<# :
cls&echo off&cd /d "%~dp0"&mode con lines=5000
powershell -NoProfile -ExecutionPolicy bypass "[IO.File]::ReadAllText(\"%~f0\",[Text.Encoding]::GetEncoding('GB2312'))|Invoke-Expression"
pause
exit
#>
$findword="qq.com";
$txtfile1=".\1.txt";
$txtfile2=".\2.txt";

function gethtml($u){
    $t='';
    for($j=1;$j -le 3;$j++){
        try{
            $req=Invoke-WebRequest -Uri $u;
            $t=$req.Content;
            break;
        }catch{
            write-host ('第'+$j.ToString()+'次获取网页源码失败');
            start-sleep -Seconds 3;
        }
    }
    return $t;
}

$newfolder=".\源码";
if(-not (test-path -literal $newfolder)){[void][IO.Directory]::CreateDirectory($newfolder);}
$enc=[Text.Encoding]::GetEncoding('GB2312');
$fs=New-Object System.IO.FileStream($txtfile2, [System.IO.FileMode]::Create);
$sw=New-Object System.IO.StreamWriter($fs, $enc);
$text=[IO.File]::ReadAllLines($txtfile1, $enc);
for($i=0;$i -lt $text.count;$i++){
    write-host $text[$i];
    $html=gethtml $text[$i];
    $outfile=$newfolder.trimend('\')+'\'+($i+1).ToString()+'.txt';
    [Io.File]::WriteAllText($outfile, $html, [Text.Encoding]::GetEncoding('UTF-8'))
    $m=[regex]::matches($html, 'href=[''"]?([^''" ]+)');
    $isexist=$false;
    foreach($it in $m){
        if($it.groups[1].value.Contains($findword)){
            $sw.WriteLine($it.groups[1].value);
            $sw.Flush();
            $isexist=$true;
        }
    }
    if($isexist){
        write-host 'match' -ForegroundColor green;
    }
}
$sw.Close();
$fs.Close();
复制代码