# algoforge_prime/core/safe_executor.py # WARNING: THIS IS A SIMULATED SAFE EXECUTOR. DO NOT USE FOR REAL UNTRUSTED CODE. # A real implementation would require robust sandboxing (Docker, nsjail, Firecracker, WASM, etc.) # This simulation is purely for demonstrating the application flow. import time import random import traceback class ExecutionResult: def __init__(self, success: bool, output: str = None, error: str = None, execution_time: float = 0.0, passed_tests: int = 0, total_tests: int = 0): self.success = success # True if code ran without crashing (not necessarily if tests passed) self.output = output # Stdout from the execution self.error = error # Stderr or exception message self.execution_time = execution_time self.passed_tests = passed_tests # Number of 'assert' statements that passed self.total_tests = total_tests # Total 'assert' statements found and attempted def __str__(self): if self.success: test_summary = f"Tests: {self.passed_tests}/{self.total_tests} passed. " if self.total_tests > 0 else "" return f"Execution OK. {test_summary}Time: {self.execution_time:.4f}s. Output: {str(self.output)[:200]}" else: return f"Execution FAILED. Error: {str(self.error)[:300]}. Time: {self.execution_time:.4f}s" def execute_python_code_with_tests(code_string: str, tests_string: str, timeout_seconds: int = 5) -> ExecutionResult: """ SIMULATES sandboxed execution of Python code against a set of assert-based tests. In a real system, this function would be a complex interface to a secure sandbox. """ print(f"SIMULATOR: Attempting to 'execute' code with tests. Timeout: {timeout_seconds}s") start_time = time.time() if not code_string.strip(): return ExecutionResult(success=False, error="No code provided to execute.", execution_time=time.time() - start_time) # Basic simulation: # 1. Check for obvious syntax issues (very naive) if "def " not in code_string and "class " not in code_string and not any(op in code_string for op in ["=", "+", "-", "*", "/"]): # If it doesn't look like defining something or doing operations, maybe it's just a malformed snippet if len(code_string) < 50 and "print" not in code_string : # very arbitrary return ExecutionResult(success=False, error="Simulated: Code appears incomplete or malformed for execution.", execution_time=time.time()-start_time) # 2. Simulate test parsing and execution test_lines = [line.strip() for line in tests_string.splitlines() if line.strip().startswith("assert")] total_tests = len(test_lines) passed_tests = 0 simulated_stdout = [] simulated_stderr = None if total_tests == 0 and tests_string.strip(): # If tests were provided but no asserts found simulated_stderr = "Simulated: Test string provided, but no 'assert' statements found." # We can still "run" the code itself without asserts # Simulate code "running" try: if "loop forever" in code_string.lower() or "while True:" in code_string and "break" not in code_string: # Timeout simulation time.sleep(timeout_seconds + 0.1) # Exceed timeout raise TimeoutError("Simulated: Code execution timed out.") # Simulate print statements if "print('hello world')" in code_string: simulated_stdout.append("hello world") if "print(1+1)" in code_string: simulated_stdout.append("2") # Simulate errors if "1/0" in code_string or "zerodivisionerror" in code_string: raise ZeroDivisionError("Simulated: division by zero") if "undefined_variable" in code_string: raise NameError("Simulated: name 'undefined_variable' is not defined") # Simulate test passing/failing (very naively) if total_tests > 0: for i, test_line in enumerate(test_lines): # This is extremely basic and not real evaluation. # A real system would execute each assert in the context of the provided code. if "==" in test_line: if "True" in test_line or "120" in test_line or "correct" in test_line.lower(): # Naive pass conditions if random.random() > 0.1: # 90% chance of passing "good" tests passed_tests += 1 simulated_stdout.append(f"Simulated Test {i+1} ({test_line[:30]}...): PASSED") else: simulated_stdout.append(f"Simulated Test {i+1} ({test_line[:30]}...): FAILED (Random Sim)") if not simulated_stderr: simulated_stderr = f"Simulated: Assertion failed on test {i+1}" elif "False" in test_line or "wrong" in test_line.lower(): # Naive fail conditions if random.random() > 0.1: # 90% chance of failing "bad" tests simulated_stdout.append(f"Simulated Test {i+1} ({test_line[:30]}...): FAILED (Expected by Sim)") if not simulated_stderr: simulated_stderr = f"Simulated: Assertion failed on test {i+1} (Expected by Sim)" else: # 10% chance of unexpectedly passing passed_tests +=1 simulated_stdout.append(f"Simulated Test {i+1} ({test_line[:30]}...): PASSED (Unexpected by Sim)") else: # Other asserts if random.random() > 0.5: # 50/50 passed_tests += 1 simulated_stdout.append(f"Simulated Test {i+1} ({test_line[:30]}...): PASSED (Random Sim)") else: simulated_stdout.append(f"Simulated Test {i+1} ({test_line[:30]}...): FAILED (Random Sim)") if not simulated_stderr: simulated_stderr = f"Simulated: Assertion failed on test {i+1} (Random Sim)" else: # Non-equality asserts, just pass some randomly if random.random() > 0.3: passed_tests +=1 # If no explicit error, but not all tests passed if total_tests > 0 and passed_tests < total_tests and not simulated_stderr: simulated_stderr = f"Simulated: {total_tests - passed_tests} out of {total_tests} assertions failed." execution_time = time.time() - start_time if execution_time > timeout_seconds: # Check after simulated work raise TimeoutError("Simulated: Code execution exceeded timeout.") return ExecutionResult( success=True if not simulated_stderr or (total_tests > 0 and passed_tests == total_tests) else False, # Success if no errors or all tests pass output="\n".join(simulated_stdout) if simulated_stdout else None, error=simulated_stderr, execution_time=execution_time, passed_tests=passed_tests, total_tests=total_tests ) except Exception as e: execution_time = time.time() - start_time tb_str = traceback.format_exc() print(f"SIMULATOR EXCEPTION: {e}\n{tb_str}") return ExecutionResult( success=False, error=f"Simulated Exception: {type(e).__name__} - {str(e)}", execution_time=execution_time, passed_tests=passed_tests, # Could be some if error happened mid-tests total_tests=total_tests )