Added support for "\@" formatting code for dates.

mailmerge.py

@@ -1,3 +1,6 @@
+import re
+import shlex
+from datetime import date, time, datetime
 from copy import deepcopy
 import re
 import warnings
@@ -51,7 +54,7 @@ def __init__(self, file, remove_empty_tables=False):
                         m = r.match(instr)
                         if m is None:
                             continue
-                        parent[idx] = Element('MergeField', name=m.group(1))
+                        parent[idx] = Element('MergeField', name=m.group(1), data=instr)
 
                 for parent in part.findall('.//{%(w)s}instrText/../..' % NAMESPACES):
                     children = list(parent)
@@ -83,7 +86,7 @@ def __init__(self, file, remove_empty_tables=False):
                         m = r.match(instr_text)
                         if m is None:
                             continue
-                        parent[idx_begin] = Element('MergeField', name=m.group(1))
+                        parent[idx_begin] = Element('MergeField', name=m.group(1), data=instr_text)
 
                         # use this so we know *where* to put the replacement
                         instr_elements[0].tag = 'MergeText'
@@ -115,7 +118,7 @@ def __get_tree_of_file(self, file):
     def write(self, file):
         # Replace all remaining merge fields with empty values
         for field in self.get_merge_fields():
-            self.merge(**{field: ''})
+            self.merge(**{field: None})
 
         with ZipFile(file, 'w', ZIP_DEFLATED) as output:
             for zi in self.zip.filelist:
@@ -253,16 +256,115 @@ def merge(self, parts=None, **replacements):
                 for part in parts:
                     self.__merge_field(part, field, replacement)
 
+    @classmethod
+    def eval_strftime(cls, dt, fmt):
+        if dt is None:
+            return ''
+        elif not isinstance(dt, (datetime, date, time, )):
+            return str(dt)
+
+        def repl(m):
+            res = ''
+            pattern = m.group(0)
+            while pattern:
+                # Years
+                if pattern[:4] in ['yyyy', 'YYYY']:
+                    res += '%Y'
+                    pattern = pattern[4:]
+                elif pattern[:2] in ['yy', 'YY']:
+                    res += '%y'
+                    pattern = pattern[2:]
+                elif pattern[:1] in ['y', 'Y']:
+                    res += '%Y'
+                    pattern = pattern[1:]
+
+                # Months
+
+                elif pattern[:4] == 'MMMM':
+                    res += '%B'
+                    pattern = pattern[4:]
+                elif pattern[:2] == 'MM':
+                    res += '%m'
+                    pattern = pattern[2:]
+                elif pattern[:1] == 'M':
+                    # Hack for non-zero padded month
+                    res += str(dt.month)
+                    pattern = pattern[1:]
+
+                # Days
+                elif pattern[:4] == 'dddd':
+                    res += '%A'
+                    pattern = pattern[4:]
+                elif pattern[:2] == 'dd':
+                    res += '%d'
+                    pattern = pattern[2:]
+                elif pattern[:1] == 'd':
+                    # Hack for non-zero padded month
+                    res += str(dt.day)
+                    pattern = pattern[1:]
+                else:
+                    break
+            return res
+
+        if fmt[0] == '"' and fmt[-1] == '"':
+            fmt = fmt[1:-1]
+
+        fmt = re.sub(r'[dmMyY]+', repl, fmt)
+        try:
+            return dt.strftime(fmt)
+        except AttributeError:
+            return str(dt)
+
+    @classmethod
+    def eval_star(cls, data):
+        # This is similiar to what Word is doing. It uses
+        # default formats based on locale. This should be
+        # consistent behaviour.
+        if isinstance(data, (datetime, )):
+            # i.e. 08/16/1988 13:42
+            return data.strftime('%x %X')
+        elif isinstance(data, (date, )):
+            # i.e. 08/16/1988
+            return data.strftime('%x')
+        elif isinstance(data, (time, )):
+            # i.e. 13:18:00
+            return data.strftime('%X')
+        elif data is None:
+            return ''
+        else:
+            return str(data)
+
+    @classmethod
+    def eval(cls, data, code):
+        params = shlex.split(code, posix=False)
+        params = params[2:]
+        evaluated = False
+        for i, param in enumerate(params):
+            if param == '\\@':
+                data = cls.eval_strftime(data, params[i + 1])
+                evaluated = True
+            elif param == '\\*':
+                data = cls.eval_star(data)
+                evaluated = True
+
+        # According to Word lack of "\* MERGEFORMAT" in MERGEFIELD is perfectly fine
+        # so we treat lack of evaluation as \* code.
+        if not evaluated:
+            data = cls.eval_star(data)
+        return data
+
     def __merge_field(self, part, field, text):
         for mf in part.findall('.//MergeField[@name="%s"]' % field):
             children = list(mf)
+            # Original code is saved in "data" attribute
+            instr = mf.attrib['data']
             mf.clear()  # clear away the attributes
             mf.tag = '{%(w)s}r' % NAMESPACES
             mf.extend(children)
 
             nodes = []
             # preserve new lines in replacement text
-            text = text or ''  # text might be None
+            text = self.eval(text, instr)
             text_parts = text.replace('\r', '').split('\n')
             for i, text_part in enumerate(text_parts):
                 text_node = Element('{%(w)s}t' % NAMESPACES)

tests/test_eval.py

@@ -0,0 +1,48 @@
+from unittest import TestCase
+from datetime import datetime
+from mailmerge import MailMerge
+
+
+class TestEval(TestCase):
+
+    def setUp(self):
+        self.today = datetime(2018, 7, 1, 10, 12, 54, 0)
+
+    def test_year(self):
+        self.assertEqual(MailMerge.eval_strftime(self.today, 'y'), '2018')
+        self.assertEqual(MailMerge.eval_strftime(self.today, 'Y'), '2018')
+        self.assertEqual(MailMerge.eval_strftime(self.today, 'yy'), '18')
+        self.assertEqual(MailMerge.eval_strftime(self.today, 'YY'), '18')
+        self.assertEqual(MailMerge.eval_strftime(self.today, 'yyyy'), '2018')
+        self.assertEqual(MailMerge.eval_strftime(self.today, 'YYYY'), '2018')
+        self.assertEqual(MailMerge.eval_strftime(self.today, 'yyyyy'), '20182018')
+        self.assertEqual(MailMerge.eval_strftime(self.today, 'YYYYY'), '20182018')
+
+    def test_month(self):
+        self.assertEqual(MailMerge.eval_strftime(self.today, 'M'), '7')
+        self.assertEqual(MailMerge.eval_strftime(self.today, 'MM'), '07')
+        self.assertEqual(MailMerge.eval_strftime(self.today, 'MMMM'), self.today.strftime('%B'))
+
+    def test_days(self):
+        self.assertEqual(MailMerge.eval_strftime(self.today, 'd'), '1')
+        self.assertEqual(MailMerge.eval_strftime(self.today, 'dd'), '01')
+        self.assertEqual(MailMerge.eval_strftime(self.today, 'dddd'), self.today.strftime('%A'))
+
+    def test_strftime_none(self):
+        self.assertEqual(MailMerge.eval_strftime(None, 'd'), '')
+
+    def test_strftime_str(self):
+        self.assertEqual(MailMerge.eval_strftime('foo', 'd'), 'foo')
+
+    def test_eval(self):
+        self.assertEqual(MailMerge.eval_star(self.today), self.today.strftime('%x %X'))
+        self.assertEqual(MailMerge.eval_star(self.today.date()), self.today.strftime('%x'))
+        self.assertEqual(MailMerge.eval_star(self.today.time()), self.today.strftime('%X'))
+
+    def test_eval_none(self):
+        self.assertEqual(MailMerge.eval(None, 'MAILMERGE Foo \\@ "y-MM-dd" MERGEFORMAT'), '')
+
+    def test_parse_code(self):
+        self.assertEqual(MailMerge.eval(self.today, 'MAILMERGE Foo \\* MERGEFORMAT'), self.today.strftime('%x %X'))
+        self.assertEqual(MailMerge.eval(self.today, 'MAILMERGE Foo'), self.today.strftime('%x %X'))
+        self.assertEqual(MailMerge.eval(self.today, 'MAILMERGE Foo \\@ "y-MM-dd" MERGEFORMAT'), '2018-07-01')