commit0602

flaskProject/.idea/.gitignore

@@ -0,0 +1,8 @@
+# Default ignored files
+/shelf/
+/workspace.xml
+# Editor-based HTTP Client requests
+/httpRequests/
+# Datasource local storage ignored files
+/dataSources/
+/dataSources.local.xml

flaskProject/.idea/flaskProject.iml

@@ -0,0 +1,21 @@
+<?xml version="1.0" encoding="UTF-8"?>
+<module type="PYTHON_MODULE" version="4">
+  <component name="Flask">
+    <option name="enabled" value="true" />
+  </component>
+  <component name="NewModuleRootManager">
+    <content url="file://$MODULE_DIR$">
+      <excludeFolder url="file://$MODULE_DIR$/.venv" />
+    </content>
+    <orderEntry type="inheritedJdk" />
+    <orderEntry type="sourceFolder" forTests="false" />
+  </component>
+  <component name="TemplatesService">
+    <option name="TEMPLATE_CONFIGURATION" value="Jinja2" />
+    <option name="TEMPLATE_FOLDERS">
+      <list>
+        <option value="$MODULE_DIR$/../flaskProject\templates" />
+      </list>
+    </option>
+  </component>
+</module>

flaskProject/.idea/inspectionProfiles/profiles_settings.xml

@@ -0,0 +1,6 @@
+<component name="InspectionProjectProfileManager">
+  <settings>
+    <option name="USE_PROJECT_PROFILE" value="false" />
+    <version value="1.0" />
+  </settings>
+</component>

flaskProject/.idea/misc.xml

@@ -0,0 +1,7 @@
+<?xml version="1.0" encoding="UTF-8"?>
+<project version="4">
+  <component name="Black">
+    <option name="sdkName" value="Python 3.8 (flaskProject)" />
+  </component>
+  <component name="ProjectRootManager" version="2" project-jdk-name="Python 3.8 (flaskProject)" project-jdk-type="Python SDK" />
+</project>

flaskProject/.idea/modules.xml

@@ -0,0 +1,8 @@
+<?xml version="1.0" encoding="UTF-8"?>
+<project version="4">
+  <component name="ProjectModuleManager">
+    <modules>
+      <module fileurl="file://$PROJECT_DIR$/.idea/flaskProject.iml" filepath="$PROJECT_DIR$/.idea/flaskProject.iml" />
+    </modules>
+  </component>
+</project>

flaskProject/app.py

@@ -0,0 +1,185 @@
+from flask import Flask, render_template, request
+import requests
+from lxml import etree
+import pymysql
+
+app = Flask(__name__)
+
+def getDB():
+    db = pymysql.connect(host='localhost',user='root',password='123456',database='douban')
+    return db
+
+
+
+@app.route('/movies')
+def movies():
+    db=getDB()
+    cursor=db.cursor()
+    sql="select film_name,director,actor,language,ranks,rating_num,summary,links from movies"
+    cursor.execute(sql)
+    data=cursor.fetchall()
+    datalist=[]
+    for item in data:
+        datalist.append(item)
+    cursor.close()
+    db.close()
+    return render_template('movies.html',movies=datalist)
+
+@app.route('/index')
+def index():
+    return render_template('index.html')
+
+
+@app.route('/index1')
+def index1():
+    uname=request.args.get('u')
+    psw=request.args.get('p')
+    if uname=="肖旺" and psw=="123":
+        return render_template('index.html')
+    else:
+        return render_template('regist.html')
+
+@app.route('/login')
+def login():
+    return render_template('login.html')
+
+@app.route('/regist')
+def regist():
+    return render_template('login.html')
+
+@app.route('/get_nbadata')
+def get_nbadata():
+    db=getDB()
+    cursor=db.cursor()
+    sql="select no,name,score,team,info from nba"
+    cursor.execute(sql)
+    data=cursor.fetchall()
+    datalist=[]
+    for item in data:
+        datalist.append(item)
+    cursor.close()
+    db.close()
+    return render_template('nba.html',nbadata=datalist)
+
+
+'''@app.route('/get_nbadata')
+def get_nbadata():
+    url='https://nba.hupu.com/stats/players'
+    header={'User-Agent':'Mozilla/5.0 (Windows NT 10.0; Win64; x64) AppleWebKit/537.36 (KHTML, like Gecko) Chrome/122.0.0.0 Safari/537.36 Edg/122.0.0.0'}
+    res=requests.get(url,headers=header)
+    e=etree.HTML(res.text)
+    names=e.xpath('//table[@class="players_table"]//tr/td[2]/a/text()')
+    teams=e.xpath('//table[@class="players_table"]//tr/td[3]/a/text()')
+    nos=e.xpath('//table[@class="players_table"]//tr/td[1]/text()')
+    nos=nos[1::]
+    scores=e.xpath('//table[@class="players_table"]//tr/td[4]/text()')
+    scores=scores[1::]
+    
+    infos=e.xpath('//table[@class="players_table"]//tr/td[2]/a/@href')
+    total_ls = []
+    infos_ls=[]
+    for no, name, team, score in zip(nos, names, teams, scores):
+        st1 = f"排名：{no}  姓名：{name}  队伍：{team}  得分：{score}"
+        total_ls.append(st1)
+    for info in infos:
+        infos_ls.append(info)
+    return render_template('nba.html',total_ls=total_ls,infos_ls=infos_ls)
+
+
+@app.route('/music')
+def get_music():
+
+    headers = {
+        'User-Agent': 'Mozilla/5.0 (Windows NT 10.0; Win64; x64) AppleWebKit/537.36 (KHTML, like Gecko) Chrome/123.0.0.0 Safari/537.36 Edg/123.0.0.0'
+    }
+    url = 'https://www.kugou.com/yy/html/rank.html'
+    res = requests.get(url=url, headers=headers)
+    e = etree.HTML(res.text)
+    music_names = e.xpath('//div[@id="rankWrap"]//ul/li/a/text()')
+    music_urls = e.xpath('//div[@id="rankWrap"]//ul/li/a/@href')
+    return render_template('music.html', music_urls=music_urls, music_names=music_names)
+'''
+
+@app.route('/music')
+def music():
+    db=getDB()
+    cursor=db.cursor()
+    sql="select name,info from music"
+    cursor.execute(sql)
+    data=cursor.fetchall()
+    datalist=[]
+    for item in data:
+        datalist.append(item)
+    cursor.close()
+    db.close()
+    return render_template('music.html',datalist=datalist)
+
+
+@app.route('/get_text')
+def get_text():
+    headers= {
+        'User-Agent': 'Mozilla/5.0 (Windows NT 10.0; Win64; x64) AppleWebKit/537.36 (KHTML, like Gecko) Chrome/123.0.0.0 Safari/537.36'
+    }
+    url = 'https://www.biqg.cc/top/'
+    res = requests.get(url=url,headers=headers)
+    e = etree.HTML(res.text)
+    text_names=e.xpath('//div[@class="wrap rank"]//li/a/text()')
+    text_urls=e.xpath('//div[@class="wrap rank"]//li/a/@href')
+    text_authors=e.xpath('//div[@class="wrap rank"]//li/text()')
+    text_authors2 = []
+    for i in range(len(text_authors)):
+        x = str(text_authors[i]).strip('/').strip('\n').rstrip('\n ')
+        if x=='':
+            text_authors2.append('匿名作者')
+        else:
+            text_authors2.append(x)
+    print(text_authors2)
+    return render_template('text.html',text_names=text_names,text_authors2=text_authors2,text_urls=text_urls)
+
+
+
+@app.route('/manh')
+def get_url_info():
+    db = getDB()
+    cursor = db.cursor()
+    sql = "select names,ranks,renqi,shoucang,authors,urls from manh"
+    cursor.execute(sql)
+    data = cursor.fetchall()
+    datalist = []
+    for item in data:
+        datalist.append(item)
+    cursor.close()
+    db.close()
+    return render_template('manh.html', datalist=datalist)
+
+@app.route('/qm')
+def get_qm_info():
+    db = getDB()
+    cursor = db.cursor()
+    sql = "select names,ranks,authors,summarys,urls from xiaoshuo"
+    cursor.execute(sql)
+    data = cursor.fetchall()
+    datalist = []
+    for item in data:
+        datalist.append(item)
+    cursor.close()
+    db.close()
+    return render_template('qm.html', datalist=datalist)
+
+
+
+@app.route('/about')
+def about():
+    return render_template('about.html')
+
+@app.route('/talk')
+def talk():
+    return render_template('talk.html')
+
+@app.route('/zhanghao')
+def zhanghao():
+    return render_template('login.html')
+
+
+if __name__=='__main__':
+    app.run()

flaskProject/getData.py

@@ -0,0 +1,71 @@
+import pymysql
+import requests
+import re
+from bs4 import BeautifulSoup
+
+def getDB():
+    db = pymysql.connect(host='localhost',user='root',password='123456',database='douban')
+    return db
+
+def Agent_info():
+    headers={
+        'Cookie':'118268"; bid=IO1Dg04kIm4; _vwo_uuid_v2=D1A645C6CFFBAF77D4DD98C62F188490B|754ea089c6c61f5af161e2912f2d4bee; __yadk_uid=NpVAgGKfMbTI28NFopWDALVWvzGMJt3S; _pk_id.100001.4cf6=095af3751c7a7a20.1681143032.; ap_v=0,6.0; __utmc=30149280; __utmc=223695111; dbcl2="279593631:HhdIjxDt0FA"; ck=XIW8; __utma=30149280.966668946.1681143033.1712632454.1712639313.6; __utmb=30149280.0.10.1712639313; __utmz=30149280.1712639313.6.2.utmcsr=accounts.douban.com|utmccn=(referral)|utmcmd=referral|utmcct=/; __utma=223695111.1690211447.1681143033.1712632454.1712639313.6; __utmb=223695111.0.10.1712639313; __utmz=223695111.1712639313.6.2.utmcsr=accounts.douban.com|utmccn=(referral)|utmcmd=referral|utmcct=/; _pk_ref.100001.4cf6=%5B%22%22%2C%22%22%2C1712639313%2C%22https%3A%2F%2Faccounts.douban.com%2F%22%5D; _pk_ses.100001.4cf6=1; push_noty_num=0; push_doumail_num=0; __gads=ID=d8a2141b0213474a-229414c42add00a6:T=1681143032:RT=1712639314:S=ALNI_Mb89dGhTs42z60R9TMxDscyQIzA8A; __gpi=UID=00000bf05307ad13:T=1681143032:RT=1712639314:S=ALNI_MbkC2b_Z_7nO1PL2HHsgHolhWs0iw; __eoi=ID=2f9ca57c63f42bd7:T=1712496871:RT=1712639314:S=AA-AfjbV9P_SdwHly0Xzv8gyJ7ZR',
+        'Host':'movie.douban.com',
+        'User-Agent':'Mozilla/5.0 (Windows NT 10.0; Win64; x64) AppleWebKit/537.36 (KHTML, like Gecko) Chrome/123.0.0.0 Safari/537.36 Edg/123.0.0.0'
+    }
+    return headers
+
+def get_url(url):
+    print("抓取网址",url)
+    headers=Agent_info()
+    request=requests.get(url,headers=headers)
+    soup=BeautifulSoup(request.text,'lxml')
+    pic=soup.find_all(attrs={'class' :'item'})
+    film_urls=[]
+    for x in pic:
+        href=x.a.get('href')
+        film_urls.append(href)
+
+    return film_urls
+
+def get_url_info(film_url,id):
+    print("抓取网址", film_url)
+    headers = Agent_info()
+    request = requests.get(film_url, headers=headers)
+    soup = BeautifulSoup(request.text, 'lxml')
+    ranks=soup.find(attrs={'class':'top250-no'}).text.split('.')[1]
+    film_name=soup.find(attrs={'property':'v:itemreviewed'}).text
+    director=soup.find(attrs={'id':'info'}).text.split('\n')[1].split(':')[1].strip()
+    actor=soup.find(attrs={'id':'info'}).text.split('\n')[3].split(':')[1].strip().split('/')
+    actor= str(actor)
+    actor = pymysql.converters.escape_string(actor)
+    language=soup.find(attrs={'id':'info'}).text.split('\n')[6].split(':')[1].strip()
+    rating_num = soup.find(attrs={'property':'v:average'}).text
+    summary = soup.find(attrs={'property': 'v:summary'}).text
+    summary=pymysql.converters.escape_string(summary)
+    sql = 'insert into movies (film_name,director,actor,language,ranks,rating_num,summary,links) values ("{}","{}","{}","{}","{}","{}","{}","{}")'.format(film_name,director,actor,language,ranks,rating_num,summary,film_url)
+    db = getDB()
+    cursor = db.cursor()
+    try:
+        cursor.execute(sql)
+        cursor.execute('insert into moviehash(movieid) values ("{}")'.format(id))
+        db.commit()
+    except Exception as e:
+        print(e)
+        db.rollback()
+    cursor.close()
+    db.close()
+
+if __name__ == '__main__':
+    print("开始抓取")
+    db=getDB()
+    cursor=db.cursor()
+    for i in range(0,50,25):
+        film_urls= get_url("https://movie.douban.com/top250?start="+str(i)+"&filter=")
+        for film_url in range(len(film_urls)):
+            id=re.search('\d\d+',film_urls[film_url]).group()
+            sql='select movieid from moviehash where movieid="{}";'.format(id)
+            cursor.execute(sql)
+            data=cursor.fetchall()
+            if not data:
+                get_url_info(film_urls[film_url],id)

flaskProject/get_music.py

@@ -0,0 +1,35 @@
+import pymysql
+import requests
+from lxml import etree
+def getDB():
+    db = pymysql.connect(host='localhost',user='root',password='123456',database='douban')
+    return db
+def Agent_info():
+    headers={
+        'User-Agent':'Mozilla/5.0 (Windows NT 10.0; Win64; x64) AppleWebKit/537.36 (KHTML, like Gecko) Chrome/123.0.0.0 Safari/537.36 Edg/123.0.0.0'
+    }
+    return headers
+def get_url_info(music_url):
+    print("抓取网址", music_url)
+    music_url = 'https://www.kugou.com/yy/html/rank.html'
+    headers = Agent_info()
+    res = requests.get(url=music_url, headers=headers)
+    e = etree.HTML(res.text)
+    music_singer =e.xpath('//div[@id="rankWrap"]//ul/li/a/@title')
+    music_urls = e.xpath('//div[@id="rankWrap"]//ul/li/a/@href')
+    for i in range(len(music_singer)):
+        sql = 'insert into music (name,info) values ("{}","{}")'.format(music_singer[i],music_urls[i])
+        db = getDB()
+        cursor = db.cursor()
+        try:
+            cursor.execute(sql)
+            db.commit()
+        except Exception as e:
+            print(e)
+            db.rollback()
+        result = cursor.fetchone()
+        cursor.close()
+        db.close()
+
+if __name__ == '__main__':
+    get_url_info(music_url= 'https://www.kugou.com/yy/html/rank.html')

flaskProject/get_nbadata.py

@@ -0,0 +1,38 @@
+import pymysql
+import requests
+from lxml import etree
+def getDB():
+    db = pymysql.connect(host='localhost',user='root',password='123456',database='douban')
+    return db
+def Agent_info():
+    headers={
+        'User-Agent':'Mozilla/5.0 (Windows NT 10.0; Win64; x64) AppleWebKit/537.36 (KHTML, like Gecko) Chrome/123.0.0.0 Safari/537.36 Edg/123.0.0.0'
+    }
+    return headers
+def get_url_info(nba_url):
+    print("抓取网址", nba_url)
+    headers = Agent_info()
+    res = requests.get(nba_url, headers=headers)
+    e = etree.HTML(res.text)
+    names = e.xpath('//table[@class="players_table"]//tr/td[2]/a/text()')
+    teams = e.xpath('//table[@class="players_table"]//tr/td[3]/a/text()')
+    nos = e.xpath('//table[@class="players_table"]//tr/td[1]/text()')
+    nos = nos[1::]
+    scores = e.xpath('//table[@class="players_table"]//tr/td[4]/text()')
+    scores = scores[1::]
+    infos = e.xpath('//table[@class="players_table"]//tr/td[2]/a/@href')
+    for i in range(len(names)):
+        sql = 'insert into nba (no,name,team,score,info) values ("{}","{}","{}","{}","{}")'.format(nos[i],names[i],teams[i],scores[i],infos[i])
+        db = getDB()
+        cursor = db.cursor()
+        try:
+            cursor.execute(sql)
+            db.commit()
+        except Exception as e:
+            print(e)
+            db.rollback()
+        result = cursor.fetchone()
+        cursor.close()
+        db.close()
+if __name__ == '__main__':
+    get_url_info(nba_url="https://nba.hupu.com/stats/players")

flaskProject/get_xiaoshuo.py

@@ -0,0 +1,62 @@
+import pymysql
+import requests
+from lxml import etree
+import re
+
+
+def getDB():
+    db = pymysql.connect(host='localhost',user='root',password='123456',database='douban')
+    return db
+
+def Agent_info():
+    headers = {
+        'User-Agent': 'Mozilla/5.0 (Windows NT 10.0; Win64; x64) AppleWebKit/537.36 (KHTML, like Gecko) Chrome/123.0.0.0 Safari/537.36'
+    }
+    return headers
+
+def get_url():
+    headers=Agent_info()
+    manh_url = 'https://www.qimao.com/paihang'
+    res = requests.get(url=manh_url, headers=headers)
+    htmldata=res.text
+    urls=re.findall('<a href="https://www.qimao.com/shuku/(\d+)/"',htmldata)
+    urls2=[]
+    for url in urls:
+        url="https://www.qimao.com/shuku/"+url+'/'
+        if url not in urls2:
+            urls2.append(url)
+    return urls2
+
+def get_url_info(manh_url):
+    print("抓取网址", manh_url)
+    headers = Agent_info()
+    res = requests.get(manh_url, headers=headers)
+    e = etree.HTML(res.text)
+    ranks = e.xpath('//*[@id="__layout"]/div/div[3]/div/div/div/div[1]/div/div[1]/div[2]/div[1]/span[2]/text()')
+    ranks=str(ranks)[1:-1]
+    names = e.xpath('//*[@id="__layout"]/div/div[3]/div/div/div/div[1]/div/div[1]/div[2]/div[1]/span[1]/text()')
+    names =str( names)[1:-1]
+    authors = e.xpath('//*[@id="__layout"]/div/div[3]/div/div/div/div[1]/div/div[1]/div[2]/div[3]/span[1]/em/a/text()')
+    authors =str(authors)[4:-4]
+    summarys = e.xpath('//*[@id="__layout"]/div/div[3]/div/div/div/div[2]/div/div[2]/div[1]/div/div[1]/div[2]/p/text()')
+    sql = 'insert into xiaoshuo (names,ranks,authors,summarys,urls) values ("{}","{}","{}","{}","{}")'.format(names,ranks,authors,summarys,manh_url)
+    db = getDB()
+    cursor = db.cursor()
+    try:
+        cursor.execute(sql)
+        db.commit()
+    except Exception as e:
+        print(e)
+        db.rollback()
+    cursor.close()
+    db.close()
+if __name__ == '__main__':
+    urls=get_url()
+    for url in urls:
+        get_url_info(url)
+
+
+
+
+
+

flaskProject/manh.py

@@ -0,0 +1,61 @@
+import pymysql
+import requests
+from lxml import etree
+
+
+def getDB():
+    db = pymysql.connect(host='localhost',user='root',password='123456',database='douban')
+    return db
+
+def Agent_info():
+    headers = {
+        'User-Agent': 'Mozilla/5.0 (Windows NT 10.0; Win64; x64) AppleWebKit/537.36 (KHTML, like Gecko) Chrome/123.0.0.0 Safari/537.36 Edg/123.0.0.0'
+    }
+    return headers
+
+def get_url():
+    headers=Agent_info()
+    manh_url = 'https://ac.qq.com/Rank/comicRank/type/top'
+    res = requests.get(url=manh_url, headers=headers)
+    e = etree.HTML(res.text)
+    hrefs = e.xpath('//div/ul/li/a/@href')
+    hrefs = hrefs[16:38]
+    for i in range(22):
+        hrefs[i] = "https://ac.qq.com/" + hrefs[i]
+    return hrefs
+
+def get_url_info(manh_url):
+    print("抓取网址", manh_url)
+    headers = Agent_info()
+    res = requests.get(manh_url, headers=headers)
+    e = etree.HTML(res.text)
+    ranks = e.xpath('//*[@id="special_bg"]/div[3]/div/div/div[2]/div/div[2]/p/strong/text()')
+    ranks=str(ranks).strip('[').strip(']')
+    names = e.xpath('//*[@id="special_bg"]/div[3]/div/div/div[2]/div/div[1]/h2/strong/text()')
+    names=str(names).strip('[').strip(']')
+    authors = e.xpath('//*[@id="special_bg"]/div[3]/div/div/div[2]/div/p/span/em/text()')
+    authors1 = e.xpath('//*[@id="special_bg"]/div[3]/div/div/div[2]/div/p/span/em/text()')
+    authors2 = e.xpath('//*[@id="special_bg"]/div[3]/div/div/div[2]/div/p/span/em/text()')
+
+    authors0 = str(authors).split(',')[0]
+    authors0 = authors0[2:-5]
+    renqi = str(authors1).split(',')[1]
+    shoucang = str(authors2).split(',')[2]
+    shoucang=shoucang[:-1]
+    print(authors)
+    sql = 'insert into manh (names,ranks,renqi,shoucang,urls,authors) values ("{}","{}","{}","{}","{}","{}")'.format(names,ranks,renqi,shoucang,manh_url,authors0)
+    db = getDB()
+    cursor = db.cursor()
+    try:
+        cursor.execute(sql)
+        db.commit()
+    except Exception as e:
+        print(e)
+        db.rollback()
+    cursor.close()
+    db.close()
+if __name__ == '__main__':
+
+    manh_urls = get_url()
+    for url in manh_urls:
+        get_url_info(url)

flaskProject/static/about.css

@@ -0,0 +1,28 @@
+body {
+    background-image: url(pic/about.jpg);
+    background-repeat:no-repeat ;
+    background-position:center;
+    background-size: cover;
+    background-attachment: fixed;
+}
+
+p{
+    text-align-all: center;
+    font-size: xx-large;
+    color: green;
+}
+
+.but{
+	width:300px;
+	min-height:40px;
+	display:block;
+	background-color:orange;
+	border:1px;
+	color:#fff;
+	padding:9px 14px;
+	font-size:15px;
+	line-height:normal;
+	border-radius:5px;
+	margin:20px;
+}
+

flaskProject/static/index.css

@@ -0,0 +1,107 @@
+body {
+    background-image: url(pic/b6.jpg);
+    background-repeat:no-repeat ;
+    background-position:center;
+    background-size: cover;
+    background-attachment: fixed;
+    font-family: Arial, Helvetica, sans-serif;
+    font-size: 12px;
+    margin: 0px auto;
+    height: 700px;
+    width: 800px;
+    border: 1px solid #00663300;
+}
+
+#navigator{
+				height:25px;
+				width:800px;
+				font-size:14px;
+				list-style-type:none;
+			}
+			#navigator li{
+				float:left;
+				width:190px;
+				height:25px;
+			}
+			#navigator li a{
+				color:#000000;
+				text-decoration:none;
+				padding-top:4px;
+				display:block;
+				width:131px;
+				height:22px;
+				text-align:center;
+				background-color:greenyellow;
+				margin-left:2px;
+			}
+			#navigator li a:hover{
+				background-color:#006633;
+				color:#FFFFFF;
+			}
+
+
+#content {
+    height:auto;
+    width:800px;
+}
+
+.but{
+	width:300px;
+	min-height:40px;
+	display:block;
+	background-color:orange;
+	border:1px;
+	color:#fff;
+	padding:9px 14px;
+	font-size:15px;
+	line-height:normal;
+	border-radius:5px;
+	margin:20px;
+}
+
+.button:hover{
+	background-color:#006633;
+	color:#FFFFFF;
+}
+
+.movie {
+  height: 200px;
+  background-image: url("pic/b1.jpg");
+  background-repeat: no-repeat;
+  background-size: cover;
+}
+
+.text {
+  height: 200px;
+  background-image: url("pic/b2.jpg");
+  background-repeat: no-repeat;
+  background-size: cover;
+}
+
+.music {
+  height: 200px;
+  background-image: url("pic/b3.jpg");
+  background-repeat: no-repeat;
+  background-size: cover;
+}
+
+.nba {
+  height: 200px;
+  background-image: url("pic/b4.jpg");
+  background-repeat: no-repeat;
+  background-size: cover;
+}
+
+.manh {
+  height: 200px;
+  background-image: url("pic/b5.jpg");
+  background-repeat: no-repeat;
+  background-size: cover;
+}
+
+.qm {
+  height: 200px;
+  background-image: url("pic/b6.jpg");
+  background-repeat: no-repeat;
+  background-size: cover;
+}

flaskProject/static/login.css

@@ -0,0 +1,68 @@
+@charset "UTF-8";
+html{
+	width:100%;
+	height:100%;
+	overflow:hidden;
+	font-style:sans-serif;
+}
+
+body{
+	width:100%;
+	height:100%;
+	font-family:'Open Sans',sans-serif;
+	margin:0;
+	background-color:black;
+}
+
+#login{
+	position:absolute;
+	top:50%;
+	left:50%;
+	margin:-150px 0 0 -150px;
+	width:300px;
+	height:300px;
+	background-color:gray;
+}
+
+#login h1{
+	color:pink;
+	text-shadow:0 0 10px;
+	letter-spacing:1px;
+	text-align:center;
+}
+
+h1{
+	font-size:2em;
+	margin:0.67em 0;
+}
+
+input{
+	width:278px;
+	height:18px;
+	margin-bottom:10px;
+	outline:none;
+	padding:10px;
+	font-size:13px;
+	color:#fff;
+	text-shadow:1px 1px 1px;
+	border-top:1px;
+	border-right:1px;
+	border-left:1px;
+	border-bottom:1px;
+	border-radius:4px;
+	background-color:green;
+}
+
+.but{
+	width:300px;
+	min-height:20px;
+	display:block;
+	background-color:orange;
+	border:1px;
+	color:#fff;
+	padding:9px 14px;
+	font-size:15px;
+	line-height:normal;
+	border-radius:5px;
+	margin:0;
+}

flaskProject/static/manh.css

@@ -0,0 +1,30 @@
+body {
+    background-image: url(pic/manh.jpg);
+    background-repeat:no-repeat ;
+    background-position:center;
+    background-size: cover;
+    background-attachment: fixed;
+}
+a{
+    text-decoration: none;
+    color: darkslateblue;
+}
+
+h1{
+    text-align: center;
+    color:darkorchid;
+}
+
+.but{
+	width:300px;
+	min-height:40px;
+	display:block;
+	background-color:orange;
+	border:1px;
+	color:#fff;
+	padding:9px 14px;
+	font-size:15px;
+	line-height:normal;
+	border-radius:5px;
+	margin:20px;
+}

flaskProject/static/movies.css

@@ -0,0 +1,27 @@
+body{
+    background-color: bisque;
+
+}
+a{
+    text-decoration: none;
+    color: darkslateblue;
+}
+
+h1{
+    text-align: center;
+    color:darkorchid;
+}
+
+.but{
+	width:300px;
+	min-height:40px;
+	display:block;
+	background-color:orange;
+	border:1px;
+	color:#fff;
+	padding:9px 14px;
+	font-size:15px;
+	line-height:normal;
+	border-radius:5px;
+	margin:20px;
+}

flaskProject/static/music.css

@@ -0,0 +1,43 @@
+
+a{
+        display:inline-block;
+        height: 29px;
+        float:right;
+        text-decoration: none;
+        margin:2px 5px
+}
+
+body {
+    background-image: url(pic/music.jpg);
+    background-repeat:no-repeat ;
+    background-position:center;
+    background-size: cover;
+    background-attachment: fixed;
+}
+
+h3{
+    color:red
+}
+a{
+    text-decoration: none;
+    color: darkslateblue;
+}
+
+h1{
+    text-align: center;
+    color:darkorchid;
+}
+
+.but{
+	width:300px;
+	min-height:40px;
+	display:block;
+	background-color:orange;
+	border:1px;
+	color:#fff;
+	padding:9px 14px;
+	font-size:15px;
+	line-height:normal;
+	border-radius:5px;
+	margin:20px;
+}

flaskProject/static/nba.css

@@ -0,0 +1,38 @@
+body {
+    background-image: url(pic/nba.jpg);
+    background-repeat:no-repeat ;
+    background-position:center;
+    background-size: cover;
+    background-attachment: fixed;
+}
+
+table{
+    border:1;
+    border-color: mediumvioletred;
+
+}
+
+a{
+    text-decoration: none;
+    color: darkslateblue;
+}
+
+
+h1{
+    text-align: center;
+    color:darkorchid;
+}
+
+.but{
+	width:300px;
+	min-height:40px;
+	display:block;
+	background-color:orange;
+	border:1px;
+	color:#fff;
+	padding:9px 14px;
+	font-size:15px;
+	line-height:normal;
+	border-radius:5px;
+	margin:20px;
+}

flaskProject/static/regist.css

@@ -0,0 +1,12 @@
+body {
+    background-image: url(pic/regist.jpg);
+    background-repeat:no-repeat ;
+    background-position:center;
+    background-size: cover;
+    background-attachment: fixed;
+
+}
+
+h1{
+    text-align: center;
+}

flaskProject/static/text.css

@@ -0,0 +1,30 @@
+body {
+    background-image: url(pic/text.jpg);
+    background-repeat:no-repeat ;
+    background-position:center;
+    background-size: cover;
+    background-attachment: fixed;
+}
+a{
+    text-decoration: none;
+    color: darkslateblue;
+}
+
+h1{
+    text-align: center;
+    color:darkorchid;
+}
+
+.but{
+	width:300px;
+	min-height:40px;
+	display:block;
+	background-color:orange;
+	border:1px;
+	color:#fff;
+	padding:9px 14px;
+	font-size:15px;
+	line-height:normal;
+	border-radius:5px;
+	margin:20px;
+}

flaskProject/templates/about.html

@@ -0,0 +1,17 @@
+<!DOCTYPE html>
+<html lang="en">
+<head>
+    <meta charset="UTF-8">
+    <title>Title</title>
+    <link rel="stylesheet" type="text/css" href="static/about.css">
+</head>
+<body>
+    <div class="btn">
+    <form method="get" action="/index" >
+        <button class="but" type="submit" >回到主页</button>
+    </form>
+        <button class="but" type="submit">联系我们</button>
+    </div>
+    <p>&nbsp &nbsp&nbsp &nbsp  我们网站提供关于爬虫的相关信息和资源，旨在帮助用户了解爬虫技术并学习如何使用它们。爬虫是一种自动的程序，用于在互联网上收集和提取数据。它可以模拟人类的浏览行为，访问网页并抓取所需的信息。虫技术在数据分析、市场研究、争情报等领域具有广泛应用。在我们的网站上，用户可以找到有关爬虫的教程、指南实例代码。我们提供入门级别的教程，帮助初学者了解爬虫的基本原理和操作步骤。我们还提供高级教程，介绍更复杂的爬虫技术和应用场景。此外，我们还提供爬虫工具和架的推和评估。用户可以了解不同的爬虫工具，选择适合己需求的工具来开展爬虫项目。我们还提供爬虫实战案例和经验分享，让用户可以学习和借鉴成功的爬虫项目。我们的目标是帮助掌握爬虫技术，并将其应用于实际项目中。我们鼓励用户遵守法律法规和道德准则，在使用爬虫技术时保护个人隐私和网络安全。感谢您选择我们的网站，我们期待为您提供有关爬虫的知识和资源。如有任何问题或需求，请随时联系我们。</p>
+</body>
+</html>

flaskProject/templates/index.html

@@ -0,0 +1,54 @@
+<!DOCTYPE html>
+<html lang="en">
+
+<head>
+    <title>index</title>
+    <link rel="stylesheet" type="text/css" href="static/index.css">
+</head>
+
+<body>
+    <div id="content" >
+        <embed src="static/pic/bgmusic.ogg" preload="auto" autostart="true" loop="true" hidden="true">
+        <div id="navigator">
+			<ul id="navigator">
+				<li><a href="/about">关于</a></li>
+				<li><a href="/zhanghao">切换账号</a></li>
+				<li><a href="/talk">论坛</a></li>
+				<li><a href="#">帮助</a></li>
+			</ul>
+		</div>
+        <div class="movie">
+        <form method="get" action="/movies">
+            <button class="but" type="submit">点我看电影</button>
+        </form>
+
+        </div>
+        <div class="nba">
+        <form method="get" action="/get_nbadata">
+            <button class="but" type="submit">点我看NBA资讯</button>
+        </form>
+        </div>
+        <div class="music">
+        <form method="get" action="/music">
+            <button class="but" type="submit">点我听音乐</button>
+        </form>
+        </div>
+        <div class="text">
+        <form method="get" action="/get_text">
+            <button class="but" type="submit">点我看小说</button>
+        </form>
+        </div>
+        <div class="manh">
+        <form method="get" action="/manh">
+            <button class="but" type="submit">点我看漫画</button>
+        </form>
+        </div>
+        <div class="qm">
+        <form method="get" action="/qm">
+            <button class="but" type="submit">点我看七猫小说</button>
+        </form>
+        </div>
+    </div>
+</body>
+
+</html>

flaskProject/templates/login.html

@@ -0,0 +1,18 @@
+<!DOCTYPE html>
+<html>
+<head>
+<meta charset="utf-8">
+<title>登录界面</title>
+<link rel="stylesheet" type="text/css" href="static/login.css">
+</head>
+<body>
+	<div id="login">
+	<h1>阿肖的爬虫项目</h1>
+	<form method="get" action="/index1">
+		<input type="text" required="required" placeholder="用户名" name="u"></input>
+		<input type="password" required="required" placeholder="密码" name="p"></input>
+		<button class="but" type="submit">登录</button>
+	</form>
+	</div>
+</body>
+</html>

flaskProject/templates/manh.html

@@ -0,0 +1,38 @@
+<!DOCTYPE html>
+<html lang="en">
+<head>
+    <meta charset="UTF-8">
+    <title>Title</title>
+    <link rel="stylesheet" type="text/css" href="static/manh.css">
+</head>
+<body>
+    <div class="row">
+         <div class="col">
+             <h1 class="section-title">腾讯漫画</h1>
+         </div>
+        <form method="get" action="/index">
+        <button class="but" type="submit">回到主页</button>
+        </form>
+     </div>
+    <table class="table" border="1">
+        <tr>
+            <td >评分</td>
+            <td style="width:20%;">名称</td>
+            <td style="width:20%;">人气</td>
+            <td style="width:20%;">收藏</td>
+            <td style="width:20%;">作者</td>
+            <td >地址</td>
+        </tr>
+        {% for manh in  datalist %}
+            <tr>
+                <td>{{ manh[1] }}</td>
+                <td>{{ manh[0] }}</td>
+                <td>{{ manh[2] }}</td>
+                <td>{{ manh[3] }}</td>
+                <td>{{ manh[4] }}</td>
+                <td><a href="{{ manh[5] }}" target="_blank">去腾讯动漫看</a></td>
+            </tr>
+        {% endfor %}
+    </table>
+</body>
+</html>

flaskProject/templates/movies.html

@@ -0,0 +1,43 @@
+<!DOCTYPE html>
+<html lang="en">
+<head>
+    <meta charset="UTF-8">
+    <title>Title</title>
+    <link rel="stylesheet" type="text/css" href="static/movies.css">
+</head>
+<body>
+     <div class="row">
+         <div class="col">
+             <h1 class="section-title">豆瓣电影top50排行榜</h1>
+         </div>
+         <form method="get" action="/index">
+        <button class="but" type="submit">回到主页</button>
+        </form>
+     </div>
+    <table class="table" border="1">
+        <tr>
+            <td >排名</td>
+            <td >评分</td>
+            <td style="width:5%;">电影中文名称</td>
+            <td style="width:5%;">导演</td>
+            <td style="width:5%;">语言</td>
+            <td style="width: 30%;">演员</td>
+            <td style="width: 50%;">简介</td>
+            <td style="width:5%;">播放地址</td>
+        </tr>
+        {% for movie in movies %}
+            <tr>
+                <td>{{ movie[4] }}</td>
+                <td>{{ movie[5] }}</td>
+                <td>{{ movie[0] }}</td>
+                <td>{{ movie[1] }}</td>
+                <td>{{ movie[3] }}</td>
+                <td>{{ movie[2] }}</td>
+                <td>{{ movie[6] }}</td>
+                <td><a href="{{ movie[7] }}" target="_blank">去豆瓣看</a></td>
+            </tr>
+        {% endfor %}
+    </table>
+    <embed src="static/pic/bgmusic.mp3" preload="auto" autostart="true" loop="true" hidden="true">
+</body>
+</html>

flaskProject/templates/music.html

@@ -0,0 +1,30 @@
+<!DOCTYPE html>
+<html lang="en">
+<head>
+    <meta charset="UTF-8">
+    <title>Title</title>
+    <link rel="stylesheet" type="text/css" href="static/music.css">
+</head>
+<body>
+     <div class="row">
+         <div class="col">
+             <h1 class="section-title">酷狗音乐排行榜</h1>
+         </div>
+         <form method="get" action="/index">
+        <button class="but" type="submit">回到主页</button>
+        </form>
+     </div>
+    <table class="table" border="1">
+        <tr>
+            <td bgcolor="#f8f8ff">歌名</td>
+            <td style="width:20%;">播放地址</td>
+        </tr>
+        {% for music in datalist %}
+            <tr>
+                <td bgcolor="#d2691e">{{ music[0] }}</td>
+                <td bgcolor="#ff4500"><a href="{{ music[1] }}" target="_blank">播放</a></td>
+            </tr>
+        {% endfor %}
+    </table>
+</body>
+</html>

flaskProject/templates/nba.html

@@ -0,0 +1,37 @@
+<!DOCTYPE html>
+<html lang="en">
+<head>
+    <meta charset="UTF-8">
+    <title>Title</title>
+    <link rel="stylesheet" type="text/css" href="static/nba.css">
+</head>
+<body>
+     <div class="row">
+         <div class="col">
+             <h1 class="section-title">NBA球员top50排行榜</h1>
+         </div>
+         <form method="get" action="/index">
+        <button class="but" type="submit">回到主页</button>
+        </form>
+     </div>
+    <table class="table" border="1">
+        <tr>
+            <td bgcolor="#ff4500">排名</td>
+            <td style="width:20%;" bgcolor="#ff4500">姓名</td>
+            <td style="width:20%;" bgcolor="#ff4500">得分</td>
+            <td style="width:20%;" bgcolor="#ff4500">队伍</td>
+            <td style="width:20%;" bgcolor="#ff4500">球员信息</td>
+
+        </tr>
+        {% for nba in nbadata %}
+            <tr>
+                <td>{{ nba[0] }}</td>
+                <td bgcolor="#f8f8ff">{{ nba[1] }}</td>
+                <td bgcolor="#f8f8ff">{{ nba[2] }}</td>
+                <td bgcolor="#f8f8ff">{{ nba[3] }}</td>
+                <td bgcolor="#f8f8ff"><a href="{{ nba[4] }}" target="_blank" >去虎扑网看</a></td>
+            </tr>
+        {% endfor %}
+    </table>
+</body>
+</html>

flaskProject/templates/qm.html

@@ -0,0 +1,36 @@
+<!DOCTYPE html>
+<html lang="en">
+<head>
+    <meta charset="UTF-8">
+    <title>Title</title>
+    <link rel="stylesheet" type="text/css" href="static/text.css">
+</head>
+<body>
+    <div class="row">
+         <div class="col">
+             <h1 class="section-title">七猫小说</h1>
+         </div>
+        <form method="get" action="/index">
+        <button class="but" type="submit">回到主页</button>
+        </form>
+     </div>
+    <table class="table" border="1">
+        <tr>
+            <td >评分</td>
+            <td style="width:20%;">名称</td>
+            <td style="width:20%;">作者</td>
+            <td style="width:20%;">简介</td>
+            <td >地址</td>
+        </tr>
+        {% for qm in  datalist %}
+            <tr>
+                <td>{{ qm[1] }}</td>
+                <td>{{ qm[0] }}</td>
+                <td>{{ qm[2] }}</td>
+                <td>{{ qm[3] }}</td>
+                <td><a href="{{ qm[4] }}" target="_blank">去七猫小说看</a></td>
+            </tr>
+        {% endfor %}
+    </table>
+</body>
+</html>

flaskProject/templates/regist.html

@@ -0,0 +1,36 @@
+<!DOCTYPE html>
+<html lang="en">
+<head>
+    <meta charset="UTF-8">
+    <meta name="viewport" content="width=device-width, initial-scale=1.0">
+    <title>Document</title>
+    <link rel="stylesheet" type="text/css" href="static/regist.css">
+    <style>
+        .but{
+	        width:300px;
+	        min-height:50px;
+	        display:block;
+	        background-color:orange;
+	        border:1px;
+	        color:#fff;
+	        font-size:15px;
+	        line-height:normal;
+	        border-radius:5px;
+	        margin-left:auto;
+            margin-right: auto;
+        }
+        h1{
+            color: deepskyblue;
+	        font-size:2em;
+	        margin:0.67em 0;
+        }
+
+    </style>
+</head>
+<body>
+    <form method="get" action="/login">
+        <h1>输入有误，请重新登录</h1>
+        <button class="but" type="submit">重新登录</button>
+    </form>
+</body>
+</html>

flaskProject/templates/talk.html

@@ -0,0 +1,56 @@
+<!DOCTYPE html>
+    <head>
+        <meta charset = "utf-8" />
+        <style>
+            body{
+                display: flex;
+                flex-direction: column;
+                align-items: center;
+                background-color: rgb(117, 113, 113);
+                color: white;
+            }
+            #main{
+                flex: 1;
+                width: 60%;
+            }
+            #typing{
+                flex: 1;
+                width: 60%;
+                height: 20vh;
+                border: red;
+                font-size:xx-large;
+            }
+            #commend{
+                flex: 1;
+                width: 80%;
+            }
+            #btn{
+                background-color: red;
+            }
+            p{
+                background-color:black;
+            }
+        </style>
+
+    </head>
+
+    <body>
+        <div id="main">
+            <h1>评论区</h1><hr>
+            <textarea id="typing">输入文字</textarea>
+            <button id="btn">发送评论</button>
+            <div id="commend"></div>
+        </div>
+    </body>
+
+    <script>
+        document.getElementById("btn").onclick=function(){send()};
+        function send(){
+            var g=new Date();
+            var value= document.getElementById("typing").value;
+            var p= document.createElement("p");
+            p.innerHTML="用户："+value+'<hr>'+g.toUTCString();
+            document.getElementById("commend").prepend(p);
+        }
+    </script>
+</html>

flaskProject/templates/text.html

@@ -0,0 +1,40 @@
+<!DOCTYPE html>
+<html lang="en">
+<head>
+    <meta charset="UTF-8">
+    <title>Title</title>
+    <link rel="stylesheet" type="text/css" href="static/text.css">
+</head>
+<body>
+    <div class="row">
+         <div class="col">
+             <h1 class="section-title">笔趣阁小说</h1>
+                 <form method="get" action="/index">
+                    <button class="but" type="submit" >回到主页</button>
+                 </form>
+         </div>
+
+     </div>
+    <table id="table1" style="float:left" border="1" bgcolor="#3cb371">
+        {% for name in text_names %}
+            <tr>
+                <td width="500px">{{ name }}</td>
+            </tr>
+        {% endfor %}
+     </table>
+    <table id="table3" style="float:left" border="1" bgcolor="#3cb371">
+        {% for author in text_authors2 %}
+            <tr>
+                <td width="500px">{{ author }}</td>
+            </tr>
+        {% endfor %}
+     </table>
+     <table id="table2" border="1" bgcolor="#faebd7">
+        {% for url in text_urls %}
+            <tr>
+                <td><a href="https://www.biqg.cc{{url}}">看小说</a></td>
+            </tr>
+        {% endfor %}
+     </table>
+</body>
+</html>

flaskProject/test.py

@@ -0,0 +1,26 @@
+import pymysql
+import requests
+from lxml import etree
+import re
+
+def get_news():
+    headers= {
+        'User-Agent': 'Mozilla/5.0 (Windows NT 10.0; Win64; x64) AppleWebKit/537.36 (KHTML, like Gecko) Chrome/123.0.0.0 Safari/537.36'
+    }
+    url = 'https://news.sina.com.cn/hotnews/#1'
+    res = requests.get(url=url,headers=headers)
+    e = res.json()
+
+    return e
+
+
+
+if __name__ == '__main__':
+    urls=get_news()
+    print(urls)
+
+
+
+
+
+