PowerShell小技巧之实现文件下载(类wget)

对linux熟悉的读者可能会对linux通过wget下载文件有印象，这个工具功能很强大，在.net环境下提到下载文件大多数人熟悉的是通过system.net.webclient进行下载，这个程序集能实现下载的功能，但是有缺陷，如果碰上类似于…/scripts/?dl=417这类的下载链接将无法正确识别文件名，下载的文件通常会被命名为dl=417这样古怪的名字，其实对应的文件名是在访问这个链接返回结果的http头中包含的。事实上微软也提供了避免这些缺陷的程序集system.net.httpwebrequest 和 httpwebresponse，本文将会使用这两个程序集来实现powershell版wget的功能。

代码不怎么复杂，基本上就是创建httpwebrequest对象，设定useragent和cookiecontainer以免在遇到设置防盗链的服务器出现无法下载的情况。然后通过httpwebrequest对象的getresponse()方法从http头中获取目标文件的大小以及文件名，以便能在下载到文件时提示当前下载进度，在下载完文件后，列出当前目录下对应的文件。代码不复杂，有任何疑问的读者可以留言给我，进行交流，下面上代码：

 =====文件名：get-webfile.ps1=====
function get-webfile {
<# author:fuhj(powershell#live.cn ,http://fuhaijun.com)
   downloads a file or page from the web
.example
  get-webfile http://mirrors.cnnic.cn/apache/couchdb/binary/win/1.4.0/setup-couchdb-1.4.0_r16b01.exe
  downloads the latest version of this file to the current directory
#>

[cmdletbinding(defaultparametersetname="nocredentials")]
   param(
      #  the url of the file/page to download
      [parameter(mandatory=$true,position=0)]
      [system.uri][alias("url")]$uri # = (read-host "the url to download")
   ,
      #  a path to save the downloaded content.
      [string]$filename
   ,
      #  leave the file unblocked instead of blocked
      [switch]$unblocked
   ,
      #  rather than saving the downloaded content to a file, output it. 
      #  this is for text documents like web pages and rss feeds, and allows you to avoid temporarily caching the text in a file.
      [switch]$passthru
   ,
      #  supresses the write-progress during download
      [switch]$quiet
   ,
      #  the name of a variable to store the session (cookies) in
      [string]$sessionvariablename
   ,
      #  text to include at the front of the useragent string
      [string]$useragent = "powershellwget/$(1.0)"
   )

   write-verbose "downloading '$uri'"
   $eap,$erroractionpreference = $erroractionpreference, "stop"
   $request = [system.net.httpwebrequest]::create($uri);
   $erroractionpreference = $eap
   $request.useragent = $(
         "{0} (powershell {1}; .net clr {2}; {3}; )" -f $useragent,
         $(if($host.version){$host.version}else{"1.0"}),
         [environment]::version,
         [environment]::osversion.tostring().replace("microsoft windows ", "win")
      )

   $cookies = new-object system.net.cookiecontainer
   if($sessionvariablename) {
      $cookies = get-variable $sessionvariablename -scope 1
   }
   $request.cookiecontainer = $cookies
   if($sessionvariablename) {
      set-variable $sessionvariablename -scope 1 -value $cookies
   }

   try {
      $res = $request.getresponse();
   } catch [system.net.webexception] {
      write-error $_.exception -category resourceunavailable
      return
   } catch {
      write-error $_.exception -category notimplemented
      return
   }

   if((test-path variable:res) -and $res.statuscode -eq 200) {
      if($filename -and !(split-path $filename)) {
         $filename = join-path (convert-path (get-location -psprovider "filesystem")) $filename
      }
      elseif((!$passthru -and !$filename) -or ($filename -and (test-path -pathtype "container" $filename)))
      {
         [string]$filename = ([regex]'(?i)filename=(.*)$').match( $res.headers["content-disposition"] ).groups[1].value
         $filename = $filename.trim("\/""'")

         $ofs = ""
         $filename = [regex]::replace($filename, "[$([regex]::escape(""$([system.io.path]::getinvalidpathchars())$([io.path]::altdirectoryseparatorchar)$([io.path]::directoryseparatorchar)""))]", "_")
         $ofs = " "

         if(!$filename) {
            $filename = $res.responseuri.segments[-1]
            $filename = $filename.trim("\/")
            if(!$filename) {
               $filename = read-host "please provide a file name"
            }
            $filename = $filename.trim("\/")
            if(!([io.fileinfo]$filename).extension) {
               $filename = $filename + "." + $res.contenttype.split(";")[0].split("/")[1]
            }
         }
         $filename = join-path (convert-path (get-location -psprovider "filesystem")) $filename
      }
      if($passthru) {
         $encoding = [system.text.encoding]::getencoding( $res.characterset )
         [string]$output = ""
      }

      [int]$goal = $res.contentlength
      $reader = $res.getresponsestream()
      if($filename) {
         try {
            $writer = new-object system.io.filestream $filename, "create"
         } catch {
            write-error $_.exception -category writeerror
            return
         }
      }
      [byte[]]$buffer = new-object byte[] 4096
      [int]$total = [int]$count = 0
      do
      {
         $count = $reader.read($buffer, 0, $buffer.length);
         if($filename) {
            $writer.write($buffer, 0, $count);
         }
         if($passthru){
            $output += $encoding.getstring($buffer,0,$count)
         } elseif(!$quiet) {
            $total += $count
            if($goal -gt 0) {
               write-progress "downloading $uri" "saving $total of $goal" -id 0 -percentcomplete (($total/$goal)*100)
            } else {
               write-progress "downloading $uri" "saving $total bytes..." -id 0
            }
         }
      } while ($count -gt 0)

      $reader.close()
      if($filename) {
         $writer.flush()
         $writer.close()
      }
      if($passthru){
         $output
      }
   }
   if(test-path variable:res) { $res.close(); }
   if($filename) {
      ls $filename
   }
}

调用方法，如下：
get-webfile http://mirrors.cnnic.cn/apache/couchdb/binary/win/1.4.0/setup-couchdb-1.4.0_r16b01.exe
这里下载couchdb的最新windows安装包。
执行效果如下图所示：

能够看到在下载文件的过程中会显示当前已下载数和总的文件大小，并且有进度条显示当前下载的进度，跟wget看起来是有些神似了。下载完毕后会显示已经下载文件的情况。