def rouge_filescores(hyp_path,ref_path):
'''
:param hyp_path: 实验结果文本
:param ref_path: 标准
:return: 实验结果文本相对于标准文本的分数
'''
files_rouge = rouge.FilesRouge(hyp_path, ref_path)
scores = files_rouge.get_scores()
return scores
def setUp(self):
self.hyp_path = './tests/hyp.txt'
self.ref_path = './tests/ref.txt'
self.data_path = './tests/data.json'
with open(self.data_path) as f:
self.data = json.load(f)
self.rouge = rouge.Rouge()
self.files_rouge = rouge.FilesRouge()
def rouge_filescores(hyp_path, ref_path):
"""
:param hyp_path: 实验结果文本
:param ref_path: 标准文本
:return: 实验结果文本相对于标准文本的分数
use the rouge matrix calculate
"""
files_rouge = rouge.FilesRouge(hyp_path, ref_path)
scores = files_rouge.get_scores()
return scores
def setUp(self):
self.hyp_path = 'predicted_sentences_task1_ref0_2.txt'
self.ref_path = 'task1_ref0_2.txt'
self.data_path = 'data2.json'
with open(self.data_path) as f:
self.data = json.load(f)
self.rouge = rouge.Rouge()
self.files_rouge = rouge.FilesRouge()
def main():
sys.setrecursionlimit(2000 * 500 + 1000)
for model in [
"ast_cleaned", "transformer_no_ast", "raw_python",
"unprocessed_ast"
]:
fr = rouge.FilesRouge("hyps/" + model + "_hyp.txt",
"refs/" + model + "_ref.test")
print(model)
print(json.dumps(fr.get_scores(avg=True), indent=2))
def run_red(reference_fn: str, candidate_fn: str) -> ROUGEResults:
r = rouge.FilesRouge(candidate_fn, reference_fn)
scores = r.get_scores(avg=True)
return {
'rouge1': {
'precision': scores['rouge-1']['p'],
'recall': scores['rouge-1']['r'],
'f1': scores['rouge-1']['f'],
},
'rouge2': {
'precision': scores['rouge-2']['p'],
'recall': scores['rouge-2']['r'],
'f1': scores['rouge-2']['f'],
},
'rougeL': {
'precision': scores['rouge-l']['p'],
'recall': scores['rouge-l']['r'],
'f1': scores['rouge-l']['f'],
},
}
os.makedirs(save_path)
pred_test_file_name = os.path.join(save_path,
f'test_{pred_file_name_suffix}.txt')
ref_test_file_name = os.path.join(save_path,
f'test_{ref_file_name_suffix}.txt')
if model_name == 'textrank':
summarizer = GensimSummarizer(model_params, file_to_process,
pred_test_file_name, ref_test_file_name,
preprocessor)
summarizer.summarize()
if model_name == 'first_sentence':
summarizer = ExtractFirstFullSentence(model_params, file_to_process,
pred_test_file_name,
ref_test_file_name, preprocessor)
summarizer.summarize()
else:
raise NotImplementedError(f'Model {model_name} not yet implemented')
r = rouge.FilesRouge(pred_test_file_name, ref_test_file_name)
scores = r.get_scores(avg=True)
print(scores)
with open(os.path.join(save_path, 'scores.json'), 'w') as f:
json.dump(scores, f, indent=2)
with open(os.path.join(save_path, 'config.json'), 'w') as f:
json.dump(config, f, indent=2)
# references = ['the quick brown fox jumped over the lazy dog', 'test']
# hypotheses = ['the fast brown fox jumped over the lazy dog', 'test']
references = [
'Man on ocean beach flying several kites on windy day',
'Two people standing next to a river holding surfboards', 'test'
]
hypotheses = [
'a person on a beach flying a kite',
'a man holding a surfboard on a beach', 'test'
]
rouge_eval = rouge.Rouge()
test_r_eval = rouge.FilesRouge("data/example/hyp.txt", "data/example/ref.txt")
b = test_r_eval.get_scores()
print(b)
scores_rouge_dict_list = rouge_eval.get_scores(hypotheses, references)
print(scores_rouge_dict_list)
rouge_1_scores = [0.0] * 3
rouge_2_scores = [0.0] * 3
rouge_l_scores = [0.0] * 3
for i, s in enumerate(scores_rouge_dict_list):
if isinstance(s, dict):
for metric, vals in s.items():
print(metric)
print(vals)
if metric == 'rouge-1':
def setUp(self):
self.hyp_path = './hyp.txt'
self.ref_path = './ref.txt'
self.rouge = rouge.Rouge()
self.files_rouge = rouge.FilesRouge(self.hyp_path, self.ref_path)