RecruitmentAssistant-A2A Evaluation Scripts

This directory contains evaluation scripts for analyzing the RecruitmentAssistant-A2A (Agent-to-Agent) project results.

Scripts

1. analyze_retry_patterns.py

Analyzes RETRY patterns in execution paths.

Features:

Counts error occurrences by location and type
Calculates RETRY rates per model
Identifies maximum RETRY attempts
Tracks errors by agent and node type

Output:

retry_analysis.json - Detailed JSON results
retry_summary.csv - Summary statistics
error_by_agent.csv - Error breakdown by agent

Usage:

python3 analyze_retry_patterns.py

2. evaluate_success.py

Calculates success rates for each model.

Success Criteria:

Reports folder exists
Contains at least 2 markdown files

Output:

success_detailed_results.json - Detailed results per session
success_rate.csv - Success rate summary

Usage:

python3 evaluate_success.py

3. evaluate_trajectory.py

Evaluates trajectory metrics against reference trajectories.

Metrics:

Exact match
In-order match
Any-order match
Precision
Recall
Single-tool use
Path diversity (unique path ratio, path entropy)

Output:

evaluation_results.csv - Trajectory evaluation results

Usage:

python3 evaluate_trajectory.py --config reference_trajectory.yaml

4. reference_trajectory.yaml

Configuration file defining reference trajectories and evaluation parameters.

Key Sections:

reference_trajectory - Ground truth execution path
target_tools - Required tools for evaluation
models - List of models to evaluate
extract_types - Node types to extract from execution paths

Configuration

Edit reference_trajectory.yaml to:

Modify reference trajectories
Add/remove models
Adjust tool evaluation criteria
Configure extraction types

Requirements

Python 3.7+
pandas
pyyaml

Install dependencies:

pip install pandas pyyaml

Data Structure

Expected directory structure:

/Users/wzr/TOSEM-2025/RESULTS/
├── ModelName/
│   └── RecruitmentAssistant-A2A/
│       └── test_results/
│           ├── session_1/
│           │   ├── execution_path.md
│           │   └── reports/
│           │       ├── file1.md
│           │       └── file2.md
│           └── session_2/
│               └── ...

Notes

All scripts output CSV files for easy analysis
JSON files contain detailed per-session information
Modify BASE_DIR in scripts if using a different results directory
The A2A architecture uses standard agent-to-agent communication patterns