refactor: replace usage of pytz with zoneinfo

docs/manual.md

@@ -143,7 +143,7 @@ Connecting to the Peregrino IP.21 data source using AspenTech Process Data REST
 ``` python
 c = tagreader.IMSClient(datasource="PER",
                         imstype="aspenone",
-                        tz="Brazil/East",
+                        tz="America/Sao_Paulo",
                         url="https://aspenone-per.equinor.com/ProcessExplorer/ProcessData/AtProcessDataREST.dll")
 c.connect()
 ```
@@ -303,7 +303,7 @@ There are two levels of determining which time zone input arguments should be in
 
 The client-provided time zone can be specified with the optional `tz` argument (string, e.g. "*US/Central*") during client creation. If it is not specified, then the default value *Europe/Oslo* is used. Note that for the most common use case where Equinor employees want to fetch data from Norwegian assets and display them with Norwegian time stamps, nothing needs to be done.
 
-*Note:* It is a good idea to update the `pytz` package rather frequently (at least twice per year) to ensure that time zone information is up to date. `pip install --upgrade pytz` .
+*Note:* Tagreader uses Python's built-in `zoneinfo` module. Keep your system time zone database up to date. On environments without an IANA database, install/update `tzdata` with `pip install --upgrade tzdata`.
 
 **Example (advanced usage)**
 
@@ -315,7 +315,7 @@ from datetime import datetime, timedelta
 from dateutil import tz
 c = tagreader.IMSClient("PINO", "pi", tz="US/Central")  # Force output data to Houston time
 c.connect()
-tzinfo = tz.gettz("Brazil/East")  # Generate timezone object for Rio local time
+tzinfo = tz.gettz("America/Sao_Paulo")  # Generate timezone object for Rio local time
 event_time = datetime(2020, 7, 20, 15, 5, 0, tzinfo=tzinfo)
 start_time = event_time - timedelta(minutes=30)
 end_time = event_time + timedelta(minutes=10)

documentation/docs/about/usage/basic-usage.md

@@ -74,7 +74,7 @@ stamps as well as the returned data shall use Rio local time, and using the loca
 ``` python
 c = tagreader.IMSClient(datasource="PER",
                         imstype="aspenone",
-                        tz="Brazil/East",
+                        tz="America/Sao_Paulo",
                         url="https://aspenone-per.equinor.com/ProcessExplorer/ProcessData/AtProcessDataREST.dll")
 c.connect()
 ```

documentation/docs/about/usage/time-zone.md

@@ -20,8 +20,8 @@ client creation. If it is not specified, then the default value *Europe/Oslo* is
 use case where Equinor employees want to fetch data from Norwegian assets and display them with Norwegian time stamps,
 nothing needs to be done.
 
-*Note:* It is a good idea to update the `pytz` package rather frequently (at least twice per year) to ensure that time
-zone information is up-to-date. `pip install --upgrade pytz` .
+*Note:* Tagreader uses Python's built-in `zoneinfo` module. Keep your system time zone database up to date. On
+environments without an IANA database, install/update `tzdata` with `pip install --upgrade tzdata`.
 
 **Example (advanced usage)**
 
@@ -36,7 +36,7 @@ from datetime import datetime, timedelta
 from dateutil import tz
 c = tagreader.IMSClient("PINO", "pi", tz="US/Central")  # Force output data to Houston time
 c.connect()
-tzinfo = tz.gettz("Brazil/East")  # Generate timezone object for Rio local time
+tzinfo = tz.gettz("America/Sao_Paulo")  # Generate timezone object for Rio local time
 event_time = datetime(2020, 7, 20, 15, 5, 0, tzinfo=tzinfo)
 start_time = event_time - timedelta(minutes=30)
 end_time = event_time + timedelta(minutes=10)

tagreader/cache.py

@@ -1,9 +1,8 @@
-from datetime import datetime, timedelta
+from datetime import datetime, timedelta, timezone
 from pathlib import Path
 from typing import Dict, List, Optional, Tuple, Union, cast
 
 import pandas as pd
-import pytz
 from diskcache import Cache
 
 from tagreader.logger import logger
@@ -21,7 +20,7 @@ def safe_tagname(tagname: str) -> str:
 def timestamp_to_epoch(timestamp: datetime) -> int:
     origin = datetime(1970, 1, 1)
     if timestamp.tzinfo is not None:
-        timestamp = timestamp.astimezone(pytz.utc).replace(tzinfo=None)
+        timestamp = timestamp.astimezone(timezone.utc).replace(tzinfo=None)
     return (timestamp - origin) // timedelta(seconds=1)
 
 

tagreader/clients.py

@@ -2,10 +2,10 @@
 from itertools import groupby
 from operator import itemgetter
 from typing import Any, Dict, List, Optional, Tuple, Union
+from zoneinfo import ZoneInfo, ZoneInfoNotFoundError
 
 import numpy as np
 import pandas as pd
-import pytz
 import requests
 
 from tagreader.cache import BucketCache, SmartCache
@@ -26,7 +26,7 @@
     list_piwebapi_sources,
 )
 
-NONE_START_TIME = datetime(1970, 1, 1, tzinfo=pytz.UTC)
+NONE_START_TIME = datetime(1970, 1, 1, tzinfo=timezone.utc)
 
 
 def list_sources(
@@ -193,7 +193,7 @@ def __init__(
         self,
         datasource: str,
         imstype: Optional[Union[str, IMSType]] = None,
-        tz: Union[tzinfo, str] = pytz.timezone("Europe/Oslo"),
+        tz: Union[tzinfo, str] = ZoneInfo("Europe/Oslo"),
         url: Optional[str] = None,
         handler_options: Dict[str, Union[int, float, str]] = {},  # noqa:
         verify_ssl: Optional[Union[bool, str]] = True,
@@ -210,10 +210,12 @@ def __init__(
                 )
 
         if isinstance(tz, str):
-            if tz in pytz.all_timezones:
-                self.tz = pytz.timezone(tz)
-            else:
-                raise ValueError(f"Invalid timezone string  Given type was {type(tz)}")
+            try:
+                self.tz = ZoneInfo(tz)
+            except ZoneInfoNotFoundError:
+                raise ValueError(
+                    f"Invalid timezone string. Given type was {type(tz)}"
+                ) from None
         elif isinstance(tz, tzinfo):
             self.tz = tz
         else:

tagreader/utils.py

@@ -7,10 +7,10 @@
 from enum import Enum
 from pathlib import Path
 from typing import Union
+from zoneinfo import ZoneInfo
 
 import certifi
 import pandas as pd
-import pytz
 import requests
 from platformdirs import user_data_dir
 
@@ -50,12 +50,12 @@ def convert_to_pydatetime(date_stamp: Union[datetime, str, pd.Timestamp]) -> dat
 
 def ensure_datetime_with_tz(
     date_stamp: Union[datetime, str, pd.Timestamp],
-    tz: tzinfo = pytz.timezone("Europe/Oslo"),
+    tz: tzinfo = ZoneInfo("Europe/Oslo"),
 ) -> datetime:
     date_stamp = convert_to_pydatetime(date_stamp)
 
     if not date_stamp.tzinfo:
-        date_stamp = tz.localize(date_stamp)
+        date_stamp = date_stamp.replace(tzinfo=tz)
 
     return date_stamp
 

tagreader/web_handlers.py

@@ -3,14 +3,13 @@
 import re
 import urllib.parse
 from abc import ABC, abstractmethod
-from datetime import datetime, timedelta
+from datetime import datetime, timedelta, timezone
 from hashlib import new as hashlib_new_method
 from json.decoder import JSONDecodeError
 from typing import Any, Dict, List, Optional, Tuple, Union
 
 import numpy as np
 import pandas as pd
-import pytz
 import requests
 import urllib3
 from Crypto.Hash import MD4 as _MD4
@@ -437,7 +436,7 @@ def _get_maps(self, tagname: str):
         return ret
 
     def _get_default_mapname(self, tagname: str):
-        (tagname, _) = self.split_tagmap(tagname)
+        tagname, _ = self.split_tagmap(tagname)
         all_maps = self._get_maps(tagname)
         for k, v in all_maps.items():
             if v:
@@ -729,7 +728,7 @@ def __init__(
 
     @staticmethod
     def _time_to_UTC_string(time: datetime) -> str:
-        return time.astimezone(pytz.UTC).strftime("%d-%b-%y %H:%M:%S")
+        return time.astimezone(timezone.utc).strftime("%d-%b-%y %H:%M:%S")
 
     @staticmethod
     def escape(s: str) -> str:
@@ -994,7 +993,7 @@ def read_tag(
         if not web_id:
             return pd.DataFrame()
 
-        (url, params) = self.generate_read_query(
+        url, params = self.generate_read_query(
             tag=web_id,
             start=start,
             end=end,

tests/test_AspenHandlerREST_connect.py

@@ -27,8 +27,8 @@
 SOURCE = "TRB"
 TAG = "AverageCPUTimeVals"
 FAKE_TAG = "so_random_it_cant_exist"
-START_TIME = datetime(2023, 5, 1, 10, 0, 0)
-STOP_TIME = datetime(2023, 5, 1, 11, 0, 0)
+START_TIME = datetime(2025, 5, 1, 10, 0, 0)
+STOP_TIME = datetime(2025, 5, 1, 11, 0, 0)
 SAMPLE_TIME = timedelta(seconds=60)
 
 

tests/test_PIHandlerREST.py

@@ -102,7 +102,7 @@ def test_generate_read_query(pi_handler: PIHandlerWeb, read_type: str) -> None:
     stop = ensure_datetime_with_tz(STOP_TIME)
     ts = timedelta(seconds=SAMPLE_TIME)
 
-    (url, params) = pi_handler.generate_read_query(
+    url, params = pi_handler.generate_read_query(
         tag=pi_handler.tag_to_web_id(tag="alreadyknowntag"),  # type: ignore[arg-type]
         start=start,
         end=stop,
@@ -176,7 +176,7 @@ def test_generate_read_query_with_status(
     stop = ensure_datetime_with_tz(STOP_TIME)
     ts = timedelta(seconds=SAMPLE_TIME)
 
-    (url, params) = pi_handler.generate_read_query(
+    url, params = pi_handler.generate_read_query(
         tag=pi_handler.tag_to_web_id("alreadyknowntag"),  # type: ignore[arg-type]
         start=start,
         end=stop,
@@ -234,7 +234,7 @@ def test_generate_read_query_long_sample_time(pi_handler: PIHandlerWeb) -> None:
     stop = ensure_datetime_with_tz(STOP_TIME)
     ts = timedelta(seconds=86410)
 
-    (url, params) = pi_handler.generate_read_query(
+    url, params = pi_handler.generate_read_query(
         tag=pi_handler.tag_to_web_id("alreadyknowntag"),  # type: ignore[arg-type]
         start=start,
         end=stop,

tests/test_bucketcache.py

@@ -541,9 +541,9 @@ def test_store_overlapping_df(cache: BucketCache) -> None:
     )
     df_expected = pd.concat(
         [
-            DF1[START_TIME_1 : START_TIME_3 - pd.Timedelta(TS, unit="s")],
+            DF1[START_TIME_1 : START_TIME_3 - TS],
             DF3[START_TIME_3:END_TIME_3],
-            DF2[END_TIME_3 + pd.Timedelta(TS, unit="s") : END_TIME_2],
+            DF2[END_TIME_3 + TS : END_TIME_2],
         ]
     )
 

tests/test_clients.py

@@ -1,8 +1,8 @@
 from datetime import datetime, timedelta
+from zoneinfo import ZoneInfo
 
 import pandas as pd
 import pytest
-import pytz
 
 from tagreader.clients import IMSClient, get_missing_intervals, get_next_timeslice
 from tagreader.utils import IMSType, ReaderType
@@ -21,32 +21,32 @@ def test_init_client_with_tzinfo() -> None:
         datasource="mock", imstype=IMSType.PIWEBAPI, cache=None, tz="US/Eastern"
     )
     print(client.tz)
-    assert client.tz == pytz.timezone("US/Eastern")
+    assert client.tz == ZoneInfo("US/Eastern")
 
     client = IMSClient(
         datasource="mock",
         imstype=IMSType.PIWEBAPI,
         cache=None,
-        tz=pytz.timezone("US/Eastern"),
+        tz=ZoneInfo("US/Eastern"),
     )
     print(client.tz)
-    assert client.tz == pytz.timezone("US/Eastern")
+    assert client.tz == ZoneInfo("US/Eastern")
 
     client = IMSClient(
         datasource="mock", imstype=IMSType.PIWEBAPI, cache=None, tz="Europe/Oslo"
     )
     print(client.tz)
-    assert client.tz == pytz.timezone("Europe/Oslo")
+    assert client.tz == ZoneInfo("Europe/Oslo")
 
     client = IMSClient(
         datasource="mock", imstype=IMSType.PIWEBAPI, cache=None, tz="US/Central"
     )
     print(client.tz)
-    assert client.tz == pytz.timezone("US/Central")
+    assert client.tz == ZoneInfo("US/Central")
 
     client = IMSClient(datasource="mock", imstype=IMSType.PIWEBAPI, cache=None)
     print(client.tz)
-    assert client.tz == pytz.timezone("Europe/Oslo")
+    assert client.tz == ZoneInfo("Europe/Oslo")
 
     with pytest.raises(ValueError):
         _ = IMSClient(
@@ -63,15 +63,15 @@ def test_init_client_with_datasource() -> None:
         datasource="mock", imstype=IMSType.PIWEBAPI, cache=None, tz="US/Eastern"
     )
     print(client.tz)
-    assert client.tz == pytz.timezone("US/Eastern")
+    assert client.tz == ZoneInfo("US/Eastern")
     client = IMSClient(
         datasource="mock", imstype=IMSType.PIWEBAPI, cache=None, tz="US/Central"
     )
     print(client.tz)
-    assert client.tz == pytz.timezone("US/Central")
+    assert client.tz == ZoneInfo("US/Central")
     client = IMSClient(datasource="mock", imstype=IMSType.PIWEBAPI, cache=None)
     print(client.tz)
-    assert client.tz == pytz.timezone("Europe/Oslo")
+    assert client.tz == ZoneInfo("Europe/Oslo")
     with pytest.raises(ValueError):
         _ = IMSClient(
             datasource="mock", imstype=IMSType.PIWEBAPI, cache=None, tz="WRONGVALUE"

tests/test_utils.py

@@ -1,9 +1,8 @@
 import datetime
 import os
+from zoneinfo import ZoneInfo
 
 import pandas as pd
-import pytz
-from pytz import timezone
 
 from tagreader.utils import ensure_datetime_with_tz, is_equinor, urljoin
 
@@ -12,37 +11,41 @@
 
 
 def test_ensure_is_datetime_string() -> None:
-    assert ensure_datetime_with_tz("10. jan. 2018 13:45:15") == timezone(
-        "Europe/Oslo"
-    ).localize(datetime.datetime(2018, 1, 10, 13, 45, 15))
-    assert ensure_datetime_with_tz("01.02.03 00:00:00") == timezone(
-        "Europe/Oslo"
-    ).localize(datetime.datetime(2003, 2, 1, 0, 0, 0))
+    assert ensure_datetime_with_tz("10. jan. 2018 13:45:15") == datetime.datetime(
+        2018, 1, 10, 13, 45, 15, tzinfo=ZoneInfo("Europe/Oslo")
+    )
+    assert ensure_datetime_with_tz("01.02.03 00:00:00") == datetime.datetime(
+        2003, 2, 1, 0, 0, 0, tzinfo=ZoneInfo("Europe/Oslo")
+    )
     assert ensure_datetime_with_tz("02.01.03 00:00:00") == ensure_datetime_with_tz(
         "2003-02-01 0:00:00am"
     )
     assert ensure_datetime_with_tz(
-        "02.01.03 00:00:00", pytz.timezone("America/Sao_Paulo")
-    ) == timezone("America/Sao_Paulo").localize(datetime.datetime(2003, 1, 2, 0, 0, 0))
+        "02.01.03 00:00:00", ZoneInfo("America/Sao_Paulo")
+    ) == datetime.datetime(2003, 1, 2, 0, 0, 0, tzinfo=ZoneInfo("America/Sao_Paulo"))
     assert ensure_datetime_with_tz(
-        "02.01.03 00:00:00", pytz.timezone("Brazil/East")
-    ) == timezone("Brazil/East").localize(datetime.datetime(2003, 1, 2, 0, 0, 0))
+        "02.01.03 00:00:00", ZoneInfo("America/Sao_Paulo")
+    ) == datetime.datetime(2003, 1, 2, 0, 0, 0, tzinfo=ZoneInfo("America/Sao_Paulo"))
     assert ensure_datetime_with_tz(
-        timezone("Brazil/East").localize(datetime.datetime(2003, 1, 2, 0, 0, 0)),
-        pytz.timezone("Brazil/East"),
-    ) == timezone("Brazil/East").localize(datetime.datetime(2003, 1, 2, 0, 0, 0))
+        datetime.datetime(2003, 1, 2, 0, 0, 0, tzinfo=ZoneInfo("America/Sao_Paulo")),
+        ZoneInfo("America/Sao_Paulo"),
+    ) == datetime.datetime(2003, 1, 2, 0, 0, 0, tzinfo=ZoneInfo("America/Sao_Paulo"))
 
 
 def test_ensure_is_datetime_pd_timestamp() -> None:
     ts = datetime.datetime(2018, 1, 10, 13, 45, 15)
-    ts_with_tz = timezone("Europe/Oslo").localize(ts)
+    ts_with_tz = datetime.datetime(
+        2018, 1, 10, 13, 45, 15, tzinfo=ZoneInfo("Europe/Oslo")
+    )
     assert ensure_datetime_with_tz(ts_with_tz) == ts_with_tz
     assert ensure_datetime_with_tz(ts) == ts_with_tz
 
 
 def test_ensure_is_datetime_datetime() -> None:
     dt = datetime.datetime(2018, 1, 10, 13, 45, 15)
-    dt_with_tz = timezone("Europe/Oslo").localize(dt)
+    dt_with_tz = datetime.datetime(
+        2018, 1, 10, 13, 45, 15, tzinfo=ZoneInfo("Europe/Oslo")
+    )
 
     assert ensure_datetime_with_tz(dt_with_tz) == dt_with_tz
     assert ensure_datetime_with_tz(dt) == dt_with_tz