基于django的协同过滤食谱推荐系统、python

1.内容简介

所用技术：scrapy、django、协同过滤

2.过程

2.1爬虫

使用scrapy爬取https://www.douguo.com网站的食谱数据

主要代码：

# 在 parse 方法中获取下一页链接并发送请求
    def parse(self, response):
        # 获取当前页面的菜谱详情链接并发送请求
        for url in response.xpath('//ul[@class="cook-list"]//a[@class="cook-img"]/@href').getall():
            pipei_url = re.sub("/0.*", "", response.url)
            tag_id = self.url_tag[unquote(pipei_url)][0]
            tag_name = self.url_tag[unquote(pipei_url)][1]
            m = {"tag_id": tag_id, "tag_name": tag_name}
            yield scrapy.Request(url=self.url_root + url, callback=self.parse_detail, meta=m)

        # 解析出下一页链接并发送请求
        next_page = response.xpath('//a[@class="anext"]/@href')
        if next_page:
            yield scrapy.Request(url=next_page[0].get().replace("http", "https"), callback=self.parse)

    def md5_encrypt(self, s):
        md5 = hashlib.md5()
        md5.update(s.encode("utf-8"))
        return md5.hexdigest()

    def parse_detail(self, response, **kwargs):
        tag_id = response.meta["tag_id"]
        tag_name = response.meta["tag_name"]
        divs = response.xpath('//div[@class="step"]/div')
        title = response.xpath("//h1/text()").get()

        # 步骤
        step_list = []
        for div in divs:
            step_img = div.xpath("a/img//@src").get()
            step_index = div.xpath('div[@class="stepinfo"]/p/text()').get()
            step_text = "\n".join(div.xpath('div[@class="stepinfo"]/text()').getall()).strip()
            step_list.append([step_img, step_index, step_text])

        # 配料
        mix_list = []
        for td in response.xpath("//table/tr/td"):
            mix_name = td.xpath('span[@class="scname"]//text()').get()
            mix_cot = td.xpath('span[@class="right scnum"]//text()').get()
            mix_list.append([mix_name, mix_cot])

        info_item = FoodInfoItem(
            tag_id=tag_id,
            tag_name=tag_name,
            food_id=self.md5_encrypt(response.url),
            food_url=response.url,
            title=title,
            step_list=str(step_list),
            img=response.xpath('//*[@id="banner"]/a/img/@src').get(),
            desc1="\n".join(response.xpath('//p[@class="intro"]/text()').getall()).strip(),
            mix_list=str(mix_list),
            all_key=tag_name + title + str(mix_list),
        )
        yield info_item

2.2django展示

2.2.1主页展示（协同过滤推荐）

 2.2.2搜索功能

 2.2.3 饮食营养构成

 2.3协同过滤推荐

基于用户的协同过滤推荐算法

def recommend_by_user_cf(user_id, similarity_matrix, N=10):
    """
    基于用户的协同过滤推荐算法
    Args:
        user_id: 目标用户ID
        similarity_matrix: 用户之间的相似度矩阵
        records: 所有用户的点菜记录，格式为 [(user_id, food_id, eat_date), ...]
        N: 推荐菜品数量
    Returns:
        recommended_foods: 推荐的菜品列表，格式为 [(food_id_1, score_1), (food_id_2, score_2), ...]
    """
    # 找到和指定用户吃过相同菜品的其他用户
    user_eating_records = EatingRecord.objects.filter(user_id=user_id).values("food_id", "eat_date")
    user_food_ids = [record["food_id"] for record in user_eating_records]
    similar_users = []
    for i in range(similarity_matrix.shape[0]):
        if i != user_id - 1:
            similarity = similarity_matrix[user_id - 1, i]
            if similarity > 0:
                # 找到该相似用户在最近M天内吃过的所有菜品
                similar_user_eating_records = (
                    EatingRecord.objects.filter(user_id=i + 1).exclude(food_id__in=user_food_ids).values("food_id", "eat_date").order_by("-eat_date")[:10]
                )
                similar_user_food_ids = [record["food_id"] for record in similar_user_eating_records]

                # 计算该相似用户与指定用户之间的相似度，并加入相似用户列表中
                similar_users.append((i + 1, similarity, similar_user_food_ids))

    # 统计所有相似用户对每个菜品的兴趣度得分
    scores = {}
    for similar_user_id, similarity, similar_user_food_ids in similar_users:
        for food_id in similar_user_food_ids:
            if food_id not in user_food_ids:
                scores[food_id] = scores.get(food_id, 0) + similarity

    # 按照得分从高到低排序，选取前N个菜品作为推荐结果
    sorted_scores = sorted(scores.items(), key=lambda x: x[1], reverse=True)
    recommended_foods = sorted_scores[:N]

    return recommended_foods

python 毕设帮助，指导，源码分享，调试部署：worthy_life_