Implemented filename guessing, headerdata first and then url

Author: rac2030

pom.xml

@@ -11,7 +11,7 @@
 
     <groupId>ch.racic.selenium.helper</groupId>
     <artifactId>SeleniumDownloadHelper</artifactId>
-    <version>0.2-SNAPSHOT</version>
+    <version>0.3-SNAPSHOT</version>
     <packaging>jar</packaging>
 
     <name>Selenium Download Helper library</name>

src/main/java/ch/racic/selenium/helper/download/FileData.java

@@ -0,0 +1,27 @@
+/*
+ * Copyleft (c) 2014. This code is for learning purposes only. Do whatever you like with it but don't take it as perfect code.
+ * Michel Racic (http://rac.su/+) => github.com/rac2030
+ */
+
+package ch.racic.selenium.helper.download;
+
+/**
+ * Transport wrapper for raw data
+ */
+public class FileData {
+    private String guessedFilename;
+    private byte[] data;
+
+    public FileData(String guessedFilename, byte[] data) {
+        this.guessedFilename = guessedFilename;
+        this.data = data;
+    }
+
+    public String getGuessedFilename() {
+        return guessedFilename;
+    }
+
+    public byte[] getData() {
+        return data;
+    }
+}

src/main/java/ch/racic/selenium/helper/download/SeleniumDownloadHelper.java

@@ -12,13 +12,10 @@
 import org.openqa.selenium.WebDriver;
 import org.openqa.selenium.htmlunit.HtmlUnitDriver;
 import org.openqa.selenium.ie.InternetExplorerDriver;
-import sun.plugin.dom.exception.BrowserNotSupportedException;
 
 import java.io.File;
 import java.io.IOException;
-import java.io.UnsupportedEncodingException;
 import java.net.URL;
-import java.nio.charset.Charset;
 
 /**
  * Created by rac on 08.06.14.
@@ -89,7 +86,7 @@ private String getDownloadJsCallScript(URL url) {
      * @param url
      * @return raw data
      */
-    public byte[] getFileFromUrl(URL url) {
+    public FileData getFileFromUrlRaw(URL url) {
         String scriptCollection;
         if (js instanceof InternetExplorerDriver) {
             scriptCollection = ieHackTestJs + ieHackJs + base64Js + dlHelperJs + getDownloadJsCallScript(url);
@@ -99,8 +96,14 @@ public byte[] getFileFromUrl(URL url) {
         } else {
             scriptCollection = ieHackTestJs + base64Js + dlHelperJs + getDownloadJsCallScript(url);
         }
-        String encodedContent = (String) js.executeScript(scriptCollection);
-        return Base64.decodeBase64(encodedContent);
+        String jsRetVal = (String) js.executeScript(scriptCollection);
+        String[] jsRetArr = jsRetVal.split(":contentstarts:", 2);
+        String encodedContent = jsRetArr[1];
+        String fileName = jsRetArr[0];
+        if (fileName.equals("")) {
+            fileName = url.getFile().replaceAll("/", "");
+        }
+        return new FileData(fileName, Base64.decodeBase64(encodedContent));
     }
 
     /**
@@ -113,9 +116,25 @@ public byte[] getFileFromUrl(URL url) {
      * @throws IOException
      */
     public File getFileFromUrl(URL url, File outputFile) throws IOException {
-        byte[] content = getFileFromUrl(url);
+        byte[] content = getFileFromUrlRaw(url).getData();
         FileUtils.writeByteArrayToFile(outputFile, content);
         return outputFile;
     }
 
+    /**
+     * Executes XHR request trough JavaScript to download the given file in the context of the current WebDriver
+     * session. This method returns a tmp file and takes a guessed filename
+     *
+     * @param url
+     * @return outpuFile
+     * @throws IOException
+     */
+    public File getFileFromUrl(URL url) throws IOException {
+        FileData fd = getFileFromUrlRaw(url);
+        byte[] content = fd.getData();
+        File tmpFile = File.createTempFile("", fd.getGuessedFilename());
+        FileUtils.writeByteArrayToFile(tmpFile, content);
+        return tmpFile;
+    }
+
 }

src/main/resources/ch/racic/selenium/helper/download/js/seleniumDownloadHelper.js

@@ -3,6 +3,7 @@
  * Michel Racic (http://rac.su/+) => github.com/rac2030
  */
 
+
 var seleniumDownloadHelper = {
     getBinary: function (url) {
         // Mozilla/Safari/IE7+
@@ -21,16 +22,29 @@ var seleniumDownloadHelper = {
         }
 
         xhr.send(null);
+
         if (xhr.status != 200) return '';
+
+        // Code taken from http://stackoverflow.com/q/16086162
+        var disp = xhr.getResponseHeader('Content-Disposition');
+        var fileName = "";
+        if (disp && disp.search('filename') != -1) {
+            var filenamePattern = "filename=\"(.*)\".*";
+            var re = new RegExp(filenamePattern);
+            if (re.test(disp)) {
+                fileName = re.exec(disp)[1];
+            }
+        }
+
         if (IE_HACK) {
-            return bin2arr(xhr.responseBody);
+            return [fileName, bin2arr(xhr.responseBody)];
         } else {
-            return xhr.responseText;
+            return [fileName, xhr.responseText];
         }
     },
 
     getB64Binary: function (url) {
         var content = seleniumDownloadHelper.getBinary(url);
-        return base64Encode(content);
+        return content[0] + ":contentstarts:" + base64Encode(content[1]);
     }
 };

src/test/java/ch/racic/selenium/helper/download/SeleniumDownloadHelperTest.java

@@ -116,8 +116,9 @@ public void testGetFileFromUrlPhantomJS() throws Exception {
     private void invokeGetFileDataFromUrl() throws Exception {
         driver.get(baseUrl + indexPage);
         SeleniumDownloadHelper sdlh = new SeleniumDownloadHelper(driver);
-        byte[] fileContent = sdlh.getFileFromUrl(new URL(baseUrl + testPdf));
-        Assert.assertArrayEquals("Raw data is correct", referenceContent, fileContent);
+        FileData testFileData = sdlh.getFileFromUrlRaw(new URL(baseUrl + testPdf));
+        Assert.assertArrayEquals("Raw data is correct", referenceContent, testFileData.getData());
+        Assert.assertTrue("Guessed name from URL is correct", testPdf.equals(testFileData.getGuessedFilename()));
     }
 
     private void invokeGetFileFromUrl() throws Exception {