Remove stale sfall female files

Author: burner1024

msg2po/conversion.py

@@ -5,10 +5,11 @@
 
 from __future__ import annotations
 
-import os
-from collections import OrderedDict
-from dataclasses import dataclass
-from typing import TYPE_CHECKING
+import os
+from collections import OrderedDict
+from dataclasses import dataclass
+from pathlib import Path
+from typing import TYPE_CHECKING
 
 if TYPE_CHECKING:
     from msg2po.indexed_po import IndexedPO
@@ -184,14 +185,18 @@ def po2file(
         # what's out path?
         female_file = get_female_filepath(output_file, dst_dir, same)
 
-        # If need to create the file
-        if same:  # if female translation is the same?
-            if female_file is None:  # don't need to copy, automatic fallback
-                logger.debug(f"Female strings are same, not copying - sfall will fallback to male {occurrence_path}")
-                return
-            else:
-                logger.debug(f"Female strings are same, copying to {female_file}")
-                copycreate(output_file, female_file)
+        # If need to create the file
+        if same:  # if female translation is the same?
+            if female_file is None:  # don't need to copy, automatic fallback
+                stale_female = get_stale_female_filepath(output_file)
+                if stale_female is not None and Path(stale_female).exists():
+                    logger.debug(f"Female strings are same, deleting stale female file {stale_female}")
+                    Path(stale_female).unlink()
+                logger.debug(f"Female strings are same, not copying - sfall will fallback to male {occurrence_path}")
+                return
+            else:
+                logger.debug(f"Female strings are same, copying to {female_file}")
+                copycreate(output_file, female_file)
         else:  # if it's different, extract separately
             if female_file is None:
                 logger.warning(
@@ -205,7 +210,7 @@ def po2file(
                     file2.writelines(lines_female)
 
 
-def get_female_filepath(path: str, dst_dir: str, same: bool = True) -> str | None:
+def get_female_filepath(path: str, dst_dir: str, same: bool = True) -> str | None:
     """Determine female file path based on format and sfall conventions."""
     # default: just add _female suffix
     female_path: str | None = path.replace(dst_dir + os.sep, dst_dir + CONFIG.female_dir_suffix + os.sep)
@@ -214,8 +219,17 @@ def get_female_filepath(path: str, dst_dir: str, same: bool = True) -> str | Non
         if "cuts" in path.split(os.sep):  # cuts dont' fallback
             female_path = path.replace(os.sep + "cuts" + os.sep, os.sep + "cuts_female" + os.sep)
         if "dialog" in path.split(os.sep) and not same:  # dialog, female translation differs
-            female_path = path.replace(os.sep + "dialog" + os.sep, os.sep + "dialog_female" + os.sep)
-    return female_path
+            female_path = path.replace(os.sep + "dialog" + os.sep, os.sep + "dialog_female" + os.sep)
+    return female_path
+
+
+def get_stale_female_filepath(path: str) -> str | None:
+    """Return the stale dialog_female path to remove when sfall falls back to male."""
+    if CONFIG.extract_format != "sfall":
+        return None
+    if "dialog" not in path.split(os.sep):
+        return None
+    return path.replace(os.sep + "dialog" + os.sep, os.sep + "dialog_female" + os.sep)
 
 
 def get_line_format(e, ext: str):

tests/test_conversion.py

@@ -1,9 +1,12 @@
 # Tests for file2po, po2file, and roundtrip conversions.
 
+from dataclasses import replace
+
 import polib
 import pytest
 
 from msg2po.conversion import file2msgstr, file2po, po2file
+from msg2po.config import CONFIG
 from msg2po.po_utils import (
     CONTEXT_FEMALE,
     EMPTY_COMMENT,
@@ -147,6 +150,24 @@ def test_translated_uses_msgstr(self, msg_file, tmp_path):
         assert "Bonjour le monde" in content
         assert "{100}{}{Bonjour le monde}" in content
 
+    def test_removes_stale_dialog_female_file_when_falling_back_to_male(self, tmp_path, monkeypatch):
+        po = polib.POFile()
+        po.append(polib.POEntry(msgid="Hello", msgstr="Bonjour", occurrences=[("dialog/test.msg", "100")]))
+
+        config = replace(CONFIG, extract_format="sfall")
+        monkeypatch.setattr("msg2po.conversion.CONFIG", config)
+
+        dst_dir = tmp_path / "french"
+        output = dst_dir / "dialog" / "test.msg"
+        stale_female = dst_dir / "dialog_female" / "test.msg"
+        stale_female.parent.mkdir(parents=True)
+        stale_female.write_text("{100}{}{STALE}\n", encoding="utf-8")
+
+        po2file(po, str(output), "utf-8", "dialog/test.msg", dst_dir=str(dst_dir))
+
+        assert output.exists()
+        assert not stale_female.exists()
+
 
 class TestFile2Msgstr:
     def test_loads_translations(self, msg_file, msg_translated_file):