ourresearch/unsub-private#61 - use formula check in _convert_blind me…

…thod for file upload

excel.py

@@ -82,6 +82,8 @@ def _convert_blind(workbook):
         with open(csv_file_name, 'w', encoding='utf-8') as csv_file:
             writer = csv.writer(csv_file, delimiter=',')
             for row in sheet.iter_rows(min_row=1):
+                if 'f' in [w.data_type for w in row]:
+                    raise RuntimeError('Uploaded files can not contain formulas')
                 writer.writerow([cell.value for cell in row])
 
         csv_file_names.append(csv_file_name)

package_input.py

@@ -533,7 +533,7 @@ def load(self, package_id, file_name, commit=False):
             self._copy_raw_to_s3(file_name, package_id, num_rows=None, error="error_reading_file")
             return {"success": False, "message": "error_reading_file", "warnings": []}
         except RuntimeError as e:
-            print("Runtime Error processing file {}".format(str(e)))
+            print("Runtime Error processing file: {}".format(str(e)))
             self._copy_raw_to_s3(file_name, package_id, num_rows=None, error="parsing_error", error_details=str(e))
             return {"success": False, "message": str(e), "warnings": []}
 

tests/test_excel.py

@@ -9,12 +9,17 @@ def test_convert_xls_to_xlsx():
     assert '.xlsx' in x
     assert os.path.isfile(x)
 
-def test_convert_spreadsheet_to_csv():
-    # FIXME: parsed=True is never used in the codebase, so just testing False for now
+def test_convert_spreadsheet_to_csv_parsed_good_file():
     x = convert_spreadsheet_to_csv('tests/test_files/counter/counter4_jr1_2018_01.xlsx', parsed=False)
     assert isinstance(x, list)
     assert isinstance(x[0], str)
     assert os.path.isfile(x[0])
      # FIXME: The below should be True, but dates are messed up in the output csv
     # assert filecmp.cmp(x[0], 'tests/test_files/counter/counter4_jr1_2018_01.csv')
 
+
+def test_convert_spreadsheet_to_csv_parsed_with_formulas():
+    # not using parsed=True, but does check for formulas
+    with pytest.raises(RuntimeError) as err:
+        convert_spreadsheet_to_csv('tests/test_files/journal_price/with-formulas.xlsx', parsed=False)
+        assert "Uploaded files can not contain formulas" in str(err.value)