From ddf8db54c6abe62175994ad89b347a87c50c0a74 Mon Sep 17 00:00:00 2001
From: Shabbir Hasan <68828793+ShabbirHasan1@users.noreply.github.com>
Date: Wed, 27 Dec 2023 11:29:21 +0530
Subject: [PATCH] Update update_user_agents.py

---
 update_user_agents.py | 77 ++++++++++++++++++++++++-------------------
 1 file changed, 43 insertions(+), 34 deletions(-)

diff --git a/update_user_agents.py b/update_user_agents.py
index 1d9b476..abb990e 100644
--- a/update_user_agents.py
+++ b/update_user_agents.py
@@ -3,73 +3,82 @@
 from github import Github
 from lxml import html
 
-user_agents_file_name = 'user_agents.json'
-user_agents_file_path = os.path.join(
-    os.path.dirname(__file__), user_agents_file_name)
+user_agents_file_name = "user_agents.json"
+user_agents_file_path = os.path.join(os.path.dirname(__file__), user_agents_file_name)
 
 _os_field_include_patterns = [
-    re.compile(r'^windows nt \d+\.\d+$', flags=re.IGNORECASE),
-    re.compile(r'^macintosh$', flags=re.IGNORECASE),
-    re.compile(r'^linux (x86_64|i686)$', flags=re.IGNORECASE),
+    re.compile(r"^windows nt \d+\.\d+$", flags=re.IGNORECASE),
+    re.compile(r"^macintosh$", flags=re.IGNORECASE),
+    re.compile(r"^linux (x86_64|i686)$", flags=re.IGNORECASE),
 ]
 _os_field_exclude_patterns = [
-    re.compile(r'\bwindows mobile\b', flags=re.IGNORECASE),
-    re.compile(r'\bxbox\b', flags=re.IGNORECASE),
-    re.compile(r'\biphone\b', flags=re.IGNORECASE),
-    re.compile(r'\bipad\b', flags=re.IGNORECASE),
-    re.compile(r'\bipod\b', flags=re.IGNORECASE),
-    re.compile(r'\bandroid\b', flags=re.IGNORECASE),
+    re.compile(r"\bwindows mobile\b", flags=re.IGNORECASE),
+    re.compile(r"\bxbox\b", flags=re.IGNORECASE),
+    re.compile(r"\biphone\b", flags=re.IGNORECASE),
+    re.compile(r"\bipad\b", flags=re.IGNORECASE),
+    re.compile(r"\bipod\b", flags=re.IGNORECASE),
+    re.compile(r"\bandroid\b", flags=re.IGNORECASE),
 ]
 
 _saved_user_agents = None
 
+
 def get_saved_user_agents():
     global _saved_user_agents
     if _saved_user_agents is None:
-        with open(user_agents_file_path, 'r') as f:
+        with open(user_agents_file_path, "r") as f:
             _saved_user_agents = json.load(f)
     return _saved_user_agents
 
+
 def get_latest_user_agents():
-    user_agents = []
-    base_url = 'https://www.whatismybrowser.com/guides/the-latest-user-agent/'
-    for browser in ('chrome', 'firefox', 'safari', 'edge'):
+    user_agents, session = [], requests.session()
+    session.headers.update(
+        {
+            "authority": "www.whatismybrowser.com",
+            "referer": "https://www.whatismybrowser.com/",
+            "User-Agent": random.choice(get_saved_user_agents()),
+        }
+    )
+    base_url = "https://www.whatismybrowser.com/guides/the-latest-user-agent/"
+    for browser in ("chrome", "firefox", "safari", "edge"):
         time.sleep(1)
-        response = requests.get(
-            ''.join((base_url, browser)),
-            headers={'User-Agent': random.choice(get_saved_user_agents())},
-        )
-        elems = html.fromstring(response.text).cssselect('td li span.code')
+        response = session.get("".join((base_url, browser)))
+        elems = html.fromstring(response.text).cssselect("td li span.code")
         browser_uas = []
         for elem in elems:
             ua = elem.text_content().strip()
-            if not ua.startswith('Mozilla/5.0 ('):
+            if not ua.startswith("Mozilla/5.0 ("):
                 continue
             browser_uas.append(ua)
         for ua in browser_uas:
-            os_type = ua[len('Mozilla/5.0 ('):ua.find(')')].lower()
-            os_fields = [p.strip() for p in os_type.split(';')]
+            os_type = ua[len("Mozilla/5.0 (") : ua.find(")")].lower()
+            os_fields = [p.strip() for p in os_type.split(";")]
 
-            if any(p.match(f) for p, f in product(
-                    _os_field_exclude_patterns, os_fields)):
+            if any(
+                p.match(f) for p, f in product(_os_field_exclude_patterns, os_fields)
+            ):
                 continue
-            if any(p.match(f) for p, f in product(
-                    _os_field_include_patterns, os_fields)):
+            if any(
+                p.match(f) for p, f in product(_os_field_include_patterns, os_fields)
+            ):
                 user_agents.append(ua)
     return user_agents
 
+
 def json_dump(obj):
-    return json.dumps(obj, indent=4).strip() + '\n'
+    return json.dumps(obj, indent=4).strip() + "\n"
+
 
 def update_files_on_github(new_user_agents_json):
-    gh = Github(os.environ['GITHUB_TOKEN'])
-    repo = gh.get_repo(os.environ['GITHUB_REPOSITORY'])
-#     for branch in ('main', 'gh-pages'):
-    branch = 'main'
+    gh = Github(os.environ["GITHUB_TOKEN"])
+    repo = gh.get_repo(os.environ["GITHUB_REPOSITORY"])
+    #     for branch in ('main', 'gh-pages'):
+    branch = "main"
     f = repo.get_contents(user_agents_file_name, ref=branch)
     repo.update_file(
         f.path,
-        message=f'Update {user_agents_file_name} on {branch} branch',
+        message=f"Update {user_agents_file_name} on {branch} branch",
         content=new_user_agents_json,
         sha=f.sha,
         branch=branch,