new_tiku_structure_v3_sci/structure/final_structure.py

# 各题型结构化
import re

from structure.option import option_structure, new_options_rank, table_option_struc
from structure.dati2slave import get_slave


def one_item_structure(xyz):
    """
    判断解析类型，解析类型为：
    if:
        1.stem不需要再做其他处理<-- 答案没有[;；],且答案不是ABCDEFG
        2.选择题类,需要把stem中的ABCD各选项内容提取出来<--答案是ABCDEFG
    else:
        都要看是否含有小题，如果含有小题，需要把小题提取出来，slave
        3.填空题类，（1）需要提取stem中下划线的个数

    选择题结构化：单选或者多选<--要把各选项是什么提取出来放在slave中
    one_item:{"stem":xxxx,"key":xxx,"parse":xxx}
    consumer: 分“高中数学”还是“全学科”；
    item_no_type：题号是否以（\d）的形式
    :return:
    """
    one_item, consumer, item_no_type = xyz
    if "【章节】" in one_item["parse"]:  # 属于后一个题的，后面须调整
        one_item["chapter"] = one_item["parse"].split("【章节】")[1].split("\n")[0]
        one_item["parse"] = one_item["parse"].replace("【章节】" + one_item["chapter"], "")
    if "【章节】" in one_item["stem"]:  # 属于后一个题的，后面须调整
        one_item["chapter"] = one_item["stem"].split("【章节】")[1].split("\n")[0]
        one_item["stem"] = one_item["stem"].replace("【章节】" + one_item["chapter"], "")
    if "【选做题】" in one_item["stem"] + one_item["key"] + one_item["parse"]:
        opt_str = re.search(r"【选做题】:'(\d+)分'", one_item["stem"] + one_item["key"] + one_item["parse"])
        one_item["option_st"] = "选做题,"+opt_str.group(1) if opt_str else  "选做题"  # 选做题开始的位置，后面的题开始是选做题
        one_item["stem"] = re.sub("【选做题】(:'(\d+)分')?", "", one_item["stem"])
        one_item["key"] = re.sub("【选做题】(:'(\d+)分')?", "", one_item["key"])
        one_item["parse"] = re.sub("【选做题】(:'(\d+)分')?", "", one_item["parse"])

    ans = one_item["key"]
    con = one_item["stem"]
    parse = re.sub(r"((?<=[\n】])|^)\s*解\s*[：:]", "", one_item["parse"])
    if not one_item["type"]:
        # one_item["errmsgs"].append("本题没有给出明确题型！")
        # return one_item
        if re.match(r"[A-Z][A-Z;；和与、､\s]*?$", ans.strip()):
            one_item["type"] = "单选题" if len(ans.strip()) == 1 else "多选题"
        elif re.search(r"[(（]\s*[)）]", one_item["stem"]) or \
                len(re.findall(r"[\n\s]\s*[A-D]\s*[.．、､]", one_item["stem"])) >= 4:
            one_item["type"] = "选择题"
        elif re.findall(r"_{2,}", one_item["stem"]):
            one_item["type"] = "填空题"
        else:
            one_item["type"] = "简答题"
    elif one_item["type"].replace("题", "") in ["单选", "多选", "选择", "不定选择"]:
        # # 存在试卷所给题型不规范或本身就是错的情况
        if len(re.findall("\n\s*[(（]\s*[1-9]\s*[）)]", "\n" + con)) > 1 \
                or (re.search(r"[(（]\s*[)）][\s\n]*(<img src=|A\s*[.．、､])", con) is None
                    and len(re.findall(r"[\n\s]\s*[A-D]\s*[.．、､]", con)) < 2):
            if re.findall(r"_{2,}", con):
                one_item["type"] = "填空题"
            else:
                one_item["type"] = "解答题"

    topic_type = one_item["type"]
    # print(topic_type)

    if topic_type.replace("题", "") in ["单选", "多选", "选择", "不定选择"]:
        one_item = option_structure(one_item, con, ans, item_no_type)
        one_item["stem"] = re.sub("(?<![/\"]>)(<br\s*/?>|\n)+\s*(?!<img)", "", one_item["stem"])
        if 'options' not in one_item:
            one_item["options"] = []
            # 表格类型的选项再解析,
            if "<table>" in one_item["stem"]:
                may_options = table_option_struc(one_item["stem"])
                if may_options:
                    one_item["options"] = may_options
                    one_item["options_rank"] = 2
                    one_item["errmsgs"] = [emg for emg in one_item["errmsgs"] if "选项格式不正确" not in emg]
        else:  # 选择题结构化成功时，对选项排列方式再换思路算
            options_rank_2 = new_options_rank(one_item["options"])
            if options_rank_2:
                one_item["options_rank"] = options_rank_2
        one_item["blank_num"] = len(re.findall(r"_{2,}", one_item["stem"]))
        one_item["answer_type"] = "选择题"
    elif consumer == 'toslave':  # 拆小题
        one_item = get_slave(one_item, con, parse, ans)
        # if ('slave' not in one_item or not one_item['slave']) and 'analy' in one_item:
        #     del one_item['analy']
        # if one_item["type"] == "多选题":
        #     one_item = option_structure(one_item, con, ans, item_no_type)
        #     if 'options' not in one_item:
        #         one_item["options"] = []
    else:  # 不拆小题，非选择题
        if "#" in ans:
            one_item["key"] = one_item["key"].replace("#", "; ")
        pattern1 = re.compile(r"([是为点]|等于|=|＝|有|存在)\s*_+((<img src=((?!/>).)*?/>|[^_;；。?!\n])+?)_+"
                              r"([cdkm上]?m?\s*.?[。.?]?\s*($|<br/>|<img src|……))")
        pattern2 = re.compile(r"((有|存在|[是为])[\u4e00-\u9fa5]{0,2})\s*_+(\d+)_+\s*([\u4e00-\u9fa5，,;；。.．])")

        if re.findall(r"_{2,}", one_item["stem"]):  # re.search("_+([^_]*?)_+", one_item['stem']):
            one_item["blank_num"] = len(re.findall(r"_{2,}", one_item["stem"]))
        else:
            # 是否只需将所有标点符号去除即可，这里容易判断错误！！！！
            if re.search("^[A-Z]{2,}$",
                         re.sub(r"[^\w＞＜≤≥≡≦≧＋－≈≠﹢﹣±㏒㏑∑∏π><=＝×÷／()（）﹙﹚\[\]﹛﹜{\}∧∨∠▰▱△∆⊙⌒"
                                r"⊆⊂⊇⊃∈∩∉∪⊕∥∣≌∽∞∝⊥∫∬∮∯Φ∅≮≯∁∴∵∷←↑→↓↖↗↘↙‖〒¤○′″￠°℃℉"
                                r"αβγδεζηθικλμνξορστυφχψωϕ％‰℅㎎㎏㎜㎝㎞㎡㎥㏄㏎㏕＄￡￥º¹²³⁴ⁿ₁₂₃₄·∶½⅓⅔¼¾⅛⅜⅝⅞"
                                r"ΑΒΓΔΕΖΗΘΙΚΜ]", "", ans)):
                one_item["type"] = "多选题"
                one_item = option_structure(one_item, con, ans, item_no_type)
                if 'options' not in one_item:
                    one_item["options"] = []
        if one_item["type"] == "填空题" and re.search("_{2,}|填正确答案", one_item['stem']) is None:
            # -----放在huanhang_wash_after中调整--------------
            # blank_ans =[]
            # while re.search(pattern1, one_item["stem"]):  # 答案直接填在____上的情况
            #     blank_con1 = re.search(pattern1, one_item["stem"])
            #     one_item["stem"] = one_item["stem"].replace(blank_con1.group(0),
            #                                           blank_con1.group(1) + "____" + blank_con1.group(5))
            #     blank_ans.append(blank_con1.group(2))
            # while re.search(pattern2, one_item["stem"]):  # 答案直接填在____上的情况
            #     blank_con1 = re.search(pattern2, one_item["stem"])
            #     one_item["stem"] = one_item["stem"].replace(blank_con1.group(0),
            #                                           blank_con1.group(1) + "____" + blank_con1.group(3))
            #     blank_ans.append(blank_con1.group(2))
            # if not ans:
            #         one_item["key"] = ";".join(blank_ans)
            # one_item["blank_num"] = len(blank_ans)
            # ----------------------------------------------
            if re.match(r"[A-Z][A-Z;；和与、､\s]*?$", ans.strip()):
                one_item["type"] = "单选题" if len(ans.strip()) == 1 else "多选题"
                one_item = option_structure(one_item, con, ans, item_no_type)
                if 'options' not in one_item:
                    one_item["options"] = []
            elif re.search(r"[(（]\s*[)）]", one_item["stem"]) or ('步骤' not in one_item["stem"] and
                    len(re.findall(r"[\n\s]\s*[A-D]\s*[.．、､]", one_item["stem"])) >= 4):
                one_item["type"] = "选择题"
                one_item = option_structure(one_item, con, ans, item_no_type)
                if 'options' not in one_item:
                    one_item["options"] = []
            elif re.findall('(有|存在|[是为＝=])[　\s]{3,}[a-zA-Z]', one_item["stem"]):
                one_item["blank_num"] = len(re.findall('(有|存在|[是为＝=])[　\s]{3,}[a-zA-Z]', one_item["stem"]))
            elif re.findall('[　\s]{3,}[a-zA-Z]\s*[,;.。；，]', one_item["stem"]):
                one_item["blank_num"] = len(re.findall('\s{3,}\n*\s*[a-zA-Z]\s*[,;.。；，．]', one_item["stem"]))
            elif re.search(pattern1, one_item["stem"]) is None and re.search(pattern2, one_item["stem"]) is None:
                stem = re.sub("<img src=.*?/>|[,，.。．、､]", "", one_item["stem"])
                if len(stem) > 2:
                    one_item["type"] = "解答题"

    # print('------------------------------------------------')
    if one_item:
        # if re.match(r"(\[.*?\])?\s*\(.*?(\d+)分\)", one_item["stem"].strip()):
        #     # print(one_item["stem"])
        #     score_info = re.match(r"(\[.*?\])?\s*\(.*?(\d+)分\)", one_item["stem"].strip())
        #     one_item["score"] = float(score_info.group(2))
        one_item["stem"] = re.sub(r"(\[.*?\])?\(.*?\d+分\)", "", one_item["stem"][:20]) + one_item["stem"][20:]
    return one_item