diff --git a/国内疫情信息.png b/国内疫情信息.png
new file mode 100644
index 0000000..710c80b
Binary files /dev/null and b/国内疫情信息.png differ
diff --git a/国外疫情.png b/国外疫情.png
new file mode 100644
index 0000000..0ecc738
Binary files /dev/null and b/国外疫情.png differ
diff --git a/国外疫情概况.png b/国外疫情概况.png
new file mode 100644
index 0000000..088fadd
Binary files /dev/null and b/国外疫情概况.png differ
diff --git a/数据采集/丁香园数据/.idea/.gitignore b/数据采集/丁香园数据/.idea/.gitignore
new file mode 100644
index 0000000..e7e9d11
--- /dev/null
+++ b/数据采集/丁香园数据/.idea/.gitignore
@@ -0,0 +1,2 @@
+# Default ignored files
+/workspace.xml
diff --git a/数据采集/丁香园数据/.idea/inspectionProfiles/profiles_settings.xml b/数据采集/丁香园数据/.idea/inspectionProfiles/profiles_settings.xml
new file mode 100644
index 0000000..105ce2d
--- /dev/null
+++ b/数据采集/丁香园数据/.idea/inspectionProfiles/profiles_settings.xml
@@ -0,0 +1,6 @@
+
+
+
+
+
+
\ No newline at end of file
diff --git a/数据采集/丁香园数据/.idea/misc.xml b/数据采集/丁香园数据/.idea/misc.xml
new file mode 100644
index 0000000..65531ca
--- /dev/null
+++ b/数据采集/丁香园数据/.idea/misc.xml
@@ -0,0 +1,4 @@
+
+
+
+
\ No newline at end of file
diff --git a/数据采集/丁香园数据/.idea/modules.xml b/数据采集/丁香园数据/.idea/modules.xml
new file mode 100644
index 0000000..60f1283
--- /dev/null
+++ b/数据采集/丁香园数据/.idea/modules.xml
@@ -0,0 +1,8 @@
+
+
+
+
+
+
+
+
\ No newline at end of file
diff --git a/数据采集/丁香园数据/.idea/vcs.xml b/数据采集/丁香园数据/.idea/vcs.xml
new file mode 100644
index 0000000..b2bdec2
--- /dev/null
+++ b/数据采集/丁香园数据/.idea/vcs.xml
@@ -0,0 +1,6 @@
+
+
+
+
+
+
\ No newline at end of file
diff --git a/数据采集/丁香园数据/.idea/丁香园数据.iml b/数据采集/丁香园数据/.idea/丁香园数据.iml
new file mode 100644
index 0000000..a193443
--- /dev/null
+++ b/数据采集/丁香园数据/.idea/丁香园数据.iml
@@ -0,0 +1,11 @@
+
+
+
+
+
+
+
+
+
+
+
\ No newline at end of file
diff --git a/数据采集/GrapSpider.py b/数据采集/丁香园数据/GrapSpider.py
similarity index 100%
rename from 数据采集/GrapSpider.py
rename to 数据采集/丁香园数据/GrapSpider.py
diff --git a/数据采集/grapdata.py b/数据采集/丁香园数据/grapdata.py
similarity index 100%
rename from 数据采集/grapdata.py
rename to 数据采集/丁香园数据/grapdata.py
diff --git a/数据采集/text.txt b/数据采集/丁香园数据/text.txt
similarity index 100%
rename from 数据采集/text.txt
rename to 数据采集/丁香园数据/text.txt
diff --git a/数据采集/百度数据/.idea/inspectionProfiles/profiles_settings.xml b/数据采集/百度数据/.idea/inspectionProfiles/profiles_settings.xml
new file mode 100644
index 0000000..105ce2d
--- /dev/null
+++ b/数据采集/百度数据/.idea/inspectionProfiles/profiles_settings.xml
@@ -0,0 +1,6 @@
+
+
+
+
+
+
\ No newline at end of file
diff --git a/数据采集/百度数据/.idea/misc.xml b/数据采集/百度数据/.idea/misc.xml
new file mode 100644
index 0000000..65531ca
--- /dev/null
+++ b/数据采集/百度数据/.idea/misc.xml
@@ -0,0 +1,4 @@
+
+
+
+
\ No newline at end of file
diff --git a/数据采集/百度数据/.idea/modules.xml b/数据采集/百度数据/.idea/modules.xml
new file mode 100644
index 0000000..a974ff4
--- /dev/null
+++ b/数据采集/百度数据/.idea/modules.xml
@@ -0,0 +1,8 @@
+
+
+
+
+
+
+
+
\ No newline at end of file
diff --git a/数据采集/百度数据/.idea/vcs.xml b/数据采集/百度数据/.idea/vcs.xml
new file mode 100644
index 0000000..b2bdec2
--- /dev/null
+++ b/数据采集/百度数据/.idea/vcs.xml
@@ -0,0 +1,6 @@
+
+
+
+
+
+
\ No newline at end of file
diff --git a/数据采集/百度数据/.idea/workspace.xml b/数据采集/百度数据/.idea/workspace.xml
new file mode 100644
index 0000000..94bb5dd
--- /dev/null
+++ b/数据采集/百度数据/.idea/workspace.xml
@@ -0,0 +1,111 @@
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+ 1596505436061
+
+
+ 1596505436061
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
\ No newline at end of file
diff --git a/数据采集/百度数据/.idea/百度数据.iml b/数据采集/百度数据/.idea/百度数据.iml
new file mode 100644
index 0000000..9710855
--- /dev/null
+++ b/数据采集/百度数据/.idea/百度数据.iml
@@ -0,0 +1,8 @@
+
+
+
+
+
+
+
+
\ No newline at end of file
diff --git a/数据采集/百度数据/DataFrame.py b/数据采集/百度数据/DataFrame.py
new file mode 100644
index 0000000..bd66254
--- /dev/null
+++ b/数据采集/百度数据/DataFrame.py
@@ -0,0 +1,99 @@
+
+
+#城市疫情信息
+class CityInfo():
+ def __init__(self):
+ self.city='' #城市
+ self.confirmed=0 #累计
+ self.died=0 #死亡
+ self.cured=0 #治愈
+ self.confirmedRelative=0 #新增
+ self.curConfirm=0 #现有
+
+ def __str__(self):
+ return "%s 累计：%d ，新增：%d ，治愈：%d ，现有：%d ，死亡：%d"%(self.city,self.confirmed,self.confirmedRelative,
+ self.cured,self.curConfirm,self.died)
+
+#省疫情信息
+class Porvince():
+ def __init__(self):
+ self.confirmed = 0,
+ self.died = 0,
+ self.crued = 0,
+ self.relativeTime = 0,
+ self.confirmedRelative = 0,
+ self.diedRelative = 0,
+ self.curedRelative = 0,
+ self.asymptomaticRelative = 0,
+ self.asymptomatic = 0,
+ self.curConfirm = 0,
+ self.curConfirmRelative = 0,
+ self.icuDisable = 0,
+ self.area = '',
+ self.pub_date = '',
+ self.subList = [],
+
+ def __str__(self):
+ return "%s 现存确诊%d, 累计确诊%d, 死亡%d, 治愈%d,更新时间%s"%(self.area,self.curConfirm,self.confirmed,self.died,self.crued,self.pub_date)
+
+#国内疫情概况
+class InsideChina():
+ def __init__(self):
+ self.curConfirm = 0,
+ self.curConfirmRelative = 0,
+ self.asymptomatic = 0,
+ self.asymptomaticRelative = 0,
+ self.unconfirmed = 0,
+ self.unconfirmedRelative = 0,
+ self.icu = 0,
+ self.icuRelative = 0,
+ self.confirmed = 0,
+ self.confirmedRelative = 0,
+ self.overseasInput = 0,
+ self.overseasInputRelative = 0,
+ self.cured = 0,
+ self.curedRelative = 0,
+ self.died = 0,
+ self.diedRelative = 0,
+ self.updatedTime = 0,
+
+ def __str__(self):
+ return "国内现有确诊%d,较昨日增加%d,无症状%d，较昨日增加%d,\n现有疑似%d，较昨日增加%d，现有重症%d，" \
+ "较昨日增加%d，\n累计确诊%d，较昨日增加%d,境外输入%d，较昨日增加%d，\n累计治愈%d，" \
+ "较昨日增加%d，累计死亡%d，较昨日增加%d"%(self.curConfirm,self.curConfirmRelative,
+ self.asymptomatic,self.asymptomaticRelative,self.unconfirmed,
+ self.unconfirmedRelative,self.icu,self.icuRelative,self.confirmed,
+ self.confirmedRelative,self.overseasInput,self.overseasInputRelative,
+ self.cured,self.curedRelative,self.died,self.diedRelative)
+
+
+#国外疫情概况
+class OutsideChina():
+ def __init__(self):
+ self.confirmed = 0,
+ self.curConfirm = 0,
+ self.confirmedRelative = 0,
+ self.cured = 0,
+ self.curedRelative = 0,
+ self.died = 0,
+ self.diedRelative = 0,
+ self.updatedTime = 0,
+
+ def __str__(self):
+ return "截止%s,国外累计确诊%d，现存确诊%d，昨日新增%d，治愈%d，昨日治愈%d，累计死亡%d，昨日死亡%d"%(self.updatedTime,self.confirmed,self.curConfirm,
+ self.confirmedRelative,self.cured,self.curedRelative,self.died,self.diedRelative)
+
+
+#国外国家疫情信息
+class ForeignCountry():
+ def __init__(self):
+ self.died = 0,
+ self.confirmed = 0,
+ self.crued = 0,
+ self.country = '',
+ self.curConfirm = 0,
+ self.confirmedRelative = 0,
+ self.pub_date = '',
+
+ def __str__(self):
+ return "%s, 累计确诊%d，现存确诊%d，累计死亡%d，累计治愈%d，昨日新增%d"%(self.country,self.confirmed,self.curConfirm,self.died,self.crued,self.confirmedRelative)
\ No newline at end of file
diff --git a/数据采集/百度数据/GrapData.py b/数据采集/百度数据/GrapData.py
new file mode 100644
index 0000000..adf2701
--- /dev/null
+++ b/数据采集/百度数据/GrapData.py
@@ -0,0 +1,185 @@
+import time
+
+import requests
+import re
+import json
+from bs4 import BeautifulSoup
+from DataFrame import *
+
+class Spider():
+ def __init__(self):
+ self.url=''
+ self.headers={"user-agent":"Mozilla/5.0 (X11; Linux x86_64) AppleWebKit/537.36 (KHTML, like Gecko) Chrome/78.0.3904.108 Safari/537.36"}
+ self.updateTime=''
+
+ #抓取国内各省市的信息
+ def grapProvince(self):
+ data=requests.get(self.url,headers=self.headers)
+ data.encoding='utf-8'
+ html=data.text
+ #构建soup对象
+ soup=BeautifulSoup(html,'html.parser')
+ #查找目标标签
+ tag=soup.find('script', attrs={'id':'captain-config'})
+ tagstr=tag.string
+ #提取出各省疫情信息
+ caselist=re.findall(r'"caseList":\[(\{.*?\})\],"dataSource"',tagstr)
+ #转换成json对象
+ provincesinfo=re.findall(r'({"confirmed".*?\]})',caselist[0])
+ #保存所有省疫情信息
+ provinces=[]
+ for i in range(34):
+ pro=json.loads(provincesinfo[i])
+ province=Porvince()
+ province.area=pro['area']
+ province.confirmedRelative=int(pro['confirmedRelative'])
+ province.confirmed=int(pro['confirmed'])
+ province.crued=int(pro['crued'])
+ province.died=int(pro['died'])
+ province.curConfirmRelative=int(pro['curConfirmRelative'])
+ province.curConfirm=int(pro['curConfirm'])
+ province.diedRelative=int(pro['diedRelative'])
+ province.curedRelative=int(pro['curedRelative'])
+ province.asymptomatic=self.dealData(pro,'asymptomatic')
+ province.asymptomaticRelative=self.dealData(pro,'asymptomaticRelative')
+ province.icuDisable=int(pro['icuDisable'])
+ province.relativeTime=self.changeTime(pro['relativeTime'])
+ province.pub_date=time.strftime("%Y-%m-%d %H:%M:%S", time.localtime(time.time()))
+ province.subList=[]
+ #处理城市信息，并加入到该省的信息中
+ citysinfo=pro['subList']
+ cities=re.findall(r'(\{.*?\})',str(citysinfo))
+ print(province)
+ for city in cities:
+ city=str(city).replace('\'',"\"")
+ cj=json.loads(city)
+ citysinfo=CityInfo()
+ citysinfo.city=cj['city']
+ citysinfo.confirmed=int(cj['confirmed'])
+ #需要对缺失数据进行处理
+ if cj['died']=='':
+ citysinfo.died=0
+ else:
+ citysinfo.died=int(cj['died'])
+ if cj['crued']=='':
+ citysinfo.cured=0
+ else:
+ citysinfo.cured=int(cj['crued'])
+ citysinfo.confirmedRelative=int(cj['confirmedRelative'])
+ try:
+ if cj['curConfirm']=='':
+ citysinfo.curConfirm=0
+ else:
+ citysinfo.curConfirm=int(cj['curConfirm'])
+ except:
+ citysinfo.curConfirm = 0
+ province.subList.append(citysinfo)
+ print(citysinfo)
+ print('-----------------------------------------------------------')
+
+ #j=json.loads(caselist[0])
+
+ #抓取国外疫情信息
+ def grapForeign(self):
+ data=requests.get(self.url,headers=self.headers)
+ data.encoding='utf-8'
+ html=data.text
+ #构建soup对象
+ soup=BeautifulSoup(html,'html.parser')
+ #查找目标标签
+ tag=soup.find('script', attrs={'id':'captain-config'})
+ tagstr=tag.string
+ globallist=re.findall(r'"globalList":\[(.*?)\],"allForeignTrend"',tagstr)
+ countries=re.findall(r'({"died":.*?"country".*?\})',str(globallist[0]))
+ foreigns=[]
+ for country in countries:
+ coun=json.loads(country)
+ foreign=ForeignCountry()
+ foreign.died=int(coun['died'])
+ foreign.confirmed=int(coun['confirmed'])
+ foreign.crued=int(coun['crued'])
+ foreign.country=coun['country']
+ foreign.curConfirm=int(coun['curConfirm'])
+ foreign.confirmedRelative=int(coun['confirmedRelative'])
+ foreign.pub_date=time.strftime("%Y-%m-%d %H:%M:%S", time.localtime(time.time()))
+ foreigns.append(foreign)
+ print(foreign)
+
+
+
+ #抓取国外疫情信息概况
+ def grapSummaryForeign(self):
+ data=requests.get(self.url,headers=self.headers)
+ data.encoding='utf-8'
+ html=data.text
+ #构建soup对象
+ soup=BeautifulSoup(html,'html.parser')
+ #查找目标标签
+ tag=soup.find('script', attrs={'id':'captain-config'})
+ tagstr=tag.string
+ #国外疫情信息
+ summaryDataIn=re.findall(r'"summaryDataOut":(\{.*?\})',tagstr)
+ #转换为json
+ DataOut=json.loads(summaryDataIn[0])
+ Outside=OutsideChina()
+ Outside.confirmed=int(DataOut['confirmed'])
+ Outside.curConfirm=int(DataOut['curConfirm'])
+ Outside.confirmedRelative=int(DataOut['confirmedRelative'])
+ Outside.cured=int(DataOut['cured'])
+ Outside.curedRelative=int(DataOut['curedRelative'])
+ Outside.died=int(DataOut['died'])
+ Outside.diedRelative=int(DataOut['diedRelative'])
+ Outside.updatedTime=time.strftime("%Y-%m-%d %H:%M:%S", time.localtime(time.time()))
+ print(Outside)
+
+
+ #抓取国内疫情信息概况
+ def grapSummaryChina(self):
+ data=requests.get(self.url,headers=self.headers)
+ data.encoding='utf-8'
+ html=data.text
+ #构建soup对象
+ soup=BeautifulSoup(html,'html.parser')
+ #查找目标标签
+ tag=soup.find('script', attrs={'id':'captain-config'})
+ tagstr=tag.string
+ #国内疫情信息
+ summaryDataIn=re.findall(r'"summaryDataIn":(\{.*?\})',tagstr)
+ #转换为json
+ DataIn=json.loads(summaryDataIn[0])
+ Inside=InsideChina()
+ Inside.curConfirm=int(DataIn['confirmed'])
+ Inside.died=int(DataIn['died'])
+ Inside.cured=int(DataIn['cured'])
+ Inside.asymptomatic=int(DataIn['asymptomatic'])
+ Inside.asymptomaticRelative=int(DataIn['asymptomaticRelative'])
+ Inside.unconfirmed=int(DataIn['unconfirmed'])
+ Inside.unconfirmedRelative=int(DataIn['unconfirmedRelative'])
+ Inside.confirmed=int(DataIn['confirmed'])
+ Inside.confirmedRelative=int(DataIn['confirmedRelative'])
+ Inside.curedRelative=int(DataIn['curedRelative'])
+ Inside.diedRelative=int(DataIn['diedRelative'])
+ Inside.icu=int(DataIn['icu'])
+ Inside.icuRelative=int(DataIn['icuRelative'])
+ Inside.overseasInput=int(DataIn['overseasInput'])
+ Inside.overseasInputRelative=int(DataIn['overseasInputRelative'])
+ Inside.curConfirmRelative=int(DataIn['curConfirmRelative'])
+ Inside.updatedTime=time.strftime("%Y-%m-%d %H:%M:%S", time.localtime(time.time()))
+ print(Inside)
+
+
+ def dealData(self,a,b):
+ try:
+ if a[b]=='':
+ return 0
+ else:
+ return a[b]
+ except:
+ return 0
+
+ #转换时间格式
+ def changeTime(self,t):
+ localt=time.localtime(float(t)/1000)
+ timestr = time.strftime("%Y-%m-%d %H:%M:%S", localt)
+ return timestr
+
diff --git a/数据采集/百度数据/Grapmain.py b/数据采集/百度数据/Grapmain.py
new file mode 100644
index 0000000..fc9de1a
--- /dev/null
+++ b/数据采集/百度数据/Grapmain.py
@@ -0,0 +1,9 @@
+from GrapData import Spider
+
+def main():
+ spider=Spider()
+ spider.url='https://voice.baidu.com/act/newpneumonia/newpneumonia/'
+ spider.grapForeign()
+
+if __name__ == '__main__':
+ main()
\ No newline at end of file
diff --git a/数据采集/百度数据/__pycache__/DataFrame.cpython-36.pyc b/数据采集/百度数据/__pycache__/DataFrame.cpython-36.pyc
new file mode 100644
index 0000000..c99e70e
Binary files /dev/null and b/数据采集/百度数据/__pycache__/DataFrame.cpython-36.pyc differ
diff --git a/数据采集/百度数据/__pycache__/GrapData.cpython-36.pyc b/数据采集/百度数据/__pycache__/GrapData.cpython-36.pyc
new file mode 100644
index 0000000..319208c
Binary files /dev/null and b/数据采集/百度数据/__pycache__/GrapData.cpython-36.pyc differ
diff --git a/数据采集/百度数据/html.txt b/数据采集/百度数据/html.txt
new file mode 100644
index 0000000..7df8ede
--- /dev/null
+++ b/数据采集/百度数据/html.txt
@@ -0,0 +1,661 @@
+
+
+
+
+
+
+
+
+
+
+
+
+ 实时更新：新型冠状病毒肺炎疫情地图
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+