factor out function to clean text for open document (#36627)

wcs/qommon/ods.py

@@ -39,6 +39,17 @@ for prefix, uri in NS.items():
     ET.register_namespace(prefix, uri)
 
 
+def clean_text(value):
+    for i in range(0x20):  # remove control characters
+        char = chr(i)
+        if char in ('\t', '\r', '\n'):
+            # only allow tab, carriage return and line feed.
+            continue
+        value = value.replace(char, '')
+    # fffe and ffff are also invalid characters
+    return value.replace('\ufffe', '').replace('\uffff', '')
+
+
 def is_number(value):
     if value and (value.startswith('0') or value.startswith('+')):
         return False
@@ -167,19 +178,10 @@ class WorkSheet(object):
 
 class WorkCell(object):
     def __init__(self, worksheet, value, formdata=None, data_field=None, native_value=None):
+        self.worksheet = worksheet
         if value is None:
             value = ''
-        value = force_text(value, 'utf-8')
-        self.worksheet = worksheet
-        for i in range(0x20):  # remove control characters
-            char = chr(i)
-            if char in ('\t', '\r', '\n'):
-                # only allow tab, carriage return and line feed.
-                continue
-            value = value.replace(char, '')
-        # fffe and ffff are also invalid characters
-        value = value.replace('\ufffe', '').replace('\uffff', '')
-        self.value = value
+        self.value = clean_text(force_text(value, 'utf-8'))
         self.formdata = formdata
         self.data_field = data_field
         self.native_value = native_value