从NLP视角看电视剧《狂飙》，会有什么发现？

def parse_comments(url):
    """
        解析HTML页面,获得评论及相关数据
    :param url:
    :return:
    """
    html = get_html(url)
    soup_comment = BeautifulSoup(html, 'html.parser')

    # 所有获取的一页数据
    data_page = []

    # 提取评论
    comments_all = soup_comment.findAll("div", "comment-item")
    for comments in comments_all:
        try:
            # 解析评论及相关数据
            comment_info = comments.find("span", "comment-info")  # 评论id相关信息
            comment_vote = comments.find("span", "comment-vote")  # 评论点赞信息
            comment_content = comments.find("span", "short").text.replace("\n", "")  # 评论内容

            # 提取需要的各字段信息
            info_list = comment_info.findAll("span")
            star_rating = info_list[1]

            user_name = comment_info.find("a").text
            video_status = info_list[0].text  # 电视剧观看状态
            comment_score = int(star_rating["class"][0][-2:])  # 评论分值
            comment_level = star_rating["title"]  # 评论等级
            comment_time = info_list[2].text.replace("\n", "").replace("    ", "")  # 评论时间
            # print(info_list)
            comment_location = info_list[3].text  # 评论位置

            comment_vote_count = int(comment_vote.find("span", "votes vote-count").text)  # 评论被点赞数

            # 获取的一条数据
            # ["用户名", "电视剧观看状态", "评论分数", "评论等级", "评论时间", "评论位置", "评论点赞数", "评论"]
            data_row = [user_name, video_status, comment_score, comment_level,
                        comment_time, comment_location,
                        comment_vote_count, comment_content]
            data_page.append(data_row)
        except:
            # 跳过解析异常的数据
            continue

    return data_page

    html = get_html(url)
    soup_info = BeautifulSoup(html, 'html.parser')

    # 获得的结果信息
    result_info_dict = {}

    # 提取评论
    info_all = soup_info.findAll("div", "info")
    for info in info_all:
        info_name = info.find("span", "name").text
        info_role = info.find("span", "role").text
        info_works_list = info.find("span", "works").findAll("a")

def generate_celebrity_graph():
    """
        构建演职员关系图谱
    :return:
    """
    df = pd.read_csv("./data/狂飙演职员信息表.csv")
    data = df.values.tolist()

    # 转换格式
    nodes = []
    links = []
    nodes_name = []

    symbolSize_dict = {"姓名": 30, "角色": 20, "饰演人物": 20, "代表作": 20}
    categories = [{"name": x} for x in symbolSize_dict.keys()]

    for row in data:
        # 姓名、角色(";"分割多个)、饰演人物(可能为空)、代表作(";"分割多个)
        name, role, role_to_play, works = row
        role_list = role.split(";")
        works_list = works.split(";")

        if name not in nodes_name:
            nodes_name.append(name)
            # 一个节点
            node = {
                "name": name,
                "symbolSize": symbolSize_dict["姓名"],
                "category": "姓名",
            }
            nodes.append(node)

        for role_temp in role_list:
            if role_temp not in nodes_name:
                nodes_name.append(role_temp)
                node = {
                    "name": role_temp,
                    "symbolSize": symbolSize_dict["角色"],
                    "category": "角色",
                }
                nodes.append(node)

            link = {
                "source": name,
                "target": role_temp
            }
            links.append(link)

            if role_temp == "演员":
                if role_to_play not in nodes_name:
                    nodes_name.append(role_to_play)
                    node = {
                        "name": role_to_play,
                        "symbolSize": symbolSize_dict["饰演人物"],
                        "category": "饰演人物",
                    }
                    nodes.append(node)

                link = {
                    "source": name,
                    "target": role_to_play
                }
                links.append(link)

        for works_temp in works_list:
            if works_temp not in nodes_name:
                nodes_name.append(works_temp)
                if works_temp == "狂飙":
                    node = {
                        "name": works_temp,
                        "symbolSize": 50,  # 特别设置
                        "category": "代表作",
                    }
                else:
                    node = {
                        "name": works_temp,
                        "symbolSize": symbolSize_dict["代表作"],
                        "category": "代表作",
                    }
                nodes.append(node)

            link = {
                "source": name,
                "target": works_temp
            }
            links.append(link)

    c = (
        Graph(init_opts=opts.InitOpts(theme=ThemeType.CHALK, width="1500px", height="1000px"))
        .add(
            "",
            nodes,
            links,
            categories,
            repulsion=1000,
            linestyle_opts=opts.LineStyleOpts(curve=0.6),
        )
        .set_global_opts(
            legend_opts=opts.LegendOpts(pos_left=100, pos_top=350, orient="vertical"),
            title_opts=opts.TitleOpts(title="人物关系图谱"),
        )
        .render("./result/演职员图谱.html")
    )
    print("演职员关系图谱，保存路径为：./result/演职员图谱.html")