EnglishPal/app/Article.py

import os
import random
import pickle_idea
from difficulty import get_difficulty_level_for_user, text_difficulty_level, user_difficulty_level
from UseSqlite import RecordQuery

path_prefix = '/var/www/wordfreq/wordfreq/'
path_prefix = './'  # comment this line in deployment


def total_number_of_essays():
    '''
        得到文章总数
        return:文章数目
    '''
    rq = RecordQuery(path_prefix + 'static/wordfreqapp.db')
    rq.instructions("SELECT * FROM article")
    rq.do()
    result = rq.get_results()
    return len(result)


def get_article_title(s):
    '''
    得到文章的标题
    '''
    return s.split('\n')[0]


def get_article_body(s):
    '''
    得到文章的内容
    '''
    lst = s.split('\n')
    lst.pop(0)  # remove the first line
    return '\n'.join(lst)


def get_today_article(user_word_list, visited_articles):
    '''
    根据用户的单词列表和阅读过的文章返回需要的文章的全部信息
    '''
    rq = RecordQuery(path_prefix + 'static/wordfreqapp.db')
    if visited_articles is None:
        visited_articles = {
            "index": 0,  # 为 article_ids 的索引
            "article_ids": []  # 之前显示文章的id列表，越后越新
        }
    if visited_articles["index"] > len(visited_articles["article_ids"])-1:  # 生成新的文章，因此查找所有的文章
        rq.instructions("SELECT * FROM article")
    else:  # 生成阅读过的文章，因此查询指定 article_id 的文章
        # 可能因为直接刷新页面导致直接去查询了'null'，因此当刷新的页面的时候，需要直接进行“上一篇”操作
        if visited_articles["article_ids"][visited_articles["index"]] == 'null':
            visited_articles["index"] -= 1
            visited_articles["article_ids"].pop()
        rq.instructions(
            f'SELECT * FROM article WHERE article_id='
            f'{visited_articles["article_ids"][visited_articles["index"]]}'
        )
    rq.do()
    result = rq.get_results()
    random.shuffle(result)

    # Choose article according to reader's level
    d1 = load_freq_history(path_prefix + 'static/frequency/frequency.p')
    d2 = load_freq_history(path_prefix + 'static/words_and_tests.p')
    d3 = get_difficulty_level_for_user(d1, d2)

    d = None
    result_of_generate_article = "not found"
    d_user = load_freq_history(user_word_list)
    # 更多的考虑，因为用户的行为是动态的。应考虑时间因素。
    user_level = user_difficulty_level(d_user, d3)
    text_level = 0
    if visited_articles["index"] > len(visited_articles["article_ids"])-1:  # 生成新的文章
        amount_of_visited_articles = len(visited_articles["article_ids"])
        amount_of_existing_articles = len(result)
        # 如果当前阅读过的文章的数量 == 存在的文章的数量，即所有的书本都阅读过了
        if amount_of_visited_articles == amount_of_existing_articles:
            result_of_generate_article = "had read all articles"
        else:
            for k in range(3):  # 最多尝试3次
                for reading in result:
                    text_level = text_difficulty_level(reading['text'], d3)
                    # 从高斯分布中得出的平均值为 0.8，站位偏差为 1 的数字
                    factor = random.gauss(0.8, 0.1)
                    # 新的文章之前没有出现过且符合一定范围的水平
                    if reading['article_id'] not in visited_articles["article_ids"] and within_range(text_level, user_level, (8.0 - user_level) * factor):
                        d = reading
                        visited_articles["article_ids"].append(d['article_id'])  # 列表添加新的文章id；下面进行
                        result_of_generate_article = "found"
                        break
                if result_of_generate_article == "found":  # 用于成功找到文章后及时退出外层循环
                    break
        if result_of_generate_article != "found":  # 阅读完所有文章，或者循环3次没有找到适合的文章，则放入空（“null”）
            visited_articles["article_ids"].append('null')
    else:  # 生成已经阅读过的文章
        d = random.choice(result)
        text_level = text_difficulty_level(d['text'], d3)
        result_of_generate_article = "found"

    today_article = None
    if d:
        today_article = {
            "user_level": f'{user_level:4.2f}',
            "text_level": f'{text_level:4.2f}',
            "date": d['date'],
            "article_title": get_article_title(d['text']),
            "article_body": get_article_body(d['text']),
            "source": d["source"],
            "question": get_question_part(d['question']),
            "answer": get_answer_part(d['question'])
        }

    return visited_articles, today_article, result_of_generate_article


def load_freq_history(path):
    '''
    加载历史路径
    '''
    d = {}
    if os.path.exists(path):
        d = pickle_idea.load_record(path)
    return d


def within_range(x, y, r):
    '''
    判断x>y并且x-y<=r
    '''
    return x > y and abs(x - y) <= r


def get_question_part(s):
    '''
    得到问题部分
    '''
    s = s.strip()
    result = []
    flag = 0
    for line in s.split('\n'):
        line = line.strip()
        if line == 'QUESTION':
            result.append(line)
            flag = 1
        elif line == 'ANSWER':
            flag = 0
        elif flag == 1:
            result.append(line)
    return '\n'.join(result)


def get_answer_part(s):
    '''
    得到答案部分
    '''
    s = s.strip()
    result = []
    flag = 0
    for line in s.split('\n'):
        line = line.strip()
        if line == 'ANSWER':
            flag = 1
        elif flag == 1:
            result.append(line)
    return '\n'.join(result)