feat: add caching for timezone offsets, significantly speeds up import

this is different from pr scrapinghub#1181. it builds a cache at install time which
can be distributed.

closes scrapinghub#533

Author: tobymao

MANIFEST.in

@@ -3,6 +3,7 @@ include CONTRIBUTING.rst
 include HISTORY.rst
 include LICENSE
 include README.rst
+include dateparser/data/dateparser_tz_cache.pkl
 include dateparser_data/settings.py
 include requirements.txt
 

dateparser/data/dateparser_tz_cache.pkl

@@ -1,4 +1,8 @@
+import os
+import pickle
+import zlib
 from datetime import datetime, timedelta, timezone, tzinfo
+from pathlib import Path
 
 import regex as re
 
@@ -54,6 +58,12 @@ def convert_to_local_tz(datetime_obj, datetime_tz_offset):
     return datetime_obj - datetime_tz_offset + local_tz_offset
 
 
+def get_local_tz_offset():
+    offset = datetime.now() - datetime.now(tz=timezone.utc).replace(tzinfo=None)
+    offset = timedelta(days=offset.days, seconds=round(offset.seconds, -1))
+    return offset
+
+
 def build_tz_offsets(search_regex_parts):
     def get_offset(tz_obj, regex, repl="", replw=""):
         return (
@@ -78,14 +88,44 @@ def get_offset(tz_obj, regex, repl="", replw=""):
                     yield get_offset(tz_obj, regex, repl=replace, replw=replacewith)
 
 
-def get_local_tz_offset():
-    offset = datetime.now() - datetime.now(tz=timezone.utc).replace(tzinfo=None)
-    offset = timedelta(days=offset.days, seconds=round(offset.seconds, -1))
-    return offset
+local_tz_offset = get_local_tz_offset()
 
+_tz_offsets = None
+_search_regex = None
+_search_regex_ignorecase = None
+
+
+def _load_offsets(cache_path, current_hash):
+    global _tz_offsets, _search_regex, _search_regex_ignorecase
+
+    try:
+        with open(cache_path, mode="rb") as file:
+            (
+                serialized_hash,
+                _tz_offsets,
+                _search_regex,
+                _search_regex_ignorecase,
+            ) = pickle.load(file)
+            if current_hash == serialized_hash:
+                return
+    except (FileNotFoundError, ValueError, TypeError):
+        pass
+
+    _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)
+
+    with open(cache_path, mode="wb") as file:
+        pickle.dump(
+            (current_hash, _tz_offsets, _search_regex, _search_regex_ignorecase),
+            file,
+        )
 
-_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)
-local_tz_offset = get_local_tz_offset()
+
+CACHE_PATH = Path(__file__).parent.joinpath("data", "dateparser_tz_cache.pkl")
+
+_load_offsets(
+    cache_path=CACHE_PATH,
+    current_hash=zlib.crc32(str(timezone_info_list).encode("utf-8")),
+)

dateparser/timezone_parser.py

@@ -1,11 +1,26 @@
 import re
+import subprocess
 
 from setuptools import find_packages, setup
+from setuptools.command import develop, install
 
 __version__ = re.search(
     r"__version__.*\s*=\s*[\"]([^\"]+)[\"]", open("dateparser/__init__.py").read()
 ).group(1)
 
+
+class PostDevelop(develop.develop):
+    def run(self):
+        subprocess.call("python 3 dateparser_scripts/timezones.py", shell=True)
+        develop.develop.run(self)
+
+
+class PostInstall(install.install):
+    def run(self):
+        subprocess.call("python3 dateparser_scripts/timezones.py", shell=True)
+        install.install.run(self)
+
+
 introduction = re.sub(
     r":members:.+|..\sautomodule::.+|:class:|:func:|:ref:",
     "",
@@ -45,6 +60,7 @@
         "fasttext": ["fasttext"],
         "langdetect": ["langdetect"],
     },
+    cmdclass={"develop": PostDevelop, "install": PostInstall},
     license="BSD",
     zip_safe=False,
     keywords="dateparser",