BNO
BNO
In[]:=
b=StartWebSession[]
Out[]=
WebSessionObject
In[]:=
WebExecute[b,"OpenWebPage""https://bnonews.com/index.php/2020/02/the-latest-coronavirus-cases/"]
Out[]=
Success
In[]:=
html=WebExecute[b,"JavascriptExecute""return document.getElementsByClassName('wp-block-table')[0].innerHTML"]
Out[]=
<tbody><tr><td><strong>MAINLAND CHINA</strong></td><td><strong>Cases</strong></td><td><strong>Deaths</strong></td><td><strong>Notes</strong></td><td><strong>Links</strong></td></tr><tr><td>Hubei province <br>(including Wuhan)</td><td>29,631</td><td>871</td><td>4,269 serious, 1,236 critical</td><td><a href="http://wjw.hubei.gov.cn/fbjd/dtyw/202002/t20200210_2022514.shtml" target="_blank" rel="noreferrer noopener" aria-label=" (opens in a new tab)">Source</a></td></tr><tr><td>Other regions</td><td>10,540</td><td>37</td><td>979 serious/critical</td><td><a href="http://www.nhc.gov.cn/yjb/s7860/202002/167a0e01b2d24274b03b2ca961107929.shtml" target="_blank" rel="noreferrer noopener" aria-label=" (opens in a new tab)">Source</a></td></tr><tr><td><strong>TOTAL</strong></td><td><strong>40,171</strong></td><td><strong>908</strong></td><td><strong>6,484 serious<br>3,281 recovered<br>23,589 suspected</strong></td><td></td></tr></tbody>
In[]:=
xml=ImportString[html,{"HTML","XMLObject"}];
In[]:=
Cases[xml,XMLElement["tr",{},list_]list,∞]/.{XMLElement["td",_,x_]x}/.{XMLElement["strong",{},x_]x}/.{XMLElement["a",{___,"href"url_,___},_]URL[url]}//.{{s_String}s}/.{XMLElement["br",{"clear""none"},{}]Nothing}/.{{}Missing["NotAvailable"]}/.{{"Hubei province ","(including Wuhan)"}"Hubei province"}/.{"MAINLAND CHINA"|"REGIONS"|"INTERNATIONAL""region"}//Grid
In[]:=
scrape[n_]:=Module[{html,xml,ds},html=WebExecute[b,"JavascriptExecute""return document.getElementsByClassName('wp-block-table')["<>ToString[n]<>"].innerHTML"];xml=ImportString[html,{"HTML","XMLObject"}];xml=Cases[xml,XMLElement["tr",{},list_]list,∞]/.{XMLElement["td",_,x_]x}/.{XMLElement["strong",{},x_]x}/.{XMLElement["a",{___,"href"url_,___},_]URL[url]}//.{{s_String}s}/.{XMLElement["br",{"clear""none"},{}]Nothing}/.{{}Missing["NotAvailable"]}/.{{"Hubei province ","(including Wuhan)"}"Hubei province"}/.{"MAINLAND CHINA"|"REGIONS"|"INTERNATIONAL""region"};ds=Dataset[Map[AssociationThread[First[xml]#]&,Most[Rest[xml]]]];ds[All,<|"region"Interpreter["AdministrativeDivision"|"Country"][#["region"]],"cases"Interpreter["Number"][StringReplace[#Cases,"*"""]],"deaths"Interpreter["Number"][#Deaths],"notes"#Notes,"links"#Links|>&]]
In[]:=
scrape[0]
Out[]=
In[]:=
scrape[1]
Out[]=
In[]:=
cc=scrape[2]
Out[]=
In[]:=
dsFinal=Join[scrape[0],scrape[1],scrape[2]]
Out[]=
