fixed get_news() start/end date and image issue

Iceloof · Mar 24, 2024 · 3a9e2f3 · 3a9e2f3
1 parent f9f81e4
commit 3a9e2f3
Show file tree

Hide file tree

Showing 3 changed files with 12 additions and 6 deletions.
diff --git a/GoogleNews/__init__.py b/GoogleNews/__init__.py
@@ -82,7 +82,7 @@ def __init__(self,lang="en",period="",start="",end="",encode="utf-8",region=None
         self.__end = end
         self.__encode = encode
         self.__exception = False
-        self.__version = '1.6.13'
+        self.__version = '1.6.14'
 
     def getVersion(self):
         return self.__version
@@ -146,6 +146,12 @@ def build_response(self):
         result = self.content.find_all("a",attrs={'data-ved': True})
         return result
 
+    def remove_after_last_fullstop(self, s):
+        # Find the last occurrence of the full stop
+        last_period_index = s.rfind('.')
+        # Slice the string up to the last full stop
+        return s[:last_period_index+1] if last_period_index != -1 else s
+
     def page_at(self, page=1):
         """
         Retrieves a specific page from google.com in the news sections into __results.
@@ -184,7 +190,7 @@ def page_at(self, page=1):
                     tmp_date = ''
                     tmp_datetime=None
                 try:
-                    tmp_desc = item.find_next_sibling('div').find('div').find_next_sibling('div').find('div').find('div').find('div').contents[0].replace('\n','')
+                    tmp_desc = self.remove_after_last_fullstop(item.find('div').find_next_sibling('div').find('div').find_next_sibling('div').find('div').find('div').find('div').text).replace('\n','')
                 except Exception:
                     tmp_desc = ''
                 try:
@@ -240,7 +246,7 @@ def get_page(self, page=1):
                     tmp_date = ''
                     tmp_datetime=None
                 try:
-                    tmp_desc = item.find_next_sibling('div').find('div').find_next_sibling('div').find('div').find('div').find('div').contents[0].replace('\n','')
+                    tmp_desc = self.remove_after_last_fullstop(item.find('div').find_next_sibling('div').find('div').find_next_sibling('div').find('div').find('div').find('div').text).replace('\n','')
                 except Exception:
                     tmp_desc = ''
                 try:
@@ -329,7 +335,7 @@ def get_news(self, key="",deamplify=False):
                         desc = 'video'
                     # image
                     try:
-                        img = article.find("figure").find("img").get("src")
+                        img = 'news.google.com'+article.find("figure").find("img").get("src")
                     except:
                         img = None
                     # site

diff --git a/setup.py b/setup.py
@@ -5,7 +5,7 @@
 
 setuptools.setup(
     name="GoogleNews",
-    version="1.6.13",
+    version="1.6.14",
     author="Hurin Hu",
     author_email="[email protected]",
     description="Google News search for Python",

diff --git a/test/test_search.py b/test/test_search.py
@@ -51,7 +51,7 @@ class TestStringMethods(unittest.TestCase):
 
   def testVersion(self):
     googlenews = GoogleNews()
-    version = '1.6.13'
+    version = '1.6.14'
     self.assertIn(version, googlenews.getVersion())
     print('Latest version matched')