import itertools import numpy as np from typing import Dict from datasets import load_dataset import testing_util as test_util DATASET = "codeparrot/apps" def evaluate_generations(generations, level=["all"]): """We take the list of code generations and try to compile them and the run their corresponding unit tests which are retrieved from the APPS dataset. Args: generations: list of code generations, in the same order as APPS dataset samples level: list of levels to evaluate, can be "all", "introductory", "interview" or "competition" Returns: results: dictionary of results, key is the problem index, value is a list of results for each generation [-2] = compile error, [-1] = runtime error [False] = failed test case [True] = passed test case """ # generations are code generations in the same order of the dataset apps_eval = load_dataset(DATASET, split="test", difficulties=level) results = {} for index in range(len(generations)): print(f"task {index}") generated_code = generations[index] sample = apps_eval[index] res = [] # loop over the generations for o_idx, o in enumerate(generated_code): curr_res = [-2] try: print("Run test") curr_res = test_util.run_test(sample, test=o, debug=False) print("\nSuccessful compilation!") fixed = [] for e in curr_res: if isinstance(e, np.ndarray): e = e.item(0) if isinstance(e, np.bool_): e = bool(e) fixed.append(e) curr_res = fixed if not np.all(curr_res): print(f"Results were not True for all test cases") #{curr_res}") except Exception as e: print(f"Compilation failed, test framework exception = {repr(e)}{e}\n") break finally: assert isinstance(curr_res, list) res.append(curr_res) results[index] = res return results def estimate_pass_at_k(num_samples, num_correct, k): """Estimates pass@k of each problem and returns them in an array.""" def estimator(n: int, c: int, k: int) -> float: """Calculates 1 - comb(n - c, k) / comb(n, k).""" if n - c < k: return 1.0 return 1.0 - np.prod(1.0 - k / np.arange(n - c + 1, n + 1)) if isinstance(num_samples, int): num_samples_it = itertools.repeat(num_samples, len(num_correct)) else: assert len(num_samples) == len(num_correct) num_samples_it = iter(num_samples) return np.array([estimator(int(n), int(c), k) for n, c in zip(num_samples_it, num_correct)]) def get_results(results: Dict, count_errors: bool = False, k_list: list = [1, 10, 100]): """ Given the results evaluated against the testcases we output some statistics. For single generations: >>> example_results = {"0": [[-2]],"1": [[False,False]],"2": [[True,True]],"3": [[False,True,False,True]], "4": [[-1,-1]]} >>> get_results(example_results, count_errors=True) number of compile errors = 1 avg = 0.2 number of runtime errors = 1 avg = 0.2 number of test cases run = 5 Test Case Average (average accuracy over problems) = 0.3 Strict Accuracy (all test cases passed / total problems) = 0.2 For multiple generations: >>> example_results = {"0": [[-2], [True, True, True]],"1": [[-1,-1, -1], [True, False, True]]} >>> get_results(example_results k_list=[1, 2]) {'pass@1': 0.25, 'pass@2': 0.5} """ metrics = {"avg_accuracy": None, "strict_accuracy": None, "pass_at_k": None} if len(results["0"]) == 1: # for single generations we compute average accuracy and stric accuracy: original APPS metrics print("Computing accuracy metrics...") res = [] per_prob_res = [] all_correct = [] for index in results: results[index] = np.array(results[index]) res.extend(results[index]) per_prob_res.append(np.mean(results[index]>0)) all_correct.append(np.all(results[index]>0)) # we count campilation and runtime errors once per pronlem compile_errors = len([e for e in res if -2 in e]) runtime_errors = len([e for e in res if -1 in e]) total_testcases = len(res) if count_errors: print(f"number of compile errors = {compile_errors} avg = {compile_errors / total_testcases}") print(f"number of runtime errors = {runtime_errors} avg = {runtime_errors / total_testcases}") print(f"number of problems evaluated = {total_testcases}") print(f"Test Case Average Accuracy (ver tests) = {np.mean(per_prob_res)}") print(f"Strict Accuracy (over problems that pass all tests) = {np.mean(all_correct)}") metrics["avg_accuracy"] = np.mean(per_prob_res) metrics["strict_accuracy"] = np.mean(all_correct) else: # for multiple generations we use pass@k metric used in the HumanEval benchmark # we use strict accuracy, a generation is valid if it has to pass all the tests print("Computing pass@k metric for multiple generations...") # total is list with nb generations per task (task=index) # correct is number of generations that passed all tests per task total = [] correct = [] for index in results: all_correct = [] for generation in results[index]: gen = np.array(generation) all_correct.append(np.all(gen>0)) total.append(len(all_correct)) correct.append(sum(all_correct)) total = np.array(total) correct = np.array(correct) ks = k_list pass_at_k = {f"pass@{k}": estimate_pass_at_k(total, correct, k).mean() for k in ks if (total >= k).all()} print(pass_at_k) metrics["pass_at_k"] = pass_at_k return metrics def compute_metrics(generations, k_list=[1, 10, 100], count_errors=True, level=["all"]): """Return metrics for the given generations. Args: generations: dict of generations, keyed by problem index k_list: list of k values to compute pass@k when using multiple generations count_errors: whether to count compilation and runtime errors when using single generations level: which level difficulty in APPS dataset was used for the given generations Returns: metrics: dict of metrics """ results = evaluate_generations(generations, level=level) metrics = get_results(results, count_errors=count_errors, k_list=k_list) return metrics