网站制作客户寻找,数据中台厂商,推广引流怎么做,可以合成装备的传奇手游目录 问题描述#xff1a;
问题解决#xff1a; 问题描述#xff1a;
原始的标注的三元组格式如下#xff1a; 需要转换的格式如下#xff1a; tips:有一个小的难点#xff1a;
1. 针对多三元组的情况#xff0c;需要额外考虑
2. 最后一个样本#xff0c;也记得需要…目录 问题描述
问题解决 问题描述
原始的标注的三元组格式如下 需要转换的格式如下 tips:有一个小的难点
1. 针对多三元组的情况需要额外考虑
2. 最后一个样本也记得需要处理 问题解决
from pdb import set_trace as stop
import os
from tqdm import trange
generated_path /public/home/hongy/qtxu/UniCOQE_20230812/data/tuple/ele/train_new_generated.txt
Unicoqe_path /public/home/hongy/qtxu/UniCOQE_20230812/data/tuple/ele/train.txtraw_data []
with open(os.path.join(generated_path), r) as f:for line in f:raw_data.append(line)with open(Unicoqe_path, w) as fw:line_id, i 0, 0text_line, label_line , for line_id in trange(len(raw_data), desc procesing data ……):cur_line raw_data[line_id]if len(cur_line.split(\t)) ! 2:label_line \n cur_lineelse:if text_line !:sent, label text_line.strip().split(\t)fw.write(sent####)label_list label_line.strip().split(\n\n)span_index []all_span ()for label_i in label_list:cur_span label_i.strip()[1:-1].split(;)sub, obj,asp cur_span[0], cur_span[1], cur_span[2]try:sub_index [int(index) for index, word in (pair.split(, 1) if in pair else [pair, ] for pair in sub.strip()[1:-1].split())] # 针对14这样的特例进行处理obj_index [int(index) for index, word in (pair.split(, 1) if in pair else [pair, ] for pair in obj.strip()[1:-1].split())]asp_index [int(index) for index, word in (pair.split()for pair in asp.strip()[1:-1].split())] except:print(text_line)stop()span_tuple (sub_index, obj_index, asp_index)span_index.append(span_tuple)fw.write(str([, .join(str(span) for span in span_index)) ]\n)text_line cur_linelabel_linefw.write(text_line.strip().split(\t)[0]####)label_list label_line.strip().split(\n\n)span_index []all_span ()for label_i in label_list:cur_span label_i.strip()[1:-1].split(;)sub, obj,asp cur_span[0], cur_span[1], cur_span[2]sub_index [int(index) for index, word in (pair.split(, 1) if in pair else [pair, ] for pair in sub.strip()[1:-1].split())] # 针对14这样的特例进行处理obj_index [int(index) for index, word in (pair.split(, 1) if in pair else [pair, ] for pair in obj.strip()[1:-1].split())]asp_index [int(index) for index, word in (pair.split()for pair in asp.strip()[1:-1].split())] span_tuple (sub_index, obj_index, asp_index)span_index.append(span_tuple)fw.write(str([, .join(str(span) for span in span_index)) ]\n)