more tests of the new parser

2020-11-26 00:08:23 +01:00
parent bd0db57604
commit f901def797
10 changed files with 146 additions and 10 deletions
--- a/src/paperless_tesseract/parsers.py
+++ b/src/paperless_tesseract/parsers.py
@@ -160,7 +160,9 @@ def strip_excess_whitespace(text):
        r"([\n\r]+)([^\S\n\r]+)", '\\1', collapsed_spaces)
    no_trailing_whitespace = re.sub(
        r"([^\S\n\r]+)$", '', no_leading_whitespace)
-    return no_trailing_whitespace
+
+    # TODO: this needs a rework
+    return no_trailing_whitespace.strip()


 def get_text_from_pdf(pdf_file):