import requests
import  re
from bs4 import BeautifulSoup
from urllib.parse import urljoin
from collections import Counter
import pandas as pd
from wordcloud import WordCloud
import matplotlib.pyplot as plt

headers = {
        "User-Agent": "Mozilla/5.0 (Windows NT 10.0; Win64; x64) AppleWebKit/537.36 (KHTML, like Gecko) Chrome/128.0.0.0 Safari/537.36 Edg/128.0.0.0",
        "cookie": "CURRENT_FNVAL=4048; buvid_fp_plain=undefined; buvid4=04DF7AEF-34D9-CC62-690A-D369B35D458509591-023061415-%2FxwqHe8zHTWav6Q4ZiB1Ag%3D%3D; enable_web_push=DISABLE; FEED_LIVE_VERSION=V_WATCHLATER_PIP_WINDOW3; PVID=1; buvid3=D5B12366-476E-6163-1D79-774D300DF97306537infoc; b_nut=1718270506; _uuid=243B710F9-1010E3-9654-E867-4A8D8BB10AB1307743infoc; header_theme_version=CLOSE; rpdid=0zbfAHMKHr|S8rGMSwG|1uI|3w1Sum1G; fingerprint=042b265e3c7da3104d09a0692278e922; CURRENT_QUALITY=80; home_feed_column=5; browser_resolution=1659-836; bili_ticket=eyJhbGciOiJIUzI1NiIsImtpZCI6InMwMyIsInR5cCI6IkpXVCJ9.eyJleHAiOjE3MjU5NDEwOTEsImlhdCI6MTcyNTY4MTgzMSwicGx0IjotMX0.j7rN8z5QOwH-7R7gPvyBxJzDLqymAWFfZeFF-QAXoTQ; bili_ticket_expires=1725941031; bp_t_offset_482950113=974463371485118464; buvid_fp=042b265e3c7da3104d09a0692278e922; b_lsid=DDE103767_191D4FCA152"
    }
vedio_sum=0

def contains_ai_or_artificial_intelligence(text):
    ai_pattern = re.compile( r'(\bai\b|人工智能|([\u4e00-\u9fff]|\s|^)ai([\u4e00-\u9fff]|\s|$))', re.IGNORECASE)
    return re.search(ai_pattern, text)

def filter_strings(strings_list):
    return [string for string in strings_list if contains_ai_or_artificial_intelligence(string)]

def top_n_strings(strings_list, n):
    filtered_strings = filter_strings(strings_list)
    # 统计每个字符串的出现次数
    counter = Counter(filtered_strings)
    # 获取出现次数最多的前n个字符串
    most_common = counter.most_common(n)
    return most_common

def dm_urls(url):
    try:
        response = requests.get(url, headers=headers)
        response.raise_for_status()
        soup = BeautifulSoup(response.text, 'html.parser')
        urls = set()
        for a_tag in soup.find_all('a', href=True):
            href = a_tag['href']
            urls.add(href)
        pattern = re.compile(r'https://api\.bilibili\.com/x/v1/dm/list\.so\?oid=\d+')
        # 过滤符合格式的URL
        specific_urls = [url1 for url1 in urls if pattern.match(url1)]
        return specific_urls
    except:
        return []  # 返回一个空列表作为默认值

def extract_urls(url):
    response = requests.get(url, headers=headers)
    response.raise_for_status()
    soup = BeautifulSoup(response.text, 'html.parser')
    urls = set()
    urlbilibili="https://"
    for a_tag in soup.find_all('a', href=True):
        href = a_tag['href']
        # 将相对URL转换为绝对URL
        full_url = urljoin(urlbilibili, href)
        urls.add(full_url)
    pattern = re.compile(r'^https://www\.bilibili\.com/video')
    # 过滤符合格式的URL
    specific_urls = [url1 for url1 in urls if pattern.match(url1)]
    if (len(specific_urls) + vedio_sum > 300):
        specific_urls = specific_urls[:6]  # 保留前六个元素
    return specific_urls

def vedio_transform_port(url):
    response = requests.get(url, headers=headers)
    response.raise_for_status()
    soup = BeautifulSoup(response.text, 'html.parser')
    page_num = []   #储存总共的分p数
    span_tag = None #用做判断有无分p的flag
    div_tags = soup.findAll("div",attrs={"class":"head-left"}) #找到class=head-left的div
    for tag in div_tags:
        span_tag = tag.findAll("span", attrs={"class": "cur-page"})  # 再从中找到class=cur-page的span
    if not span_tag:  # 值为None则为单个视频
        return 0
    else:
        for tag in span_tag:
            pattern = re.compile(r'\((\d+)/(\d+)\)')
            match = pattern.search(tag.text)  # 使用tag.text提取文本内容
            if match:  # 确保找到匹配项
                page_num = int(match.group(2))  # 获取总页数
    return page_num

list=[]

for page_num in range(1,9):
    urls = extract_urls(f"https://search.bilibili.com/video?keyword=%E5%B7%B4%E9%BB%8E%E5%A5%A5%E8%BF%90%E4%BC%9A&from_source=webtop_search&spm_id_from=333.1007&search_source=2&page={page_num}")
    mod_urls=set()
    vedio_sum +=len(urls)
    for ori_url in urls:
        clean_url = re.sub(r'\?.*$', '', ori_url)
        p_sum=vedio_transform_port(ori_url)
        if (p_sum == 0):
            modified_url = clean_url.replace("bilibili.com", "ibilibili.com")
            mod_urls.add(modified_url)
        else:
            for p_num in range(1, p_sum + 1):
                ori2_url = f"{clean_url}?p={p_num}"
                modified_url = ori2_url.replace("bilibili.com", "ibilibili.com")
                mod_urls.add(modified_url)

    for url in mod_urls:
        dmurls=dm_urls(url)
        if dmurls:  # 检查列表是否为空
            dmurl = dmurls[0]
            response = requests.get(dmurl, headers=headers)
            response.encoding = 'utf-8'
            html = response.text
            soup = BeautifulSoup(html, "lxml-xml")
            # 提取所有评论
            comments = soup.find_all('d')
            for comment in comments:
                comment_text = comment.get_text()
                list.append(comment_text)

result = top_n_strings(list,len(list))
# 创建 DataFrame
df = pd.DataFrame(result, columns=['Element', 'Count'])
# 写入 Excel 文件
df.to_excel('output.xlsx', index=False)

result8=top_n_strings(list,8)
for item, count in result8:
    print(f'{item}: {count}')

df = pd.read_excel('output.xlsx')
# 创建词云输入格式
text = ' '.join(df.apply(lambda row: f"{row['Element']} " * row['Count'], axis=1))
# 生成词云
wordcloud = WordCloud(width=800, height=400, background_color='white',font_path=r'C:\Windows\Fonts\FZSTK.TTF').generate(text)
wordcloud.to_file('wordcloud5.png')
# 显示词云图
plt.figure(figsize=(10, 5))
plt.imshow(wordcloud, interpolation='bilinear')
plt.axis('off')
plt.show()