Fixing issue preventing the correct read of images by full_html and reveal exporters.

Currently to see and images in the notebook inside a md cell you can use the prefix files/ to read the content of the directory where the ipynb lives. But, if you do a conversion using full_html or reveal exporters, the are not rendered ok because the previous version of this function erased `/files/` instead of `files/`. With this fix, the html files produced by full_html and reveal exporters can read the images in a correct way.

IPython/nbconvert/filters/strings.py

@@ -65,14 +65,14 def strip_dollars(text):
 
 def rm_fake(text):
     """
-    Remove all occurrences of '/files/' from text
+    Remove all occurrences of 'files/' from text
     
     Parameters
     ----------
     text : str
-        Text to remove '/files/' from
+        Text to remove 'files/' from
     """
-    return text.replace('/files/', '')
+    return text.replace('files/', '')
 
 
 def python_comment(text):