new_tiku_structure_v3_art/structure/stems_to_groups.py

#!/usr/bin/env/python
# -*- coding:utf-8 -*-

import re
from pprint import pprint


def regroup(res_list, item_groups, ans_groups):
    """
    将多个题共用一个题干的情况进行slave重组,如地理
    item_groups中的groups_data,key值表示带公共题干的试题位置,从0开始计;
    value值:'fei'表示本题不是小题多问;
            '\d-\d'表示哪几个题合成slave；
            ''空表示带公共题干试题开始位置，没有slave范围
    例：item_groups: {'is_groups': 1, 'groups_data':
                    {0: 'fei', 5: '', 8: '', 11: 'fei', 12: '', 15: '', 19: '20-21', 23: 'fei'}}
    :param res_list:
    :param item_groups:
    :param ans_groups:
    :return:
    """
    new_res_dict = []
    groups_data = item_groups["groups_data"]
    start_no = list(groups_data.keys())  # 与真实题号差1
    start_no.sort()  # 排序

    def takefirst(elem):
        return int(elem.split("-")[0])

    ans_start_no = []
    if ans_groups:
        ans_start_no = list(ans_groups.keys())
        ans_start_no.sort(key=takefirst)  # 排序

    contained_no = []  # 答案参与重组的题号
    for ans_no in ans_start_no:
        st1, ed1 = ans_no.split("-")
        contained_no.extend(list(range(int(st1)-1, int(ed1))))
    not_contained_no = set(range(len(res_list))) - set(contained_no)

    added_nos = []  # 已经slave了的真实题号
    # 开始是单层题型结构时
    temp_no = -1
    while groups_data and groups_data[start_no[0]] == "fei":
        if start_no[0] - temp_no == 1:
            new_res_dict.append(res_list[start_no[0]])
        elif start_no[0] - temp_no > 1:
            new_res_dict.extend(res_list[temp_no+1: start_no[0]+1])
        temp_no = start_no[0]
        del start_no[0]
    if start_no[0] > 0:
        new_res_dict.extend(res_list[temp_no+1:start_no[0]])

    one_group = {}
    alone_item_nos = []
    print("start_no:", start_no)
    fei_no = {}
    for n, group_no in enumerate(start_no):
        # print("added_nos:", added_nos, group_no)
        if "com_stem" not in res_list[group_no]:   # 不带"com_stem"
            if group_no == start_no[-1] and groups_data[group_no] == "fei":  # 最后一个不管
                continue
            if groups_data[group_no] == "fei":  # 可能存在多个fei
                if n - 1 in fei_no:
                    new_res_dict.extend(res_list[fei_no[n - 1] + 1: group_no + 1])
                else:
                    new_res_dict.append(res_list[group_no])
                added_nos.append(group_no+1)
                fei_no[n] = group_no
                continue
            # 其他情况
            new_res_dict.append(res_list[group_no])
            added_nos.append(group_no+1)
            # continue
        else:
            # 遇到带"com_stem"的试题
            one_group["com_stem"] = res_list[group_no]["com_stem"]
            del res_list[group_no]["com_stem"]

        if "-" in groups_data[group_no]:  # slave范围知道时
            st, end = groups_data[group_no].split("-")
            if not added_nos:  # 开始
                if n + 1 < len(start_no) and start_no[n + 1] <= int(end):  # 公共题文中的结束序号信息有误，以下一个题的key值为主
                    one_group["slave"] = res_list[int(st) - 1: start_no[n + 1]]
                    added_nos.append(start_no[n + 1])
                else:
                    added_nos.append(int(end))
                    if len(res_list) < int(end):
                        st = int(st) - int(res_list[0]["topic_num"]) + 1
                        end = int(end) - int(res_list[0]["topic_num"]) + 1
                    one_group["slave"] = res_list[int(st) - 1:int(end)]
            elif int(st) <= added_nos[-1]:  # 公共题文中的初始序号信息有误
                if n + 1 < len(start_no):  # 不是最后一组
                    if int(end) < start_no[n + 1]:
                        one_group["slave"] = res_list[added_nos[-1]:int(end)]
                        added_nos.append(int(end))
                    else:  # 结束序号有误，
                        one_group["slave"] = res_list[added_nos[-1]: start_no[n + 1]]
                        added_nos.append(start_no[n + 1])
                else:
                    if int(end) >= added_nos[-1]:
                        one_group["slave"] = res_list[added_nos[-1]:int(end)]
                        added_nos.append(int(end))
                    else:  # end值出错
                        if str(group_no+1) + "-" in "#".join(ans_groups.keys()):
                            end = re.search("[^#]{}-(\d+)[$#]".format(group_no), "#".join(ans_groups.keys())).group(1)
                            one_group["slave"] = res_list[group_no: int(end) + 1]
                        else:
                            endp = [m for m, j in enumerate(res_list[added_nos[-1]:])
                                    if j["type"] != res_list[added_nos[-1]]["type"]]
                            if endp:
                                one_group["slave"] = res_list[added_nos[-1]:endp[0] + len(res_list[:added_nos[-1]])]
                                added_nos.append(endp[0] + len(res_list[:added_nos[-1]]))
                            else:
                                one_group["slave"] = res_list[group_no:]
            else:
                added_nos.append(int(end))
                one_group["slave"] = res_list[int(st) - 1:int(end)]
                if int(st) > added_nos[-1] + 1:
                    new_res_dict.extend(res_list[added_nos[-1]:int(st) - 1])
        else:  # salve范围不知道时
            # print("yyy:", group_no, start_no, groups_data)
            if group_no != start_no[-1]:  # 不是最后一个
                if groups_data[group_no] == "fei":  # 可能存在多个fei
                    if n - 1 in fei_no:
                        new_res_dict.append(res_list[fei_no[n-1]+1: group_no+1])
                    else:
                        new_res_dict.append(res_list[group_no])
                    added_nos.append(group_no+1)
                    fei_no[n] = group_no
                    continue
                elif "#" + str(group_no + 1) + "-" in "#" + "#".join(ans_groups.keys()):   # 以答案的序号为准
                    aa = ("#" + "#".join(ans_groups.keys())).split("#{}-".format(group_no + 1))
                    end = aa[-1].split("#", maxsplit=1)[0]
                    one_group["slave"] = res_list[group_no: int(end)]
                    added_nos.append(int(end))
                    if int(end) < start_no[n+1]:  # 中间单独的题目
                        alone_item_nos.append([int(end), start_no[n + 1]])
                        added_nos.append(start_no[n + 1])
                    # new_res_dict.extend(res_list[int(end)+1:])
                else:
                    one_group["slave"] = res_list[group_no: start_no[n+1]]
                    added_nos.append(start_no[n+1])
            else:
                if groups_data[group_no] == "fei":
                    continue
                elif "#{}-".format(group_no + 1) in "#" + "#".join(ans_groups.keys()):  # 以答案的序号为准
                    aa = ("#" + "#".join(ans_groups.keys())).split("#{}-".format(group_no + 1))
                    end = aa[-1].split("#", maxsplit=1)[-1]
                    one_group["slave"] = res_list[group_no: int(end)]
                    added_nos.append(int(end))
                else:
                    # 只根据题型来判断截止点，不靠谱，先按作答类型
                    endp = []
                    if added_nos:
                        endp = [m for m, j in enumerate(res_list[added_nos[-1]:]) if "answer_type" in j and
                                "作文" in j["answer_type"] and j["answer_type"] != res_list[added_nos[-1]]["answer_type"]]
                        if not endp or (endp[0] <= 1 and len(res_list[added_nos[-1]:]) > 6) or endp[0] > 6:  # 默认最多小题是6个
                            endp = [m for m, j in enumerate(res_list[added_nos[-1]:]) if "answer_type" in j
                                    and "answer_type" in res_list[added_nos[-1]] and
                                    j["answer_type"] != res_list[added_nos[-1]]["answer_type"]]
                    # print("endp:", endp, )
                    if endp and endp[0] > 1:
                        one_group["slave"] = res_list[added_nos[-1]:endp[0] + len(res_list[:added_nos[-1]])]
                        added_nos.append(endp[0] + len(res_list[:added_nos[-1]]))
                        # new_res_dict.extend(res_list[added_nos[-1]:])
                    else:
                        one_group["slave"] = res_list[group_no:]
                        added_nos.append(len(res_list))

        one_group["type"] = one_group["slave"][0]["type"] if one_group["slave"] else ""
        one_group["que_num"] = len(one_group["slave"])
        if one_group["slave"]:
            if one_group["slave"][-1]["topic_num"] != one_group["slave"][0]["topic_num"]:
                one_group["topic_num"] = "{}-{}".format(one_group["slave"][0]["topic_num"],
                                                        one_group["slave"][-1]["topic_num"])
            else:
                one_group["topic_num"] = one_group["slave"][0]["topic_num"]
        else:
            one_group["topic_num"] = ""

        # ---------小题答案拆分----------
        one_group = ans_regroup(ans_start_no, one_group, ans_groups)
        if "key" not in one_group:
            one_group["key"] = ""
        if "parse" not in one_group:
            one_group["parse"] = ""

        new_res_dict.append(one_group)
        if alone_item_nos:  # 把中间单独的不参与重组的题目也加上
            for alone_no in alone_item_nos:
                new_res_dict.extend(res_list[alone_no[0]: alone_no[1]])
            alone_item_nos = []
        one_group = {}

    # 把末尾不参与重组的题目也加上
    if added_nos[-1] < len(res_list):
        new_res_dict.extend(res_list[added_nos[-1]:])
    # ---------------------题目重组end----------------------------------
    if not_contained_no:  # 答案不参与重组的题号
        for one_no in not_contained_no:
            for idx, one_res in enumerate(new_res_dict):
                if one_no+1 == one_res["topic_num"]:
                    parse_split2group(one_res)

    # ------对重组后的每个大题更新题型结构，并对公共题文初步添加缩进-------
    for one_res in new_res_dict:
        if "com_stem" in one_res:  # 公共题文中暂不考虑填空个数
            # 添加缩进属性<p style="text-indent: 2em">、居中属性<p style="text-align:center">
            # new_com_stem = suojin(one_res["com_stem"])
            new_com_stem = one_res["com_stem"]
            new_com_stem = re.sub(r"(完成|回答)下?[面列]?的?第?(\d{1,2})[-到至第～~－]+?(\d{1,2})小?题",
                                  r"\1下面小题", new_com_stem)
            new_com_stem = new_com_stem.replace("  ", "&nbsp;&nbsp;")  # 允许手动调整的空格保留
            # 字体设置：带缩进的行均设置为楷体，其他是宋体
            new_com_stem = re.sub(r'(<p style="text-indent:.*?">\n*|<p style="text-align: center;?">\n*'
                                  r'|<p style="text-align: right;?">\n*)([^\n]+?)', r'\1<span style="font-family:楷体;">\2',
                                  new_com_stem)
            one_res["stem"] = new_com_stem + "\n" + one_res["stem"] if "stem" in one_res else new_com_stem
            del one_res["com_stem"]
        elif "slave" in one_res and one_res["slave"] and "stem" in one_res:  # 带小问的试题题文也设置一下字体
            lw_com_stem = re.sub(r'(<p style="text-indent:.*?">\n*|<p style="text-align: center;?">\n*'
                                  r'|<p style="text-align: right;?">\n*)([^\n]+?)', r'\1<span style="font-family:楷体;">\2',
                                  one_res["stem"])
            one_res["stem"] = lw_com_stem

        one_res["topic_num"] = str(one_res["topic_num"])
        if "slave" in one_res:
            one_res['type'] = '小题多问类'
        elif "options" in one_res:
            one_res['type'] = '选择类'
        else:
            one_res['type'] = '解答类'
            if re.search("(阅读|针对|结合).{,4}[资材]料|(\n|^)\s*材料一\s", one_res['stem']):
                one_res["stem"] = re.sub(r'(<p style="text-indent:.*?">\n*|<p style="text-align: center;?">\n*'
                                         r'|<p style="text-align: right;?">\n*)([^\n]+?)',
                                         r'\1<span style="font-family:楷体;">\2',
                                         one_res["stem"])
            # ind_label = '<p style="text-indent: 2em">'
            # if "【范文】" in one_res['key']:  # "写作"
            #     anss = re.split("\n+", one_res['key'])
            #     ids = [n for n, a in enumerate(anss) if "【范文】" in a][0]
            #     may_title = anss[ids].replace("【范文】", "").strip()
            #     if not may_title:
            #         ids += 1
            #     may_title = anss[ids].strip()
            #     if 0 < len(may_title) < 5:
            #         new_ans = "\n".join(anss[:ids]) + '<p style="text-align:center">' + anss[ids] + "</p>" \
            #                   + ind_label + ('</p>' + ind_label).join(anss) + "</p>"
            #     else:
            #         new_ans = ind_label + '</p><p style="text-indent: 2em">'.join(anss) + "</p>"
            #     one_res['key'] = new_ans
            # elif re.search(r"(阅读|针对).{,4}[资材]料|(\n|^)\s*材料一\s", one_res['stem']) \
            #         and "text-indent: 2em" not in one_res['stem']:
            #     one_res['stem'] = suojin(one_res['stem'])
    return new_res_dict


def ans_regroup(ans_start_no, one_group, ans_groups):
    """
    答案重组
    ans_start_no：ans_groups中的题号组
    :return:
    """
    if ans_start_no:
        for k in ans_start_no:
            if k == one_group["topic_num"]:
                st1, end1 = k.split("-")  # 真实题号组
                # --------------------------解析----------------------------
                parse_list = []
                if len(re.findall("【详解】", ans_groups[k]["parse"])) > 1:
                    parse_list = re.split("【详解】", ans_groups[k]["parse"])
                    comm_parse, parse_list = parse_list[0], parse_list[1:]
                else:
                    t_seq_no = list(range(int(st1), int(end1) + 1))
                    t_seq_no = list(map(str, t_seq_no))
                    if any([True if len(no) > 1 else False for no in t_seq_no]):
                        parse_list = re.split(r"(?<=[】\n])\s*(" + "|".join(t_seq_no) + r")\s*[、.．､]",
                                              "\n" + ans_groups[k]["parse"])
                        comm_parse, parse_list = parse_list[0], parse_list[1:]
                        parse_list = [pr for idn, pr in enumerate(parse_list) if idn % 2 == 1]
                    else:
                        parse_list = re.split(r"(?<=[】\n])\s*[" + "".join(t_seq_no) + r"]\s*[、.．､]",
                                              "\n" + ans_groups[k]["parse"])
                        comm_parse, parse_list = parse_list[0], parse_list[1:]

                if len(parse_list) > 1:
                    if len(parse_list) == int(end1) + 1 - int(st1):
                        if comm_parse:
                            one_group["parse"] = comm_parse
                        for i in range(len(parse_list)):
                            pr = parse_list[i].strip()
                            if i == len(parse_list) - 1 and re.search("\n\s*[【参考]*?译文\s*[】:：]|\n\s*【点睛】", pr):
                                pr, hd, one_group["parse"] = re.split("\n\s*([【参考]*?译文\s*[】:：]|【点睛】)",
                                                                      pr, maxsplit=1)
                                one_group["parse"] = hd + one_group["parse"]
                            one_group["slave"][i]["parse"] = pr
                            if "slave" in one_group["slave"][i]:  # 解析再拆-->小问解析
                                slave_parse_list = re.split("(?<=[\s\n])[（(]\s*\d{1,2}[）)]", "\n" + pr)
                                if len(slave_parse_list) - 1 == len(one_group["slave"][i]["slave"]):
                                    for pi in range(len(slave_parse_list) - 1):
                                        one_group["slave"][i]["slave"][pi]["parse"] = slave_parse_list[pi + 1].strip()
                                    one_group["slave"][i]["parse"] = slave_parse_list[0].strip()
                    else:
                        # 就将各题解析合在一起
                        one_group["parse"] = ans_groups[k]["parse"]
                else:
                    one_group['parse'] = ans_groups[k]["parse"]
                # --------------------------答案----------------------------
                ans_list = re.split("(?<=[】\n])\d{1,2}\s*[、.．､]|\s{2,}\d{1,2}\s*[、.．､]",
                                    "\n" + ans_groups[k]["key"])[1:]
                if len(ans_list) > 1:
                    if len(ans_list) == int(end1) + 1 - int(st1):
                        for j in range(len(ans_list)):
                            one_group["slave"][j]["key"] = ans_list[j].strip()
                            if "slave" in one_group["slave"][j]:  # 答案再拆
                                slave_ans = re.sub(r"([（(]\s*\d\s*[）)])\s*[、.．､，,：:]\s*\1", r"\1", ans_list[j])
                                slave_ans_list = re.split("(?<=[\s\n])[（(]\s*\d{1,2}[）)]", "\n" + slave_ans.strip())
                                if len(slave_ans_list) - 1 == len(one_group["slave"][j]["slave"]):
                                    for aj in range(len(slave_ans_list) - 1):
                                        one_group["slave"][j]["slave"][aj]["key"] = slave_ans_list[aj + 1].strip()
                                    one_group["slave"][j]["key"] = slave_ans_list[0].strip()
                    else:
                        one_group["key"] = ans_groups[k]["key"]
                    # ans_start_no.remove(k)
                    # break
                else:
                    one_group['key'] = ans_groups[k]["key"]
                # 先暂时不去掉
                # for si, s in enumerate(one_group["slave"]):
                #     if "errmsgs" in s:
                #         del one_group["slave"][si]["errmsgs"]
    else:  # ans_groups为空时
        # 针对答案在后面且【答案】1.xx 2.xx \n【解析】1.xx 2.xx  \n【答案】3.xx 4.xx \n【解析】3.xx 4.xx
        #                 或1.xx 2.xx \n【解析】1.xx 2.xx  \n  3.xx 4.xx \n【解析】3.xx 4.xx
        if (one_group["slave"][0]["parse"] in ["略", ""] or one_group["slave"][0]["key"] in ["略", "", "见解析"]) \
            and ("-"in str(one_group["topic_num"]) and len(one_group["slave"]) > 1
                 and one_group["slave"][-1]["parse"].strip()):
            st1, end1 = one_group["topic_num"].split("-")  # 真实题号组
            t_seq_no = list(range(int(st1), int(end1) + 1))
            t_seq_no = list(map(str, t_seq_no))
            parse_list = re.split(r"(?<=[】\n])\s*(" + "|".join(t_seq_no) + r")\s*[、.．､]",
                                  "\n" + one_group["slave"][-1]["parse"])
            comm_parse, parse_list = parse_list[0], parse_list[1:]
            parse_list = [pr.strip() for idn, pr in enumerate(parse_list) if idn % 2 == 1]
            if len(parse_list) in [int(end1) + 1 - int(st1), int(end1) - int(st1)]:
                if comm_parse:
                    one_group["parse"] = comm_parse
                for ni, pr in enumerate(parse_list):
                    if ni == int(end1) - int(st1):  # 最后一个
                        pr = re.sub("\n\s*【答案】$", "", pr)
                        if re.search("\n\s*[【参考]*?译文\s*[】:：]|\n\s*【点睛】", pr):
                            pr, hd, one_group["parse"] = re.split("\n\s*([【参考]*?译文\s*[】:：]|【点睛】)",
                                                                  pr, maxsplit=1)
                            one_group["parse"] = hd + one_group["parse"]
                    if one_group["slave"][ni]["key"] in ["略", "", "见解析"]:
                        one_group["slave"][ni]["key"] = one_group["slave"][ni]["parse"]
                    one_group["slave"][ni]["parse"] = pr
    return one_group


def suojin(item_str):
    """
    文本缩进处理
    :param item_str:
    :return:
    """
    ind_label = '<p style="text-indent: 2em">'
    con_list = re.split("\n+", item_str.strip())
    if len(con_list) > 1 and re.search("(阅读|针对).{,4}[资材]料", con_list[0]):
        new_con = con_list[0] + ind_label + ('</p>' + ind_label).join(con_list[1:]) + "</p>"
    else:
        new_con = ind_label + ('</p>' + ind_label).join(con_list) + "</p>"
    new_con = re.sub(r'<p style="text-indent: 2em">(\s*<img .+?)</p>($|<p style="text-indent: 2em">)',
                     r'\1\n\2', new_con, flags=re.S).strip()
    return new_con


def parse_split2group(item_list):
    """
    有slave的题目将外层的解析拆入salve中
    :return:
    """
    # print(item_list)
    raw_item_list = item_list.copy()
    flag = 0
    # print(item_list)
    if "com_stem" in item_list and "slave" in item_list and len(item_list["slave"]) == 1:  # 嵌套
        item_list = item_list["slave"][0]
        flag = 1

    if "slave" in item_list and (item_list["key"] or item_list["parse"]) and \
               any([True if not (s["key"] + s["parse"]).strip() else False for s in item_list["slave"]]):
        # 解析
        parse_list = re.split(r"(?<=[\s\n】])[（(]\s*[\dl]{1,2}\s*[）)]", "\n" + item_list["parse"].strip())
        if len(parse_list) - 1 == len(item_list["slave"]):
            for pi in range(len(parse_list) - 1):
                item_list["slave"][pi]["parse"] = parse_list[pi + 1].strip()
            item_list["parse"] = parse_list[0].strip()
        # 答案
        ans = re.sub(r"([（(]\s*\d\s*[）)])\s*[、.．､，,：:]\s*(\1)", r"\2", item_list["key"])
        ans_list = re.split("(?<=[\s\n】])[（(]\s*[\dl]{1,2}\s*[）)]", "\n" + ans.strip())
        if len(ans_list) - 1 == len(item_list["slave"]):
            for aj in range(len(ans_list) - 1):
                item_list["slave"][aj]["key"] = ans_list[aj + 1].strip()
            item_list["key"] = ans_list[0].strip()
    # 2021-12-21
    if "com_stem" in item_list:
        item_list["stem"] = item_list["com_stem"].strip() + "<br/>" + item_list["stem"] \
            if "stem" in item_list else item_list["com_stem"]
        del item_list["com_stem"]
    if flag:
        raw_item_list["slave"] = [item_list]
        item_list = raw_item_list
    return item_list


def regroup_old(res_list, item_groups):
    """
    将多个题共用一个题干的情况进行slave重组,如地理
    :param res_list:  拆分为小题后的结果
    :return:
    """
    new_res_dict = []
    start_no = [i for i in item_groups.keys() if i != "pos"]
    if not start_no:
        return res_list

    def takefirst(elem):
        return int(elem.split("-")[0])

    start_no.sort(key=takefirst)  # 排序
    print(start_no)
    one_group = {}
    added_nos = []  # 已经slave了的题号
    for n, group_no in enumerate(start_no):
        one_group["common_stem"] = item_groups[group_no]
        st, end = group_no.split("-")  # 真实题号组
        if not added_nos:  # 开始
            if item_groups["pos"][n + 1] <= int(end):  # 公共题文中的结束序号信息有误
                one_group["slave"] = res_list[int(st) - 1:item_groups["pos"][n + 1] - 1]
                added_nos.append(item_groups["pos"][n + 1] - 1)
            else:
                added_nos.append(int(end))
                one_group["slave"] = res_list[int(st) - 1:int(end)]
        elif int(st) <= added_nos[-1]:  # 公共题文中的初始序号信息有误
            if n + 1 < len(item_groups["pos"]):  # 不是最后一组
                if int(end) < item_groups["pos"][n + 1]:
                    one_group["slave"] = res_list[added_nos[-1]:int(end)]
                    added_nos.append(int(end))
                else:  # 结束序号有误，以pos为主
                    one_group["slave"] = res_list[added_nos[-1]:item_groups["pos"][n + 1] - 1]
                    added_nos.append(item_groups["pos"][n + 1] - 1)
            else:
                if int(end) >= added_nos[-1]:
                    one_group["slave"] = res_list[added_nos[-1]:int(end)]
                    added_nos.append(int(end))
                else:  # end值出错
                    endp = [m for m, j in enumerate(res_list[added_nos[-1]:])
                            if j["type"] != res_list[added_nos[-1]]["type"]]
                    if endp:
                        one_group["slave"] = res_list[added_nos[-1]:endp[0] + len(res_list[:added_nos[-1]])]
                        added_nos.append(endp[0] + len(res_list[:added_nos[-1]]))
        else:
            added_nos.append(int(end))
            one_group["slave"] = res_list[int(st) - 1:int(end)]
            if int(st) > added_nos[-1] + 1:
                new_res_dict.extend(res_list[added_nos[-1]:int(st) - 1])

        one_group["type"] = one_group["slave"][0]["type"]
        one_group["que_num"] = len(one_group["slave"])
        new_res_dict.append(one_group)
        one_group = {}
    if added_nos[-1] < len(res_list):
        new_res_dict.extend(res_list[added_nos[-1]:])

    return new_res_dict