Merge branch 'master' into wiktextract-typing

tatuylonen · Jan 2, 2024 · 8c87081 · 8c87081
2 parents befd5d8 + 3b6e0af
commit 8c87081
Show file tree

Hide file tree

Showing 3 changed files with 35 additions and 6 deletions.
diff --git a/src/wiktextract/extractor/en/page.py b/src/wiktextract/extractor/en/page.py
@@ -3024,9 +3024,12 @@ def skip_template_fn(name, ht):
             stack.pop()
 
         if len(redirect_list) > 0:
-            new_page_data = base_data.copy()
-            new_page_data["redirects"] = redirect_list
-            page_datas.append(new_page_data)
+            if len(pos_data) > 0:
+                pos_data["redirects"] = redirect_list
+            else:
+                new_page_data = base_data.copy()
+                new_page_data["redirects"] = redirect_list
+                page_datas.append(new_page_data)
 
     def extract_examples(others, sense_base):
         """Parses through a list of definitions and quotes to find examples.

diff --git a/tests/test_long.py b/tests/test_long.py
@@ -76,8 +76,8 @@ def test_long(self):
                 words[word] += 1
                 lang = data.get("lang", "")
                 self.assertGreater(len(lang), 0)
+                # redirect Chinese character maybe not have pos
                 pos = data.get("pos", "")
-                self.assertGreater(len(pos), 0)
                 langs[lang] += 1
                 poses[pos] += 1
                 if data.get("translations"):

diff --git a/tests/test_page.py b/tests/test_page.py
@@ -2,12 +2,10 @@
 #
 # Copyright (c) 2021 Tatu Ylonen.  See file LICENSE and https://ylonen.org
 
-import json
 import unittest
 from unittest.mock import patch
 
 from wikitextprocessor import Page, Wtp
-
 from wiktextract.config import WiktionaryConfig
 from wiktextract.page import parse_page
 from wiktextract.thesaurus import close_thesaurus_db
@@ -494,6 +492,34 @@ def test_zh_see(self, mock_get_page):
                 }
             ],
         )
+        data = parse_page(
+            self.wxr,
+            "车",
+            """
+==Chinese==
+===Glyph origin===
+From cursive script of 車.
+
+===Definitions===
+{{zh-see|車|s}}
+{{zh-see|龺|ss}}
+            """,
+        )
+        self.assertEqual(
+            data,
+            [
+                {
+                    "etymology_templates": [],
+                    "etymology_text": "From cursive script of 車.",
+                    "lang": "Chinese",
+                    "lang_code": "zh",
+                    "redirects": ["車", "龺"],
+                    "word": "车",
+                    "pos": "character",
+                    "senses": [{"tags": ["no-gloss"]}],
+                }
+            ],
+        )
 
     @patch(
         "wikitextprocessor.Wtp.get_page",