Script to find format differences (#1011)

humitos · web-flow · commit f497c72b7bc5 · 2021-03-19T19:37:42.000+01:00
diff --git a/distutils/setupscript.po b/distutils/setupscript.po
@@ -1225,7 +1225,7 @@ msgid ""
 "The ``long_description`` field is used by PyPI when you publish a package, "
 "to build its project page."
 msgstr ""
-"PyPI utiliza el campo ``descripción larga`` cuando publica un paquete para "
+"PyPI utiliza el campo ``long_description`` cuando publica un paquete para "
 "construir su página de proyecto."
 
 #: ../Doc/distutils/setupscript.rst:621
diff --git a/scripts/format_differences.py b/scripts/format_differences.py
@@ -0,0 +1,56 @@
+import collections
+import os
+import glob
+
+from pprint import pprint
+
+import polib  # fades
+
+PO_DIR = os.path.abspath(
+    os.path.join(
+        os.path.dirname(__file__),
+        '..',
+    ))
+
+
+
+DELIMITERS = ("``", "*")
+
+def has_delimiters(x):
+    for d in DELIMITERS:
+        if d in x:
+            return True
+    return False
+
+def main():
+    files_with_differences = collections.defaultdict(list)
+
+    for i, pofilename in enumerate(glob.glob(PO_DIR + '**/**/*.po')):
+        po = polib.pofile(pofilename)
+        if po.percent_translated() < 85:
+            continue
+
+        for entry in po:
+            words = []
+            wordsid = wordsstr = list()
+
+            if has_delimiters(entry.msgid):
+                wordsid = [word for word in entry.msgid.split() if has_delimiter(word)]
+
+            if has_delimiters(entry.msgstr):
+                wordsstr = [word for word in entry.msgstr.split() if has_delimiter(word)]
+
+            if len(wordsid) != len(wordsstr):
+                key = pofilename.replace(PO_DIR, '')
+                files_with_differences[key].append({
+                    'occurrences': entry.occurrences,
+                    'words': {
+                        'original': wordsid,
+                        'translated': wordsstr,
+                    },
+                })
+
+    return files_with_differences
+
+
+pprint(main())
