Collapsing excess whitespace after OCR

src/documents/consumer.py

@@ -283,7 +283,7 @@ class Consumer(object):
             r = " ".join(r)
 
         # Strip out excess white space to allow matching to go smoother
-        return re.sub(r"\s+", " ", r)
+        return strip_excess_whitespace(r)
 
     def _store(self, text, doc, thumbnail):
 
@@ -360,6 +360,13 @@ class Consumer(object):
         return Document.objects.filter(checksum=checksum).exists()
 
 
+def strip_excess_whitespace(text):
+    collapsed_spaces = re.sub(r"([^\S\r\n]+)", " ", text)
+    no_leading_whitespace = re.sub("([\n\r]+)([^\S\n\r]+)", '\\1', collapsed_spaces)
+    no_trailing_whitespace = re.sub("([^\S\n\r]+)$", '', no_leading_whitespace)
+    return no_trailing_whitespace
+
+
 def image_to_string(args):
     img, lang = args
     ocr = pyocr.get_available_tools()[0]