Sort the PDFMiner text objects along the x axis before applying the grouping algorithm, to avoid missing columns

Author: ollynowell

camelot/parsers/stream.py

@@ -129,6 +129,7 @@ def _group_rows(text, row_tol=2):
         rows = []
         temp = []
 
+        text.sort(key=lambda x: (-x.y0, x.x0))
         for t in text:
             # is checking for upright necessary?
             # if t.get_text().strip() and all([obj.upright for obj in t._objs if