Spaces:
Sleeping
Sleeping
File size: 6,467 Bytes
af25c62 |
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65 66 67 68 69 70 71 72 73 74 75 76 77 78 79 80 81 82 83 84 85 86 87 88 89 90 91 92 93 94 95 96 97 98 99 100 101 102 103 104 105 106 107 108 109 110 111 112 113 114 115 116 117 118 119 120 121 122 123 124 125 126 127 128 129 130 131 132 133 134 135 136 137 138 139 140 141 142 143 144 145 146 147 148 149 150 151 152 153 154 155 156 157 158 159 160 161 162 163 164 165 166 167 168 169 170 171 172 173 174 175 176 177 178 179 180 181 182 183 184 185 186 187 188 189 190 191 192 193 194 195 196 197 |
"""
Quick Test Script for RGB RAG Evaluation
Tests a small sample to verify everything works before full evaluation.
"""
import os
import sys
# Add parent directory to path
sys.path.insert(0, os.path.dirname(os.path.abspath(__file__)))
from src.llm_client import GroqLLMClient
from src.data_loader import RGBDataLoader
from src.evaluator import RGBEvaluator
from src.prompts import format_prompt, get_prompt_template
def quick_test(num_samples: int = 3):
"""
Run a quick test with a few samples.
Args:
num_samples: Number of samples to test per task.
"""
print("="*60)
print("RGB RAG Evaluation - Quick Test")
print("="*60)
# Check for API key
api_key = os.getenv("GROQ_API_KEY")
if not api_key:
print("\n❌ ERROR: GROQ_API_KEY environment variable not set!")
print("\nTo set it:")
print(" 1. Get your free API key from: https://console.groq.com/")
print(" 2. Create a .env file with: GROQ_API_KEY=your_key_here")
print(" 3. Or set environment variable: $env:GROQ_API_KEY='your_key_here'")
return False
print(f"\n✓ API Key found: {api_key[:8]}...")
# Initialize components
print("\nInitializing components...")
try:
client = GroqLLMClient(model="llama-3.1-8b-instant") # Use fast model for testing
loader = RGBDataLoader()
evaluator = RGBEvaluator()
print(f"✓ Using model: {client.model}")
except Exception as e:
print(f"❌ Error initializing: {e}")
return False
# Test 1: Simple generation
print("\n" + "-"*60)
print("Test 1: Simple Generation")
print("-"*60)
try:
response = client.generate("What is 2 + 2? Answer with just the number.")
print(f" Prompt: 'What is 2 + 2?'")
print(f" Response: '{response}'")
print(" ✓ LLM connection working!")
except Exception as e:
print(f" ❌ Error: {e}")
return False
# Test 2: Noise Robustness sample
print("\n" + "-"*60)
print("Test 2: Noise Robustness Sample")
print("-"*60)
try:
samples = loader.load_noise_robustness(max_samples=num_samples)
sample = samples[0]
prompt = format_prompt(
question=sample.question,
documents=sample.documents,
template=get_prompt_template("default")
)
print(f" Question: {sample.question[:60]}...")
print(f" Expected: {sample.answer[:60]}...")
response = client.generate(prompt)
print(f" Response: {response[:100]}...")
is_correct = evaluator.is_correct(response, sample.answer)
print(f" Correct: {is_correct}")
print(" ✓ Noise robustness test complete!")
except Exception as e:
print(f" ❌ Error: {e}")
# Test 3: Negative Rejection sample
print("\n" + "-"*60)
print("Test 3: Negative Rejection Sample")
print("-"*60)
try:
samples = loader.load_negative_rejection(max_samples=num_samples)
sample = samples[0]
prompt = format_prompt(
question=sample.question,
documents=sample.documents,
template=get_prompt_template("negative")
)
print(f" Question: {sample.question[:60]}...")
response = client.generate(prompt)
print(f" Response: {response[:100]}...")
is_rejection = evaluator.is_rejection(response)
print(f" Rejected: {is_rejection} (should reject since docs don't have answer)")
print(" ✓ Negative rejection test complete!")
except Exception as e:
print(f" ❌ Error: {e}")
# Test 4: Information Integration sample
print("\n" + "-"*60)
print("Test 4: Information Integration Sample")
print("-"*60)
try:
samples = loader.load_information_integration(max_samples=num_samples)
sample = samples[0]
prompt = format_prompt(
question=sample.question,
documents=sample.documents,
template=get_prompt_template("default")
)
print(f" Question: {sample.question[:60]}...")
print(f" Expected: {sample.answer[:60]}...")
response = client.generate(prompt)
print(f" Response: {response[:100]}...")
is_correct = evaluator.is_correct(response, sample.answer)
print(f" Correct: {is_correct}")
print(" ✓ Information integration test complete!")
except Exception as e:
print(f" ❌ Error: {e}")
# Test 5: Counterfactual Robustness sample
print("\n" + "-"*60)
print("Test 5: Counterfactual Robustness Sample")
print("-"*60)
try:
samples = loader.load_counterfactual_robustness(max_samples=num_samples)
sample = samples[0]
prompt = format_prompt(
question=sample.question,
documents=sample.documents,
template=get_prompt_template("counterfactual")
)
print(f" Question: {sample.question[:60]}...")
print(f" Correct Answer: {sample.answer}")
print(f" Fake Answer (in docs): {sample.counterfactual_answer}")
response = client.generate(prompt)
print(f" Response: {response[:100]}...")
detects_error = evaluator.detects_error(response, sample.counterfactual_answer)
corrects_error = evaluator.corrects_error(response, sample.answer, sample.counterfactual_answer)
print(f" Detects Error: {detects_error}")
print(f" Corrects Error: {corrects_error}")
print(" ✓ Counterfactual robustness test complete!")
except Exception as e:
print(f" ❌ Error: {e}")
print("\n" + "="*60)
print("Quick Test Complete!")
print("="*60)
print("\nTo run full evaluation:")
print(" python run_evaluation.py --max-samples 50")
print("\nOr for a complete run:")
print(" python run_evaluation.py")
return True
if __name__ == "__main__":
import argparse
parser = argparse.ArgumentParser(description="Quick test for RGB evaluation")
parser.add_argument("-n", "--num-samples", type=int, default=3,
help="Number of samples to test per task")
args = parser.parse_args()
quick_test(args.num_samples)
|