)((故|因[而此]|所以)\s*[：:]?\s*答案分?别?([为是]|填)?\s*[：:]\s*(.+?))(\n|$)', parse) # # print("ans_s:",ans_s.group(5)) # # print(ans_s.group(0)) # if ans_s.group(5) and ans_s.group(5).count("【ⅳ】") > 1: # ans_summarize = ans_s.group(2) # ans_s_index = parse.index(ans_summarize) if ans_s.group(1) == '
' or not ans_s.group(1).strip() \ # or ans_s.group(1).strip() is None else parse.index(ans_s.group(1)) # # ans_summarize = [ans_s.group(2), ans_s_index] # parse = parse.replace(ans_summarize, "") # # elif ans_s.group(5) and "【ⅳ】" in ans_s.group(5): 小问答案里也可以有“故答案为” # # aa5 = ans_s.group(5).replace("【ⅳ】", "") # # parse = parse.replace(ans_s.group(5), aa5) # ----------------------------------------------- parse = re.sub("(【ⅳ】\s*解答?\s[:：])\s*【ⅳ】", r"\1", parse) little_parse = re.split(r"【ⅳ】", parse) if len(little_parse) > 1: # if re.search("\n\s*(【分析】|分析\s*[：:]).+?", "\n"+little_parse[0].strip()): # analy_comment.append(little_parse[0]) if len(little_parse[0].strip()) >= 5: analy_comment.append(little_parse[0]) parse_list = little_parse[1:] # print(analy_comment) # ---------------------------拆分后组合---------------------------------- # 解析不拆分时，小问也要拆; # 小问解析个数与小问一致时才组合的 one_item = split2little_con(con_list, ans_list, parse_list, one_item) if parse_split and "slave" in one_item: # 解析需拆分 # 里层答案/解析存在时，外层就不需要了 if one_item["slave"][0]["parse"].strip(): one_item['parse'] = "\n".join(analy_comment).strip() if one_item["slave"][0]["key"].strip(): one_item['key'] = ans_summarize.strip() else: # 不存在多问时 if re.findall(r"_{2,}", one_item["stem"]): one_item["blank_num"] = len(re.findall(r"_{2,}", one_item["stem"])) one_item["answer_type"] = "填空题" # if parse_split and re.search("^[A-Z]{2,}$", re.sub("\W", "", ans)): # one_item["type"] = "多选题" elif len(re.findall(r"[\n\s\u4e00-\u9fa5]\s*[A-D]\s*[.．、､]", one_item["stem"])) >= 3: # 增加对选项的拆分处理 one_item = option_structure(one_item, con, ans, 1) one_item["answer_type"] = "选择题" if 'options' not in one_item: one_item["options"] = [] elif re.search("[(（]\s+[）)]\s*$", one_item["stem"]) or one_item["type"] == "判断题": one_item["answer_type"] = "判断题" if re.match("【?(对的?|正确的?|[T√])】?$", one_item["key"].strip()): one_item["key"] = "正确" elif re.match("【?(错误?的?|不对的?|不正确的?|[F×])】?$", one_item["key"].strip()): one_item["key"] = "错误" elif re.search("[横划画]线处填写", one_item["stem"]) and "com_stem" in one_item: one_item["answer_type"] = "填空题" if "com_stem" in one_item: blank_num = len(re.findall(r"_{2,}", one_item["com_stem"])) if blank_num > 0: one_item["blank_num"] = blank_num else: one_item["answer_type"] = "解答题" else: one_item["answer_type"] = "解答题" return one_item def split2little_con(con_list, ans_list, parse_list, one_item): """ 将按小问切分开的题干、答案、解析 进行 【结构化组合】 :param con_list:切开了小问的题干 :param ans_list:切开了小问的答案 :param parse_list:切开了小问的解析 :param one_item: 初步切开的一道题目 :param is_sub_item: 答案是否按小题号获取（还是按照空的个数获取）的标志 :param ans_summarize: 解析中的综述 [内容，索引] :return: """ # print(con_list) # print(ans_list) # print(parse_list) # print('***********************') if len(con_list) > 1: if con_list[0] == "": # 说明全是小题，没有总题文 one_item["stem"] = "" else: one_item["stem"] = con_list[0] slave = [] for index, s in enumerate(con_list[1:]): # 以题干拆分为主 blank_num = len(re.findall(r"_{2,}", s)) s = re.sub(r"[(（]\d+分[)）]", "", s[:9]) + s[9:] one_slave = {"slave_no": "（%s）" % (index + 1), "stem": s, "key": "", "parse": "", "answer_type": "解答题", "errmsgs": [], } if len(con_list) - len(parse_list) == 1: one_slave["parse"] = parse_list[index] # 按索引取解析 if isinstance(ans_list, list) and len(con_list) - len(ans_list) == 1: one_slave["key"] = ans_list[index] # 判断小题干是否可以是选择题（含选择题时，作答类型也不一定是选择题） if len(re.findall(r"[\n\s\u4e00-\u9fa5]\s*[A-D]\s*[.．、､]", s)) >= 3: raw_ans = one_slave["key"] raw_stem = one_slave["stem"] one_slave = option_structure(one_slave, s, one_slave["key"], 1, is_slave=1) one_slave["answer_type"] = "选择题" if "options" not in one_slave or not one_slave["options"]: one_slave["key"] = raw_ans # 选择题解析不成功时，答案还原 elif blank_num > 1: one_slave["answer_type"] = "填空题" one_slave["key"] = raw_ans one_slave["stem"] = raw_stem del one_slave["options"], one_slave["options_rank"] if "options" not in one_slave or not one_slave["options"]: if blank_num > 0: one_slave["blank_num"] = blank_num one_slave["answer_type"] = "填空题" elif re.search("[(（]\s+[）)]\s*$", s): one_slave["answer_type"] = "判断题" elif re.search("[横划画]线处填写", s): one_slave["answer_type"] = "填空题" blank_num = len(re.findall(r"_{2,}", one_item["stem"])) if blank_num > 0: one_slave["blank_num"] = blank_num else: one_slave["answer_type"] = "解答题" # if "errmsgs" in one_slave: # del one_slave["errmsgs"] # 对带小题的大题，对每个小题的答案重新再提取一次 # if one_slave["parse"].strip() and (not ans_list or "key" not in one_slave or not one_slave["key"]): # or one_slave["key"] == '见解析' # new_ans = only_parse_split(one_slave["parse"], one_item["type"], one_slave["stem"], # reparse_n=2) # 再解析 # if new_ans["key"] == "见解析": # new_ans["key"] = "" # if new_ans["key"] and not new_ans["parse"]: # pass # else: # one_slave["key"] = new_ans["key"] # one_slave["parse"] = new_ans["parse"] slave.append(one_slave) one_item["slave"] = slave one_item["slave_no"] = "1-{}".format(len(slave)) if len(slave) > 1 else "1" # 有slave的话,就把外面的ans和parse字段给删除掉 # if slave: # pass # # del one_item["key"] # # if parse_list: # # del one_item["parse"] # else: # one_item["stem"] = old_con # if ans_summarize: # one_item["parse"] = one_item["parse"][:ans_summarize[1]] + '\n' + ans_summarize[0] + '\n' + one_item["parse"][ans_summarize[1]:] return one_item