From 92fb4841a2dad7adb3125b99834888fc15df38f8 Mon Sep 17 00:00:00 2001
From: Beda Kosata <beda.kosata@gmail.com>
Date: Thu, 7 Sep 2023 14:03:09 +0200
Subject: [PATCH] postpone timezone regex evaluation until first use - shaves
 off time from package import

---
 dateparser/timezone_parser.py | 42 +++++++++++++++++++++++++++++------
 dateparser/utils/__init__.py  |  6 ++---
 2 files changed, 38 insertions(+), 10 deletions(-)

diff --git a/dateparser/timezone_parser.py b/dateparser/timezone_parser.py
index c7a284ecb..20193b3ea 100644
--- a/dateparser/timezone_parser.py
+++ b/dateparser/timezone_parser.py
@@ -32,8 +32,8 @@ def __getinitargs__(self):
 
 
 def pop_tz_offset_from_string(date_string, as_offset=True):
-    if _search_regex_ignorecase.search(date_string):
-        for name, info in _tz_offsets:
+    if TzRegexCache.search_regex_ignorecase().search(date_string):
+        for name, info in TzRegexCache.tz_offsets():
             timezone_re = info["regex"]
             timezone_match = timezone_re.search(date_string)
             if timezone_match:
@@ -47,7 +47,7 @@ def pop_tz_offset_from_string(date_string, as_offset=True):
 
 
 def word_is_tz(word):
-    return bool(_search_regex.match(word))
+    return bool(TzRegexCache.search_regex().match(word))
 
 
 def convert_to_local_tz(datetime_obj, datetime_tz_offset):
@@ -85,8 +85,36 @@ def get_local_tz_offset():
     return offset
 
 
-_search_regex_parts = []
-_tz_offsets = list(build_tz_offsets(_search_regex_parts))
-_search_regex = re.compile("|".join(_search_regex_parts))
-_search_regex_ignorecase = re.compile("|".join(_search_regex_parts), re.IGNORECASE)
+class TzRegexCache:
+    _ready = False
+    _search_regex_parts = []
+    _tz_offsets = []
+    _search_regex = None
+    _search_regex_ignorecase = None
+
+    @classmethod
+    def prepare(cls):
+        if not cls._ready:
+            cls._search_regex_parts = []
+            cls._tz_offsets = list(build_tz_offsets(cls._search_regex_parts))
+            cls._search_regex = re.compile("|".join(cls._search_regex_parts))
+            cls._search_regex_ignorecase = re.compile("|".join(cls._search_regex_parts), re.IGNORECASE)
+            cls._ready = True
+
+    @classmethod
+    def tz_offsets(cls):
+        cls.prepare()
+        return cls._tz_offsets
+
+    @classmethod
+    def search_regex(cls):
+        cls.prepare()
+        return cls._search_regex
+
+    @classmethod
+    def search_regex_ignorecase(cls):
+        cls.prepare()
+        return cls._search_regex_ignorecase
+
+
 local_tz_offset = get_local_tz_offset()
diff --git a/dateparser/utils/__init__.py b/dateparser/utils/__init__.py
index 023c5fbb3..5c3679fe0 100644
--- a/dateparser/utils/__init__.py
+++ b/dateparser/utils/__init__.py
@@ -9,7 +9,7 @@
 from pytz import UTC, UnknownTimeZoneError, timezone
 from tzlocal import get_localzone
 
-from dateparser.timezone_parser import StaticTzInfo, _tz_offsets
+from dateparser.timezone_parser import StaticTzInfo, TzRegexCache
 
 
 def strip_braces(date_string):
@@ -73,7 +73,7 @@ def get_timezone_from_tz_string(tz_string):
     try:
         return timezone(tz_string)
     except UnknownTimeZoneError as e:
-        for name, info in _tz_offsets:
+        for name, info in TzRegexCache.tz_offsets():
             if info["regex"].search(" %s" % tz_string):
                 return StaticTzInfo(name, info["offset"])
         else:
@@ -104,7 +104,7 @@ def apply_tzdatabase_timezone(date_time, pytz_string):
 
 
 def apply_dateparser_timezone(utc_datetime, offset_or_timezone_abb):
-    for name, info in _tz_offsets:
+    for name, info in TzRegexCache.tz_offsets():
         if info["regex"].search(" %s" % offset_or_timezone_abb):
             tz = StaticTzInfo(name, info["offset"])
             return utc_datetime.astimezone(tz)