gh-130057: Pygettext: Support translator comments (GH-130061)

Lib/test/test_tools/i18n_data/comments.pot

@@ -0,0 +1,110 @@
+# SOME DESCRIPTIVE TITLE.
+# Copyright (C) YEAR ORGANIZATION
+# FIRST AUTHOR <EMAIL@ADDRESS>, YEAR.
+#
+msgid ""
+msgstr ""
+"Project-Id-Version: PACKAGE VERSION\n"
+"POT-Creation-Date: 2000-01-01 00:00+0000\n"
+"PO-Revision-Date: YEAR-MO-DA HO:MI+ZONE\n"
+"Last-Translator: FULL NAME <EMAIL@ADDRESS>\n"
+"Language-Team: LANGUAGE <LL@li.org>\n"
+"MIME-Version: 1.0\n"
+"Content-Type: text/plain; charset=UTF-8\n"
+"Content-Transfer-Encoding: 8bit\n"
+"Generated-By: pygettext.py 1.5\n"
+
+
+#: comments.py:4
+msgid "foo"
+msgstr ""
+
+#. i18n: This is a translator comment
+#: comments.py:7
+msgid "bar"
+msgstr ""
+
+#. i18n: This is a translator comment
+#. i18n: This is another translator comment
+#: comments.py:11
+msgid "baz"
+msgstr ""
+
+#. i18n: This is a translator comment
+#. with multiple
+#. lines
+#: comments.py:16
+msgid "qux"
+msgstr ""
+
+#. i18n: This is a translator comment
+#: comments.py:21
+msgid "quux"
+msgstr ""
+
+#. i18n: This is a translator comment
+#. with multiple lines
+#. i18n: This is another translator comment
+#. with multiple lines
+#: comments.py:27
+msgid "corge"
+msgstr ""
+
+#: comments.py:31
+msgid "grault"
+msgstr ""
+
+#. i18n: This is another translator comment
+#: comments.py:36
+msgid "garply"
+msgstr ""
+
+#: comments.py:40
+msgid "george"
+msgstr ""
+
+#. i18n: This is another translator comment
+#: comments.py:45
+msgid "waldo"
+msgstr ""
+
+#. i18n: This is a translator comment
+#. i18n: This is also a translator comment
+#. i18n: This is another translator comment
+#: comments.py:50
+msgid "waldo2"
+msgstr ""
+
+#. i18n: This is a translator comment
+#. i18n: This is another translator comment
+#. i18n: This is yet another translator comment
+#. i18n: This is a translator comment
+#. with multiple lines
+#: comments.py:53 comments.py:56 comments.py:59 comments.py:63
+msgid "fred"
+msgstr ""
+
+#: comments.py:65
+msgid "plugh"
+msgstr ""
+
+#: comments.py:67
+msgid "foobar"
+msgstr ""
+
+#. i18n: This is a translator comment
+#: comments.py:71
+msgid "xyzzy"
+msgstr ""
+
+#: comments.py:72
+msgid "thud"
+msgstr ""
+
+#. i18n: This is a translator comment
+#. i18n: This is another translator comment
+#. i18n: This is yet another translator comment
+#: comments.py:78
+msgid "foos"
+msgstr ""
+

Lib/test/test_tools/i18n_data/comments.py

@@ -0,0 +1,78 @@
+from gettext import gettext as _
+
+# Not a translator comment
+_('foo')
+
+# i18n: This is a translator comment
+_('bar')
+
+# i18n: This is a translator comment
+# i18n: This is another translator comment
+_('baz')
+
+# i18n: This is a translator comment
+# with multiple
+# lines
+_('qux')
+
+# This comment should not be included because
+# it does not start with the prefix
+# i18n: This is a translator comment
+_('quux')
+
+# i18n: This is a translator comment
+# with multiple lines
+# i18n: This is another translator comment
+# with multiple lines
+_('corge')
+
+# i18n: This comment should be ignored
+
+_('grault')
+
+# i18n: This comment should be ignored
+
+# i18n: This is another translator comment
+_('garply')
+
+# i18n: comment should be ignored
+x = 1
+_('george')
+
+# i18n: This comment should be ignored
+x = 1
+# i18n: This is another translator comment
+_('waldo')
+
+# i18n: This is a translator comment
+x = 1  # i18n: This is also a translator comment
+# i18n: This is another translator comment
+_('waldo2')
+
+# i18n: This is a translator comment
+_('fred')
+
+# i18n: This is another translator comment
+_('fred')
+
+# i18n: This is yet another translator comment
+_('fred')
+
+# i18n: This is a translator comment
+# with multiple lines
+_('fred')
+
+_('plugh')  # i18n: This comment should be ignored
+
+_('foo'  # i18n: This comment should be ignored
+  'bar')  # i18n: This comment should be ignored
+
+# i18n: This is a translator comment
+_('xyzzy')
+_('thud')
+
+
+## i18n: This is a translator comment
+# # i18n: This is another translator comment
+### ###    i18n: This is yet another translator comment
+_('foos')

Lib/test/test_tools/test_i18n.py

@@ -87,7 +87,8 @@ class Test_pygettext(unittest.TestCase):
         self.maxDiff = None
         self.assertEqual(normalize_POT_file(expected), normalize_POT_file(actual))
 
-    def extract_from_str(self, module_content, *, args=(), strict=True, with_stderr=False):
+    def extract_from_str(self, module_content, *, args=(), strict=True,
+                         with_stderr=False, raw=False):
         """Return all msgids extracted from module_content."""
         filename = 'test.py'
         with temp_cwd(None):
@@ -98,10 +99,11 @@ class Test_pygettext(unittest.TestCase):
                 self.assertEqual(res.err, b'')
             with open('messages.pot', encoding='utf-8') as fp:
                 data = fp.read()
-        msgids = self.get_msgids(data)
+        if not raw:
+            data = self.get_msgids(data)
         if not with_stderr:
-            return msgids
+            return data
-        return msgids, res.err
+        return data, res.err
 
     def extract_docstrings_from_str(self, module_content):
         """Return all docstrings extracted from module_content."""
@@ -381,7 +383,8 @@ class Test_pygettext(unittest.TestCase):
                 contents = input_file.read_text(encoding='utf-8')
                 with temp_cwd(None):
                     Path(input_file.name).write_text(contents)
-                    assert_python_ok('-Xutf8', self.script, '--docstrings', input_file.name)
+                    assert_python_ok('-Xutf8', self.script, '--docstrings',
+                                     '--add-comments=i18n:', input_file.name)
                     output = Path('messages.pot').read_text(encoding='utf-8')
 
                 expected = output_file.read_text(encoding='utf-8')
@@ -437,6 +440,51 @@ class Test_pygettext(unittest.TestCase):
             "*** test.py:3: Variable positional arguments are not allowed in gettext calls\n"
         )
 
+    def test_extract_all_comments(self):
+        """
+        Test that the --add-comments option without an
+        explicit tag extracts all translator comments.
+        """
+        for arg in ('--add-comments', '-c'):
+            with self.subTest(arg=arg):
+                data = self.extract_from_str(dedent('''\
+                # Translator comment
+                _("foo")
+                '''), args=(arg,), raw=True)
+                self.assertIn('#. Translator comment', data)
+
+    def test_comments_with_multiple_tags(self):
+        """
+        Test that multiple --add-comments tags can be specified.
+        """
+        for arg in ('--add-comments={}', '-c{}'):
+            with self.subTest(arg=arg):
+                args = (arg.format('foo:'), arg.format('bar:'))
+                data = self.extract_from_str(dedent('''\
+                # foo: comment
+                _("foo")
+
+                # bar: comment
+                _("bar")
+
+                # baz: comment
+                _("baz")
+                '''), args=args, raw=True)
+                self.assertIn('#. foo: comment', data)
+                self.assertIn('#. bar: comment', data)
+                self.assertNotIn('#. baz: comment', data)
+
+    def test_comments_not_extracted_without_tags(self):
+        """
+        Test that translator comments are not extracted without
+        specifying --add-comments.
+        """
+        data = self.extract_from_str(dedent('''\
+        # Translator comment
+        _("foo")
+        '''), raw=True)
+        self.assertNotIn('#.', data)
+
 
 def update_POT_snapshots():
     for input_file in DATA_DIR.glob('*.py'):
@@ -444,7 +492,8 @@ def update_POT_snapshots():
         contents = input_file.read_bytes()
         with temp_cwd(None):
             Path(input_file.name).write_bytes(contents)
-            assert_python_ok('-Xutf8', Test_pygettext.script, '--docstrings', input_file.name)
+            assert_python_ok('-Xutf8', Test_pygettext.script, '--docstrings',
+                             '--add-comments=i18n:', input_file.name)
             output = Path('messages.pot').read_text(encoding='utf-8')
 
         output = normalize_POT_file(output)

Misc/NEWS.d/next/Tools-Demos/2025-02-12-23-24-37.gh-issue-130057.TKUKI6.rst

@@ -0,0 +1 @@
+Add support for translator comments in :program:`pygettext.py`.

Tools/i18n/pygettext.py

@@ -46,6 +46,12 @@ Options:
     --extract-all
         Extract all strings.
 
+    -cTAG
+    --add-comments=TAG
+        Extract translator comments.  Comments must start with TAG and
+        must precede the gettext call.  Multiple -cTAG options are allowed.
+        In that case, any comment matching any of the TAGs will be extracted.
+
     -d name
     --default-domain=name
         Rename the default output file from messages.pot to name.pot.
@@ -141,7 +147,9 @@ import importlib.util
 import os
 import sys
 import time
+import tokenize
 from dataclasses import dataclass, field
+from io import BytesIO
 from operator import itemgetter
 
 __version__ = '1.5'
@@ -302,12 +310,30 @@ class Message:
     msgctxt: str | None
     locations: set[Location] = field(default_factory=set)
     is_docstring: bool = False
+    comments: list[str] = field(default_factory=list)
 
-    def add_location(self, filename, lineno, msgid_plural=None, *, is_docstring=False):
+    def add_location(self, filename, lineno, msgid_plural=None, *,
+                     is_docstring=False, comments=None):
         if self.msgid_plural is None:
             self.msgid_plural = msgid_plural
         self.locations.add(Location(filename, lineno))
         self.is_docstring |= is_docstring
+        if comments:
+            self.comments.extend(comments)
+
+
+def get_source_comments(source):
+    """
+    Return a dictionary mapping line numbers to
+    comments in the source code.
+    """
+    comments = {}
+    for token in tokenize.tokenize(BytesIO(source).readline):
+        if token.type == tokenize.COMMENT:
+            # Remove any leading combination of '#' and whitespace
+            comment = token.string.lstrip('# \t')
+            comments[token.start[0]] = comment
+    return comments
 
 
 class GettextVisitor(ast.NodeVisitor):
@@ -316,10 +342,18 @@ class GettextVisitor(ast.NodeVisitor):
         self.options = options
         self.filename = None
         self.messages = {}
+        self.comments = {}
+
+    def visit_file(self, source, filename):
+        try:
+            module_tree = ast.parse(source)
+        except SyntaxError:
+            return
 
-    def visit_file(self, node, filename):
         self.filename = filename
-        self.visit(node)
+        if self.options.comment_tags:
+            self.comments = get_source_comments(source)
+        self.visit(module_tree)
 
     def visit_Module(self, node):
         self._extract_docstring(node)
@@ -372,14 +406,51 @@ class GettextVisitor(ast.NodeVisitor):
             msg_data[arg_type] = arg.value
 
         lineno = node.lineno
-        self._add_message(lineno, **msg_data)
+        comments = self._extract_comments(node)
+        self._add_message(lineno, **msg_data, comments=comments)
+
+    def _extract_comments(self, node):
+        """Extract translator comments.
+
+        Translator comments must precede the gettext call and
+        start with one of the comment prefixes defined by
+        --add-comments=TAG. See the tests for examples.
+        """
+        if not self.options.comment_tags:
+            return []
+
+        comments = []
+        lineno = node.lineno - 1
+        # Collect an unbroken sequence of comments starting from
+        # the line above the gettext call.
+        while lineno >= 1:
+            comment = self.comments.get(lineno)
+            if comment is None:
+                break
+            comments.append(comment)
+            lineno -= 1
+
+        # Find the first translator comment in the sequence and
+        # return all comments starting from that comment.
+        comments = comments[::-1]
+        first_index = next((i for i, comment in enumerate(comments)
+                            if self._is_translator_comment(comment)), None)
+        if first_index is None:
+            return []
+        return comments[first_index:]
+
+    def _is_translator_comment(self, comment):
+        return comment.startswith(self.options.comment_tags)
 
     def _add_message(
             self, lineno, msgid, msgid_plural=None, msgctxt=None, *,
-            is_docstring=False):
+            is_docstring=False, comments=None):
         if msgid in self.options.toexclude:
             return
 
+        if not comments:
+            comments = []
+
         key = self._key_for(msgid, msgctxt)
         message = self.messages.get(key)
         if message:
@@ -388,6 +459,7 @@ class GettextVisitor(ast.NodeVisitor):
                 lineno,
                 msgid_plural,
                 is_docstring=is_docstring,
+                comments=comments,
             )
         else:
             self.messages[key] = Message(
@@ -396,6 +468,7 @@ class GettextVisitor(ast.NodeVisitor):
                 msgctxt=msgctxt,
                 locations={Location(self.filename, lineno)},
                 is_docstring=is_docstring,
+                comments=comments,
             )
 
     @staticmethod
@@ -435,6 +508,10 @@ def write_pot_file(messages, options, fp):
 
     for key, locations in sorted_keys:
         msg = messages[key]
+
+        for comment in msg.comments:
+            print(f'#. {comment}', file=fp)
+
         if options.writelocations:
             # location comments are different b/w Solaris and GNU:
             if options.locationstyle == options.SOLARIS:
@@ -473,9 +550,9 @@ def main():
     try:
         opts, args = getopt.getopt(
             sys.argv[1:],
-            'ad:DEhk:Kno:p:S:Vvw:x:X:',
+            'ac::d:DEhk:Kno:p:S:Vvw:x:X:',
-            ['extract-all', 'default-domain=', 'escape', 'help',
+            ['extract-all', 'add-comments=?', 'default-domain=', 'escape',
-             'keyword=', 'no-default-keywords',
+             'help', 'keyword=', 'no-default-keywords',
              'add-location', 'no-location', 'output=', 'output-dir=',
              'style=', 'verbose', 'version', 'width=', 'exclude-file=',
              'docstrings', 'no-docstrings',
@@ -501,6 +578,7 @@ def main():
         excludefilename = ''
         docstrings = 0
         nodocstrings = {}
+        comment_tags = set()
 
     options = Options()
     locations = {'gnu' : options.GNU,
@@ -513,6 +591,8 @@ def main():
             usage(0)
         elif opt in ('-a', '--extract-all'):
             options.extractall = 1
+        elif opt in ('-c', '--add-comments'):
+            options.comment_tags.add(arg)
         elif opt in ('-d', '--default-domain'):
             options.outfile = arg + '.pot'
         elif opt in ('-E', '--escape'):
@@ -558,6 +638,8 @@ def main():
             finally:
                 fp.close()
 
+    options.comment_tags = tuple(options.comment_tags)
+
     # calculate escapes
     make_escapes(not options.escape)
 
@@ -600,12 +682,7 @@ def main():
             with open(filename, 'rb') as fp:
                 source = fp.read()
 
-        try:
+        visitor.visit_file(source, filename)
-            module_tree = ast.parse(source)
-        except SyntaxError:
-            continue
-
-        visitor.visit_file(module_tree, filename)
 
     # write the output
     if options.outfile == '-':