我有一个基于jquery的单页web应用程序。它通过AJAX调用与基于rest的web服务通信。

我正在努力实现以下目标:

向REST url提交包含JSON数据的POST。 如果请求指定了JSON响应,则返回JSON。 如果请求指定PDF/XLS/etc响应,则返回可下载的二进制文件。

我有1和2工作现在,和客户端jquery应用程序通过创建基于JSON数据的DOM元素在网页中显示返回的数据。我还让#3从web服务的角度工作,这意味着如果给定正确的JSON参数,它将创建并返回二进制文件。但我不确定在客户端javascript代码中处理#3的最佳方法。

是否有可能从这样的ajax调用中获得一个可下载的文件?如何让浏览器下载并保存文件?

$.ajax({
    type: "POST",
    url: "/services/test",
    contentType: "application/json",
    data: JSON.stringify({category: 42, sort: 3, type: "pdf"}),
    dataType: "json",
    success: function(json, status){
        if (status != "success") {
            log("Error loading data");
            return;
        }
        log("Data loaded!");
    },
    error: function(result, status, err) {
        log("Error loading data");
        return;
    }
});

服务器响应以下报头:

Content-Disposition:attachment; filename=export-1282022272283.pdf
Content-Length:5120
Content-Type:application/pdf
Server:Jetty(6.1.11)

另一种想法是生成PDF并将其存储在服务器上,然后返回包含文件URL的JSON。然后,在ajax成功处理程序中发出另一个调用,执行如下操作:

success: function(json,status) {
    window.location.href = json.url;
}

但这样做意味着我需要对服务器进行多次调用,我的服务器需要构建可下载的文件,将它们存储在某个地方,然后定期清理存储区域。

一定有更简单的方法。想法吗?


编辑:在检查文档$。ajax,我看到响应dataType只能是xml, html,脚本,json, jsonp,文本之一,所以我猜没有办法直接下载一个文件使用ajax请求,除非我嵌入二进制文件使用数据URI方案建议在@VinayC答案(这不是我想做的事情)。

所以我想我的选择是:

Not use ajax and instead submit a form post and embed my JSON data into the form values. Would probably need to mess with hidden iframes and such. Not use ajax and instead convert my JSON data into a query string to build a standard GET request and set window.location.href to this URL. May need to use event.preventDefault() in my click handler to keep browser from changing from the application URL. Use my other idea above, but enhanced with suggestions from the @naikus answer. Submit AJAX request with some parameter that lets web-service know this is being called via an ajax call. If the web service is called from an ajax call, simply return JSON with a URL to the generated resource. If the resource is called directly, then return the actual binary file.

我越想越喜欢最后一个选项。通过这种方式,我可以获得关于请求的信息(生成时间、文件大小、错误消息等),并且可以在开始下载之前对这些信息采取行动。缺点是服务器上需要额外的文件管理。

还有其他方法吗?我应该知道这些方法的优缺点吗?


当前回答

$scope.downloadSearchAsCSV = function(httpOptions) {
  var httpOptions = _.extend({
    method: 'POST',
    url:    '',
    data:   null
  }, httpOptions);
  $http(httpOptions).then(function(response) {
    if( response.status >= 400 ) {
      alert(response.status + " - Server Error \nUnable to download CSV from POST\n" + JSON.stringify(httpOptions.data));
    } else {
      $scope.downloadResponseAsCSVFile(response)
    }
  })
};
/**
 * @source: https://github.com/asafdav/ng-csv/blob/master/src/ng-csv/directives/ng-csv.js
 * @param response
 */
$scope.downloadResponseAsCSVFile = function(response) {
  var charset = "utf-8";
  var filename = "search_results.csv";
  var blob = new Blob([response.data], {
    type: "text/csv;charset="+ charset + ";"
  });

  if (window.navigator.msSaveOrOpenBlob) {
    navigator.msSaveBlob(blob, filename); // @untested
  } else {
    var downloadContainer = angular.element('<div data-tap-disabled="true"><a></a></div>');
    var downloadLink      = angular.element(downloadContainer.children()[0]);
    downloadLink.attr('href', window.URL.createObjectURL(blob));
    downloadLink.attr('download', "search_results.csv");
    downloadLink.attr('target', '_blank');

    $document.find('body').append(downloadContainer);

    $timeout(function() {
      downloadLink[0].click();
      downloadLink.remove();
    }, null);
  }

  //// Gets blocked by Chrome popup-blocker
  //var csv_window = window.open("","","");
  //csv_window.document.write('<meta name="content-type" content="text/csv">');
  //csv_window.document.write('<meta name="content-disposition" content="attachment;  filename=data.csv">  ');
  //csv_window.document.write(response.data);
};

其他回答

解决方案

内容-性格依恋似乎对我很管用:

self.set_header("Content-Type", "application/json")
self.set_header("Content-Disposition", 'attachment; filename=learned_data.json')

解决方案

应用程序/八进制

我有类似的事情发生在我的JSON,对我来说,在服务器端,我设置头 自我。set_header(“内容类型”、“application / json”) 但是当我把它改成:

self.set_header("Content-Type", "application/octet-stream")

它会自动下载。

还要知道,为了让文件仍然保持.json后缀,你需要在文件名头文件:

self.set_header("Content-Disposition", 'filename=learned_data.json')

I know this kind of old, but I think I have come up with a more elegant solution. I had the exact same problem. The issue I was having with the solutions suggested were that they all required the file being saved on the server, but I did not want to save the files on the server, because it introduced other problems (security: the file could then be accessed by non-authenticated users, cleanup: how and when do you get rid of the files). And like you, my data was complex, nested JSON objects that would be hard to put into a form.

What I did was create two server functions. The first validated the data. If there was an error, it would be returned. If it was not an error, I returned all of the parameters serialized/encoded as a base64 string. Then, on the client, I have a form that has only one hidden input and posts to a second server function. I set the hidden input to the base64 string and submit the format. The second server function decodes/deserializes the parameters and generates the file. The form could submit to a new window or an iframe on the page and the file will open up.

这涉及到更多的工作,可能还有更多的处理,但总的来说,使用这个解决方案我感觉好多了。

代码是c# /MVC

    public JsonResult Validate(int reportId, string format, ReportParamModel[] parameters)
    {
        // TODO: do validation

        if (valid)
        {
            GenerateParams generateParams = new GenerateParams(reportId, format, parameters);

            string data = new EntityBase64Converter<GenerateParams>().ToBase64(generateParams);

            return Json(new { State = "Success", Data = data });
        }

        return Json(new { State = "Error", Data = "Error message" });
    }

    public ActionResult Generate(string data)
    {
        GenerateParams generateParams = new EntityBase64Converter<GenerateParams>().ToEntity(data);

        // TODO: Generate file

        return File(bytes, mimeType);
    }

在客户端

    function generate(reportId, format, parameters)
    {
        var data = {
            reportId: reportId,
            format: format,
            params: params
        };

        $.ajax(
        {
            url: "/Validate",
            type: 'POST',
            data: JSON.stringify(data),
            dataType: 'json',
            contentType: 'application/json; charset=utf-8',
            success: generateComplete
        });
    }

    function generateComplete(result)
    {
        if (result.State == "Success")
        {
            // this could/should already be set in the HTML
            formGenerate.action = "/Generate";
            formGenerate.target = iframeFile;

            hidData = result.Data;
            formGenerate.submit();
        }
        else
            // TODO: display error messages
    }

Another approach instead of saving the file on the server and retrieving it, is to use .NET 4.0+ ObjectCache with a short expiration until the second Action (at which time it can be definitively dumped). The reason that I want to use JQuery Ajax to do the call, is that it is asynchronous. Building my dynamic PDF file takes quite a bit of time, and I display a busy spinner dialog during that time (it also allows other work to be done). The approach of using the data returned in the "success:" to create a Blob does not work reliably. It depends on the content of the PDF file. It is easily corrupted by data in the response, if it is not completely textual which is all that Ajax can handle.

简而言之,没有更简单的方法了。您需要发出另一个服务器请求来显示PDF文件。尽管有一些替代方案,但它们并不完美,也不能在所有浏览器上运行:

看看数据URI方案。如果二进制数据很小,那么你可以使用javascript在URI中打开窗口传递数据。 Windows/IE唯一的解决方案是使用. net控件或FileSystemObject将数据保存到本地文件系统并从那里打开它。

我一直在尝试使用blobs的另一种选择。我已经设法让它下载文本文档,我下载了PDF文件(但他们是损坏的)。

使用blob API,你可以做以下事情:

$.post(/*...*/,function (result)
{
    var blob=new Blob([result]);
    var link=document.createElement('a');
    link.href=window.URL.createObjectURL(blob);
    link.download="myFileName.txt";
    link.click();

});

这是IE 10+, Chrome 8+, FF 4+。参见https://developer.mozilla.org/en-US/docs/Web/API/URL.createObjectURL

它只能在Chrome、Firefox和Opera中下载文件。它使用锚标记上的下载属性强制浏览器下载它。