Dump HTML snapshots with unique file names based on item key

This change also alters dump() to return the path and file name, see
discussion in #75

doc/index.rst

@@ -315,9 +315,9 @@ Retrieving Files
       A convenient wrapper around :py:meth:`Zotero.file()`. Writes an attachment to disk using the optional path and filename.
       If neither are supplied, the file is written to the current working
       directory, and a :py:meth:`Zotero.item()` call is first made to determine the attachment
-      filename. No error checking is done regarding the path.
+      filename. No error checking is done regarding the path. If successful, the full path including the file name is returned.
 
-      .. note:: HTML snapshots will be dumped as zip files, with "zip" appended to the file name.
+      .. note:: HTML snapshots will be dumped as zip files. These will be named with their API item key, and a .zip extension.
 
       .. code-block:: python
 

pyzotero/zotero.py

@@ -645,17 +645,18 @@ def dump(self, itemkey, filename=None, path=None):
         Dump a file attachment to disk, with optional filename and path
         """
         if not filename:
-            filename = self.item(itemkey)["data"]["filename"]
+            try:
+                filename = self.item(itemkey)['data']['filename']
+            except TypeError:
+                filename = "{i}.zip".format(i=itemkey)
         if path:
             pth = os.path.join(path, filename)
         else:
             pth = filename
-        file = self.file(itemkey)
-        if self.snapshot:
-            self.snapshot = False
-            pth = pth + ".zip"
-        with open(pth, "wb") as f:
-            f.write(file)
+        to_write = self.file(itemkey)
+        with open(pth, 'wb') as f:
+            f.write(to_write)
+        return pth
 
     @retrieve
     def children(self, item, **kwargs):