Spaces:
Running
Running
File size: 17,161 Bytes
c2ea5ed 7bc750c c2ea5ed 7bc750c |
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 99 100 101 102 103 104 105 106 107 108 109 110 111 112 113 114 115 116 117 118 119 120 121 122 123 124 125 126 127 128 129 130 131 132 133 134 135 136 137 138 139 140 141 142 143 144 145 146 147 148 149 150 151 152 153 154 155 156 157 158 159 160 161 162 163 164 165 166 167 168 169 170 171 172 173 174 175 176 177 178 179 180 181 182 183 184 185 186 187 188 189 190 191 192 193 194 195 196 197 198 199 200 201 202 203 204 205 206 207 208 209 210 211 212 213 214 215 216 217 218 219 220 221 222 223 224 225 226 227 228 229 230 231 232 233 234 235 236 237 238 239 240 241 242 243 244 245 246 247 248 249 250 251 252 253 254 255 256 257 258 259 260 261 262 263 264 265 266 267 268 269 270 271 272 273 274 275 276 277 278 279 280 281 282 283 284 285 286 287 288 289 290 291 292 293 294 295 296 297 298 299 300 301 302 303 304 305 306 307 308 309 310 311 312 313 314 315 316 317 318 319 320 321 322 323 324 325 326 327 328 329 330 331 332 333 334 335 336 337 338 339 340 341 342 343 344 345 346 347 348 349 350 351 352 353 354 355 356 357 358 359 360 361 362 363 364 365 366 367 368 369 370 371 372 373 374 375 376 377 378 379 380 381 382 383 384 385 386 387 |
"""
Service for test-related operations
"""
import os
import json
import time
import logging
import traceback
import threading
from typing import Dict, List, Any, Optional
from pathlib import Path
from backend.server_config import TEST_RESULTS_FILE, PROJECT_ROOT
logger = logging.getLogger("agent_monitoring_server.services.test")
class TestService:
"""Service for test-related operations"""
@staticmethod
def test_relation(tester, relation_id: str, model: str, system_prompt: Optional[str] = None) -> Dict[str, Any]:
"""Test a specific relation"""
try:
result = tester.test_relation(relation_id, model, system_prompt)
return result
except Exception as e:
logger.error(f"Error testing relation: {str(e)}")
logger.error(traceback.format_exc())
raise
@staticmethod
def test_relation_with_jailbreak(
tester,
relation_id: str,
model: str,
jailbreak_index: int,
system_prompt: Optional[str] = None
) -> Dict[str, Any]:
"""Test a relation with a jailbreak attack"""
try:
result = tester.test_relation_with_jailbreak(relation_id, model, jailbreak_index, system_prompt)
return result
except Exception as e:
logger.error(f"Error testing relation with jailbreak: {str(e)}")
logger.error(traceback.format_exc())
raise
@staticmethod
def test_relations_by_type(
tester,
relation_type: str,
model: str,
system_prompt: Optional[str] = None
) -> Dict[str, Any]:
"""Test all relations of a specific type"""
try:
# Get relations of the specified type
relations = tester.list_relations(relation_type)
if not relations:
return {
"message": "No relations found of the specified type",
"results": {}
}
# Get relation IDs
relation_ids = [r["id"] for r in relations]
# Test the relations and save results
results = tester.test_multiple_relations(relation_ids, model, system_prompt)
tester.save_results(TEST_RESULTS_FILE, results)
return {
"message": f"Tested {len(relation_ids)} relations of type {relation_type}",
"relation_count": len(relation_ids),
"results": results
}
except Exception as e:
logger.error(f"Error testing relations by type: {str(e)}")
logger.error(traceback.format_exc())
raise
@staticmethod
def run_perturbation_test(
tester,
knowledge_graph: str,
model: str = "gpt-5-mini",
perturbation_type: str = "jailbreak",
relation_type: str = "",
max_jailbreaks: int = 5,
relation_limit: int = 5
) -> Dict[str, Any]:
"""Run perturbation tests on a knowledge graph (using database)"""
try:
# Get a database session
from backend.database.utils import get_db, get_knowledge_graph
session = next(get_db())
try:
# Get the knowledge graph from database
kg = get_knowledge_graph(session, knowledge_graph)
if not kg:
raise FileNotFoundError(f"Knowledge graph '{knowledge_graph}' not found in database")
# Get the knowledge graph content
knowledge_graph_content = kg.graph_data
logger.info(f"Retrieved knowledge graph {knowledge_graph} from database for testing")
# Generate timestamp for this test run
timestamp = int(time.time())
output_path = f"perturbation_results_{timestamp}.json"
# Limit jailbreak techniques if specified
if perturbation_type == 'jailbreak' and max_jailbreaks > 0:
if len(tester.jailbreak_techniques) > max_jailbreaks:
logger.info(f"Limiting jailbreak techniques to {max_jailbreaks}")
tester.jailbreak_techniques = tester.jailbreak_techniques[:max_jailbreaks]
# Start the test in a separate thread
def run_tests():
try:
# Run the tests
tester.run_tests(
knowledge_graph=knowledge_graph_content,
output_file=output_path,
model=model,
perturbation_type=perturbation_type,
relation_type=relation_type,
relation_limit=relation_limit,
max_jailbreaks=max_jailbreaks
)
# Add metadata to the results file
try:
with open(output_path, 'r') as f:
results = json.load(f)
# Add test metadata
results["test_metadata"] = {
"timestamp": timestamp,
"knowledge_graph_id": kg.id,
"knowledge_graph_file": knowledge_graph,
"model": model,
"perturbation_type": perturbation_type,
"relation_type": relation_type,
"max_jailbreaks": max_jailbreaks,
"relation_limit": relation_limit
}
# Write back the updated results
with open(output_path, 'w') as f:
json.dump(results, f, indent=2)
logger.info(f"Added metadata to results file {output_path}")
except Exception as e:
logger.error(f"Error adding metadata to results: {str(e)}")
logger.info(f"Perturbation tests completed. Results saved to {output_path}")
except Exception as e:
logger.error(f"Error running perturbation tests: {str(e)}")
logger.error(traceback.format_exc())
# Create an error results file
error_results = {
"error": True,
"error_message": str(e),
"timestamp": timestamp,
"test_metadata": {
"timestamp": timestamp,
"knowledge_graph_id": kg.id,
"knowledge_graph_file": knowledge_graph,
"model": model,
"perturbation_type": perturbation_type,
"relation_type": relation_type,
"max_jailbreaks": max_jailbreaks,
"relation_limit": relation_limit,
"status": "failed"
}
}
with open(output_path, 'w') as f:
json.dump(error_results, f, indent=2)
# Start the thread
threading.Thread(target=run_tests).start()
# Return response with test information
return {
"status": "in_progress",
"message": f"Perturbation testing started. Results will be saved to {output_path}",
"output_file": output_path,
"timestamp": timestamp,
"knowledge_graph_id": kg.id
}
finally:
session.close()
except Exception as e:
logger.error(f"Error starting perturbation test: {str(e)}")
logger.error(traceback.format_exc())
raise
@staticmethod
def get_test_results() -> Dict[str, Any]:
"""Get test results from the most recent test"""
try:
if not os.path.exists(TEST_RESULTS_FILE):
# Create a simple "no results" response
return {
"status": "no_results",
"message": "No test results are currently available. Run a perturbation test first."
}
try:
with open(TEST_RESULTS_FILE, 'r') as f:
results = json.load(f)
# Add download URL for the results file if it doesn't have an error
if not results.get("error", False):
# If there's an output file stored in the results, add a download link
if "output_file" in results:
output_file = results["output_file"]
results["download_url"] = f"/download/{output_file}"
logger.info(f"Added download URL for {output_file}")
# If there's no output file but we can guess it from timestamp
elif "timestamp" in results:
timestamp = int(results.get("timestamp", time.time()))
guessed_file = f"perturbation_results_{timestamp}.json"
if os.path.exists(guessed_file):
results["output_file"] = guessed_file
results["download_url"] = f"/download/{guessed_file}"
logger.info(f"Added download URL for guessed file {guessed_file}")
return results
except json.JSONDecodeError as e:
# The test results file exists but is not valid JSON
logger.error(f"Invalid JSON in test results file: {str(e)}")
return {
"error": True,
"error_message": f"Test results file contains invalid JSON: {str(e)}",
"timestamp": time.time()
}
except Exception as e:
logger.error(f"Error retrieving test results: {str(e)}")
raise
@staticmethod
def get_test_history() -> Dict[str, List[Dict[str, Any]]]:
"""Get history of all tests"""
try:
# Get all perturbation result files
test_dir = 'datasets/test_results'
Path(test_dir).mkdir(parents=True, exist_ok=True)
result_files = [f for f in os.listdir(test_dir)
if f.startswith('perturbation_results_') and f.endswith('.json')]
# Collect metadata from each file
tests = []
for file in result_files:
try:
with open(os.path.join(test_dir, file), 'r') as f:
data = json.load(f)
# Add output file name to the data
data['output_file'] = file
tests.append(data)
except Exception as e:
logger.error(f"Error reading test result file {file}: {str(e)}")
return {"tests": tests}
except Exception as e:
logger.error(f"Error getting test history: {str(e)}")
return {"tests": [], "error": str(e)}
@staticmethod
def get_specific_test_result(test_id: Optional[str] = None, test_file: Optional[str] = None) -> Dict[str, Any]:
"""Get a specific test result by ID or filename"""
try:
# If file is specified, load directly
if test_file:
file_path = str(PROJECT_ROOT / 'datasets' / 'test_results' / test_file)
if not os.path.exists(file_path):
raise FileNotFoundError(f"Test result file {test_file} not found")
with open(file_path, 'r') as f:
result = json.load(f)
# Add download URL
result['download_url'] = f"/download/{test_file}"
return result
# Otherwise find by ID
elif test_id:
test_dir = 'datasets/test_results'
result_files = [f for f in os.listdir(test_dir)
if f.startswith('perturbation_results_') and f.endswith('.json')]
for file in result_files:
try:
with open(os.path.join(test_dir, file), 'r') as f:
data = json.load(f)
# Check if this is the test we're looking for
if str(data.get('timestamp', '')) == test_id or data.get('id', '') == test_id:
result = data
result['download_url'] = f"/download/{file}"
result['output_file'] = file
return result
except Exception as e:
logger.error(f"Error reading test result file {file}: {str(e)}")
raise FileNotFoundError(f"Test result with ID {test_id} not found")
else:
raise ValueError("Either test_id or test_file must be provided")
except Exception as e:
logger.error(f"Error getting specific test result: {str(e)}")
raise
@staticmethod
def get_progress_status(output_file: str) -> Dict[str, Any]:
"""Get progress status for a test"""
try:
# Construct the progress file path
progress_file = f"progress_{output_file}"
test_dir = 'datasets/test_results'
# Make sure the test_results directory exists
Path(test_dir).mkdir(parents=True, exist_ok=True)
progress_path = os.path.join(test_dir, progress_file)
if not os.path.exists(progress_path):
# Create an empty progress file as a placeholder
try:
with open(progress_path, 'w') as f:
json.dump({
"status": "initializing",
"overall_progress_percentage": 0,
"current_jailbreak": "Preparing...",
"last_tested_relation": "Preparing...",
"created_at": time.time()
}, f)
logger.info(f"Created placeholder progress file: {progress_path}")
except Exception as e:
logger.error(f"Failed to create progress file: {str(e)}")
raise
# Read the progress file
with open(progress_path, 'r') as f:
progress_data = json.load(f)
# Add metadata about the file itself
progress_data["progress_file"] = progress_file
progress_data["last_updated"] = os.path.getmtime(progress_path)
return progress_data
except Exception as e:
logger.error(f"Error getting progress status: {str(e)}")
logger.error(traceback.format_exc())
raise
@staticmethod
def check_progress_file(output_file: str) -> Dict[str, Any]:
"""Check if a progress file exists without returning 404"""
try:
# Construct the progress file path
progress_file = f"progress_{output_file}"
test_dir = 'datasets/test_results'
# Make sure the test_results directory exists
Path(test_dir).mkdir(parents=True, exist_ok=True)
progress_path = os.path.join(test_dir, progress_file)
exists = os.path.exists(progress_path)
return {
"exists": exists,
"file": progress_file,
"path": progress_path,
"last_modified": os.path.getmtime(progress_path) if exists else None,
"status": "ready_to_create" if not exists else "exists"
}
except Exception as e:
logger.error(f"Error checking progress file: {str(e)}")
logger.error(traceback.format_exc())
return {
"exists": False,
"error": str(e)
}
|