the entire evaluation pipeline, showcasing the framework's capabilities for benchmarking agent performance.

Highlights

• New Agent Implementation: Introduced my_agents/agents/echo_agent.py, a minimal agent that simply echoes the input prompt with a 'You said: ' prefix. This serves as a basic example for demonstrating agent evaluation.
• Evaluation Pipeline Script: Added my_agents/agent_battle.py, which orchestrates the evaluation process. This script loads test prompts, dynamically loads and runs the EchoAgent, and then uses a custom scorer to evaluate the agent's generated responses against expected outputs.
• Custom Relevancy Scorer: Implemented src/judgeval/scorers/answer_relevancy_scorer.py, a custom AnswerRelevancyScorer. This scorer calculates the similarity between the agent's generated output and the expected output using difflib.SequenceMatcher, providing a custom metric for evaluation.
• Example Prompts and Documentation: Included my_agents/examples/prompts.json with sample input/expected output pairs for testing. Comprehensive documentation is provided in Readme_Agent and readme_agent.md, detailing the project structure, how to run the evaluation, and example output.

Using Gemini Code Assist

The full guide for Gemini Code Assist can be found on our documentation page, here are some quick tips.

Invoking Gemini

You can request assistance from Gemini at any point in your pull request via creating an issue comment (i.e. comment on the pull request page) using either /gemini <command> or @gemini-code-assist <command>. Below is a summary of the supported commands.

Feature	Command	Description
Code Review	/gemini review	Performs a code review for the current pull request in its current state.
Pull Request Summary	/gemini summary	Provides a summary of the current pull request in its current state.
Comment	@gemini-code-assist	Responds in comments when explicitly tagged, both in issue comments and review comments.
Help	/gemini help	Displays a list of available commands.

Customization

To customize Gemini Code Assist for GitHub experience, repository maintainers can create a configuration file and/or provide a custom code review style guide (such as PEP-8 for Python) by creating and adding files to a .gemini/ folder in the base of the repository. Detailed instructions can be found here.

Limitations & Feedback

Gemini Code Assist is currently in preview and may make mistakes. Please leave feedback on any instances where its feedback is incorrect or counter productive. You can react with 👍 and 👎 on @gemini-code-assist comments to provide feedback.

You can also get AI-powered code generation, chat, as well as code reviews directly in the IDE at no cost with the Gemini Code Assist IDE Extension.

Footnotes

1. Review the Privacy Notices, Generative AI Prohibited Use Policy, Terms of Service, and learn how to configure Gemini Code Assist in GitHub here. Gemini can make mistakes, so double check it and use code with caution. ↩

[gemini-code-assist]

Code Review

The pull request introduces an EchoAgent and an evaluation pipeline with a custom relevancy scorer. The changes include the agent implementation, example prompts, an evaluation script, and a custom scorer. The review focuses on improving code clarity and maintainability, particularly regarding dynamic attribute assignment and the use of asynchronous methods.

[praneetha0909]

Thanks @gemini-code-assist for the helpful suggestions. I've updated the code to store actual_output in additional_metadata and removed the unused async method to reduce confusion. Let me know if anything else needs to be addressed!

[gemini-code-assist]

Thanks for the update, @praneetha0909! I've reviewed the changes:

• The update to store actual_output in example.additional_metadata in my_agents/agent_battle.py is a good improvement for managing custom data.
• Commenting out the unused a_score_example method in src/judgeval/scorers/answer_relevancy_scorer.py effectively removes the confusion as intended.

These changes look good and address the points you mentioned. Based on the current state of the pull request, I don't see any further issues that need to be addressed from my side. Feel free to merge when you're ready!

[JCamyre]

Thanks for raising this PR! We don't include agents in this repo, will be closing this PR. Thanks for using Judgeval to evaluate your agent, feel free to leave any feedback you have.