# # Copyright (c) 2018, Salesforce, Inc. # All rights reserved. # # Redistribution and use in source and binary forms, with or without # modification, are permitted provided that the following conditions are met: # # * Redistributions of source code must retain the above copyright notice, this # list of conditions and the following disclaimer. # # * Redistributions in binary form must reproduce the above copyright notice, # this list of conditions and the following disclaimer in the documentation # and/or other materials provided with the distribution. # # * Neither the name of the copyright holder nor the names of its # contributors may be used to endorse or promote products derived from # this software without specific prior written permission. # # THIS SOFTWARE IS PROVIDED BY THE COPYRIGHT HOLDERS AND CONTRIBUTORS "AS IS" # AND ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE # IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE ARE # DISCLAIMED. IN NO EVENT SHALL THE COPYRIGHT HOLDER OR CONTRIBUTORS BE LIABLE # FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL # DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS OR # SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION) HOWEVER # CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT LIABILITY, # OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY OUT OF THE USE # OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE. from subprocess import Popen, PIPE import logging import re import os import string import numpy as np import collections from multiprocessing import Pool, cpu_count from contextlib import closing from .utils.lang_utils import * from .tasks.generic_dataset import Query from pyrouge import Rouge155 from sacrebleu import corpus_bleu def to_lf(s, table): aggs = [y.lower() for y in Query.agg_ops] agg_to_idx = {x: i for i, x in enumerate(aggs)} conditionals = [y.lower() for y in Query.cond_ops] headers_unsorted = [(y.lower(), i) for i, y in enumerate(table['header'])] headers = [(y.lower(), i) for i, y in enumerate(table['header'])] headers.sort(reverse=True, key=lambda x: len(x[0])) condition_s, conds = None, [] if 'where' in s: s, condition_s = s.split('where', 1) s = ' '.join(s.split()[1:-2]) sel, agg = None, 0 for col, idx in headers: if col == s: sel = idx if sel is None: s = s.split() agg = agg_to_idx[s[0]] s = ' '.join(s[1:]) for col, idx in headers: if col == s: sel = idx full_conditions = [] if not condition_s is None: condition_s = ' ' + condition_s + ' ' for idx, col in enumerate(headers): condition_s = condition_s.replace(' ' + col[0] + ' ', ' Col{} '.format(col[1])) condition_s = condition_s.strip() for idx, col in enumerate(conditionals): new_s = [] for t in condition_s.split(): if t == col: new_s.append('Cond{}'.format(idx)) else: new_s.append(t) condition_s = ' '.join(new_s) s = condition_s conds = re.split('(Col\d+ Cond\d+)', s) if len(conds) == 0: conds = [s] conds = [x for x in conds if len(x.strip()) > 0] full_conditions = [] for i, x in enumerate(conds): if i % 2 == 0: x = x.split() col_num = int(x[0].replace('Col', '')) opp_num = int(x[1].replace('Cond', '')) full_conditions.append([col_num, opp_num]) else: x = x.split() if x[-1] == 'and': x = x[:-1] x = ' '.join(x) if 'Col' in x: new_x = [] for t in x.split(): if 'Col' in t: idx = int(t.replace('Col', '')) t = headers_unsorted[idx][0] new_x.append(t) x = new_x x = ' '.join(x) if 'Cond' in x: new_x = [] for t in x.split(): if 'Cond' in t: idx = int(t.replace('Cond', '')) t = conditionals[idx] new_x.append(t) x = new_x x = ' '.join(x) full_conditions[-1].append(x) logical_form = {'sel': sel, 'conds': full_conditions, 'agg': agg} return logical_form def computeLFEM(greedy, answer, args): answer = [x[0] for x in answer] count = 0 correct = 0 text_answers = [] for idx, (g, ex) in enumerate(zip(greedy, answer)): count += 1 text_answers.append([ex['answer'].lower()]) try: lf = to_lf(g, ex['table']) gt = ex['sql'] conds = gt['conds'] lower_conds = [] for c in conds: lc = c lc[2] = str(lc[2]).lower() lower_conds.append(lc) gt['conds'] = lower_conds correct += lf == gt except Exception as e: continue return correct / count * 100, text_answers def score(answer, gold): if len(gold) > 0: gold = set.union(*[simplify(g) for g in gold]) answer = simplify(answer) tp, tn, sys_pos, real_pos = 0, 0, 0, 0 if answer == gold: if not ('unanswerable' in gold and len(gold) == 1): tp += 1 else: tn += 1 if not ('unanswerable' in answer and len(answer) == 1): sys_pos += 1 if not ('unanswerable' in gold and len(gold) == 1): real_pos += 1 return np.array([tp, tn, sys_pos, real_pos]) def simplify(answer): return set(''.join(c for c in t if c not in string.punctuation) for t in answer.strip().lower().split()) - {'the', 'a', 'an', 'and', ''} # http://nlp.cs.washington.edu/zeroshot/evaluate.py def computeCF1(greedy, answer): scores = np.zeros(4) for g, a in zip(greedy, answer): scores += score(g, a) tp, tn, sys_pos, real_pos = scores.tolist() total = len(answer) if tp == 0: p = r = f = 0.0 else: p = tp / float(sys_pos) r = tp / float(real_pos) f = 2 * p * r / (p + r) return f * 100, p * 100, r * 100 def normalize_text(s): """Lower text and remove punctuation, articles and extra whitespace.""" def remove_articles(text): return re.sub(r'\b(a|an|the)\b', ' ', text) def white_space_fix(text): return ' '.join(text.split()) def remove_punc(text): exclude = set(string.punctuation) return ''.join(ch for ch in text if ch not in exclude) def lower(text): return text.lower() return white_space_fix(remove_articles(remove_punc(lower(s)))) def f1_score(prediction, ground_truth): prediction_tokens = prediction.split() ground_truth_tokens = ground_truth.split() common = collections.Counter(prediction_tokens) & collections.Counter(ground_truth_tokens) num_same = sum(common.values()) if num_same == 0: return 0 precision = 1.0 * num_same / len(prediction_tokens) recall = 1.0 * num_same / len(ground_truth_tokens) f1 = (2 * precision * recall) / (precision + recall) return f1 def fm_score(prediction, ground_truth): pred_funcs = get_functions(prediction) ground_truth = get_functions(ground_truth) common = collections.Counter(pred_funcs) & collections.Counter(ground_truth) if not len(ground_truth): return 1.0 return len(common) / len(ground_truth) def dm_score(prediction, ground_truth): pred_funcs = get_devices(prediction) ground_truth = get_devices(ground_truth) common = collections.Counter(pred_funcs) & collections.Counter(ground_truth) if not len(ground_truth): return 1.0 return len(common) / len(ground_truth) def exact_match(prediction, ground_truth): return prediction == ground_truth def metric_max_over_ground_truths(metric_fn, prediction, ground_truths): scores_for_ground_truths = [] for idx, ground_truth in enumerate(ground_truths): score = metric_fn(prediction, ground_truth) scores_for_ground_truths.append(score) return max(scores_for_ground_truths) def computeF1(outputs, targets): return sum([metric_max_over_ground_truths(f1_score, o, t) for o, t in zip(outputs, targets)])/len(outputs) * 100 def computeEM(outputs, targets): outs = [metric_max_over_ground_truths(exact_match, o, t) for o, t in zip(outputs, targets)] return sum(outs)/len(outputs) * 100 def computeFM(outputs, targets): outs = [metric_max_over_ground_truths(fm_score, o, t) for o, t in zip(outputs, targets)] return sum(outs) / len(outputs) * 100 def computeDM(outputs, targets): outs = [metric_max_over_ground_truths(dm_score, o, t) for o, t in zip(outputs, targets)] return sum(outs) / len(outputs) * 100 def computeBLEU(outputs, targets): targets = [[t[i] for t in targets] for i in range(len(targets[0]))] return corpus_bleu(outputs, targets, lowercase=True).score class Rouge(Rouge155): """Rouge calculator class with custom command-line options.""" # See full list of options here: # https://github.com/andersjo/pyrouge/blob/master/tools/ROUGE-1.5.5/README.txt#L82 DEFAULT_OPTIONS = [ '-a', # evaluate all systems '-n', 4, # max-ngram '-x', # do not calculate ROUGE-L '-2', 4, # max-gap-length '-u', # include unigram in skip-bigram '-c', 95, # confidence interval '-r', 1000, # number-of-samples (for resampling) '-f', 'A', # scoring formula '-p', 0.5, # 0 <= alpha <=1 '-t', 0, # count by token instead of sentence '-d', # print per evaluation scores ] def __init__(self, n_words=None, keep_files=False, options=None): if options is None: self.options = self.DEFAULT_OPTIONS.copy() else: self.options = options if n_words: options.extend(["-l", n_words]) stem = "-m" in self.options super(Rouge, self).__init__( n_words=n_words, stem=stem, keep_files=keep_files) def _run_rouge(self): # Get full options options = ( ['-e', self._rouge_data] + list(map(str, self.options)) + [os.path.join(self._config_dir, "settings.xml")]) logging.info("Running ROUGE with options {}".format(" ".join(options))) # print([self._rouge_bin] + list(options)) pipes = Popen([self._rouge_bin] + options, stdout=PIPE, stderr=PIPE) std_out, std_err = pipes.communicate() div_by_zero_error = std_err.decode("utf-8").\ startswith("Illegal division by zero") if pipes.returncode == 0 or div_by_zero_error: # Still returns the correct output even with div by zero return std_out else: raise ValueError( std_out.decode("utf-8") + "\n" + std_err.decode("utf-8")) def computeROUGE(greedy, answer): rouges = compute_rouge_scores(greedy, answer) if len(rouges) > 0: avg_rouges = {} for key in rouges[0].keys(): avg_rouges[key] = sum( [r.get(key, 0.0) for r in rouges]) / len(rouges) * 100 else: avg_rouges = None return avg_rouges def split_sentences(txt, splitchar=".", include_splitchar=False): """Split sentences of a text based on a given EOS char.""" out = [s.split() for s in txt.strip().split(splitchar) if len(s) > 0] return out def compute_rouge_scores(summs, refs, splitchar='.', options=None, parallel=True): assert len(summs) == len(refs) options = [ '-a', # evaluate all systems '-c', 95, # confidence interval '-m', # use Porter stemmer '-n', 2, # max-ngram '-w', 1.3, # weight (weighting factor for WLCS) ] rr = Rouge(options=options) rouge_args = [] for summ, ref in zip(summs, refs): letter = "A" ref_dict = {} for r in ref: ref_dict[letter] = [x for x in split_sentences(r, splitchar) if len(x) > 0] letter = chr(ord(letter) + 1) s = [x for x in split_sentences(summ, splitchar) if len(x) > 0] rouge_args.append((s, ref_dict)) if parallel: with closing(Pool(cpu_count()//2)) as pool: rouge_scores = pool.starmap(rr.score_summary, rouge_args) else: rouge_scores = [] for s, a in rouge_args: rouge_scores.append(rr.score_summary(s, ref_dict)) return rouge_scores def to_delta_state(line): delta_state = {'inform': {}, 'request': {}} try: if line == 'None' or line.strip() == '' or line.strip() == ';': return delta_state inform, request = [[y.strip() for y in x.strip().split(',')] for x in line.split(';')] inform_pairs = {} for i in inform: try: k, v = i.split(':') inform_pairs[k.strip()] = v.strip() except: pass delta_state = {'inform': inform_pairs, 'request': request} except: pass finally: return delta_state def update_state(state, delta): for act, slot in delta.items(): state[act] = slot return state def dict_cmp(d1, d2): def cmp(a, b): for k1, v1 in a.items(): if k1 not in b: return False else: if v1 != b[k1]: return False return True return cmp(d1, d2) and cmp(d2, d1) def computeDialogue(greedy, answer): examples = [] for idx, (g, a) in enumerate(zip(greedy, answer)): examples.append((a[0][0], g, a[0][1], idx)) examples.sort() turn_request_positives = 0 turn_goal_positives = 0 joint_goal_positives = 0 ldt = None for ex in examples: if ldt is None or ldt.split('_')[:-1] != ex[0].split('_')[:-1]: state, answer_state = {}, {} ldt = ex[0] delta_state = to_delta_state(ex[1]) answer_delta_state = to_delta_state(ex[2]) state = update_state(state, delta_state['inform']) answer_state = update_state(answer_state, answer_delta_state['inform']) if dict_cmp(state, answer_state): joint_goal_positives += 1 if delta_state['request'] == answer_delta_state['request']: turn_request_positives += 1 if dict_cmp(delta_state['inform'], answer_delta_state['inform']): turn_goal_positives += 1 joint_goal_em = joint_goal_positives / len(examples) * 100 turn_request_em = turn_request_positives / len(examples) * 100 turn_goal_em = turn_goal_positives / len(examples) * 100 answer = [(x[-1], x[-2]) for x in examples] answer.sort() answer = [[x[1]] for x in answer] return joint_goal_em, turn_request_em, turn_goal_em, answer def compute_metrics(greedy, answer, requested_metrics, args=None): metric_keys = [] metric_values = [] if not isinstance(answer[0], list): answer = [[a] for a in answer] if 'lfem' in requested_metrics: lfem, answer = computeLFEM(greedy, answer, args) metric_keys += ['lfem'] metric_values += [lfem] if 'joint_goal_em' in requested_metrics: joint_goal_em, request_em, turn_goal_em, answer = computeDialogue(greedy, answer) avg_dialogue = (joint_goal_em + request_em) / 2 metric_keys += ['joint_goal_em', 'turn_request_em', 'turn_goal_em', 'avg_dialogue'] metric_values += [joint_goal_em, request_em, turn_goal_em, avg_dialogue] em = computeEM(greedy, answer) metric_keys += ['em'] metric_values += [em] if 'bleu' in requested_metrics: bleu = computeBLEU(greedy, answer) metric_keys.append('bleu') metric_values.append(bleu) if 'avg_rouge' in requested_metrics: rouge = computeROUGE(greedy, answer) metric_keys += ['rouge1', 'rouge2', 'rougeL', 'avg_rouge'] avg_rouge = (rouge['rouge_1_f_score'] + rouge['rouge_2_f_score'] + rouge['rouge_l_f_score']) / 3 metric_values += [rouge['rouge_1_f_score'], rouge['rouge_2_f_score'], rouge['rouge_l_f_score'], avg_rouge] norm_greedy = [normalize_text(g) for g in greedy] norm_answer = [[normalize_text(a) for a in al] for al in answer] nf1 = computeF1(norm_greedy, norm_answer) nem = computeEM(norm_greedy, norm_answer) metric_keys.extend(['nf1', 'nem']) metric_values.extend([nf1, nem]) if 'fm' in requested_metrics: function_accuracy = computeFM(greedy, answer) metric_keys.append('fm') metric_values.append(function_accuracy) if 'dm' in requested_metrics: device_accuracy = computeDM(greedy, answer) metric_keys.append('dm') metric_values.append(device_accuracy) if 'corpus_f1' in requested_metrics: corpus_f1, precision, recall = computeCF1(norm_greedy, norm_answer) metric_keys += ['corpus_f1', 'precision', 'recall'] metric_values += [corpus_f1, precision, recall] metric_dict = dict(zip(metric_keys, metric_values)) metric_dict = collections.OrderedDict((key, metric_dict[key]) for key in requested_metrics) return metric_dict, answer