const fs=require('fs');
const [inp,out]=process.argv.slice(2);
let h=fs.readFileSync(inp,'utf8');
// tokenize
const re=/<(\/?)([a-zA-Z0-9]+)([^>]*?)(\/?)>/g;
let res='',last=0,stack=[],strike=0,m;
function emit(t){ if(strike===0) res+=t; }
while((m=re.exec(h))){
  emit(h.slice(last,m.index)); last=re.lastIndex;
  const close=m[1]==='/', tag=m[2].toLowerCase(), attrs=m[3], selfc=m[4]==='/'||/\/\s*$/.test(attrs);
  if(tag==='span'||tag==='del'||tag==='strike'||tag==='s'){
    if(close){ const x=stack.pop(); if(x) strike--; }
    else if(!selfc){ const isS=tag!=='span'||/line-through/.test(attrs); stack.push(isS); if(isS) strike++; }
  } else if(['p','div','br','tr','li','h1','h2','h3','h4'].includes(tag)) emit('\n');
  else if(tag==='td') emit('\t');
}
emit(h.slice(last));
res=res.replace(/&nbsp;|&#160;/g,' ').replace(/&amp;/g,'&').replace(/&lt;/g,'<').replace(/&gt;/g,'>').replace(/&#8212;|&mdash;/g,'—').replace(/&#(\d+);/g,(a,b)=>String.fromCharCode(+b)).replace(/&quot;/g,'"').replace(/&#39;|&rsquo;/g,"'");
res=res.replace(/[ \t]+\n/g,'\n').replace(/\n{3,}/g,'\n\n');
fs.writeFileSync(out,res);