From 4d4c7019aa2063a0b8742949021a028a3a96b795 Mon Sep 17 00:00:00 2001
From: Alexander Vasarab <alexander@wylark.com>
Date: Thu, 7 Jan 2021 21:09:25 -0800
Subject: [PATCH] Implement time zone support

---
 README.md                   | 27 +++++++++++++++++++++++++++
 examples/config.example.ini |  1 +
 infoex-autowx.py            | 35 ++++++++++++++++++++++++++---------
 requirements.txt            |  1 +
 4 files changed, 55 insertions(+), 9 deletions(-)

diff --git a/README.md b/README.md
index bebbdcb..62d15bc 100644
--- a/README.md
+++ b/README.md
@@ -83,6 +83,7 @@ options.
 `station_id = # the NRCS/MesoWest identifier for a particular station #`  
 `desired_data = # a comma-delimited list of fields you're interested in #`  
 `units = # either english or metric -- only applies when type is mesowest #`  
+`tz = # any entry from the Olson tz database e.g. America/Denver #`  
 `path = # the filesystem path to the Python program -- only applies when type is python #`  
 
 `[infoex]`  
@@ -206,6 +207,32 @@ or a remote web page which requires some custom parsing.
 Please see the program located at examples/custom-wx.example.py for a
 complete description of what's required.
 
+A note on time zones
+--------------------
+
+This program is aware of time zones via the pytz library. The way in
+which NRCS and MesoWest deal with time zones differs as follows:
+
+NRCS expects the request to come in the appropriate time zone, and the
+data retrieved will be in the same time zone (no transformation
+required before sending to InfoEx).
+
+MesoWest expects the request to come in UTC, and the data retrieved will
+be in the same time zone (transformation from UTC to the desired time
+zone is required before sending to InfoEx).
+
+As long as you specify the correct timezone in your configuration file,
+all will be handled correctly. The list of time zones comes from the
+Olson tz database. See that for more information.
+
+If you specify an invalid time zone, the program will exit and inform
+you of such.
+
+Lastly, InfoEx itself is timezone aware. If you notice that the data
+which makes it into your operation is inaccurate, start your
+investigation with time zone-related issues and move on only once you've
+ruled this out as a cause of the inaccuracy.
+
 A note on supported measurements
 --------------------------------
 
diff --git a/examples/config.example.ini b/examples/config.example.ini
index 17544ae..0408788 100644
--- a/examples/config.example.ini
+++ b/examples/config.example.ini
@@ -4,6 +4,7 @@ token = # MesoWest API token (ignored for NRCS) #
 station_id = # NRCS/MesoWest Station ID #
 desired_data = # Comma-separated list of values #
 units = # (english|metric) (ignored for NRCS) #
+tz = # any Olson tz database entries e.g. America/Denver #
 
 [infoex]
 host = # InfoEx FTP host address #
diff --git a/infoex-autowx.py b/infoex-autowx.py
index dfb22a7..15adf08 100755
--- a/infoex-autowx.py
+++ b/infoex-autowx.py
@@ -33,6 +33,8 @@ import time
 from ftplib import FTP
 from argparse import ArgumentParser
 
+import pytz
+
 import requests
 
 import zeep
@@ -115,6 +117,17 @@ def setup_config(config):
         if station['provider'] == 'python':
             station['path'] = config['station']['path']
 
+        tz = 'America/Los_Angeles'
+
+        if 'tz' in config['station']:
+            tz = config['station']['tz']
+
+        try:
+            station['tz'] = pytz.timezone(tz)
+        except pytz.exceptions.UnknownTimeZoneError:
+            LOG.critical("%s is not a valid timezone", tz)
+            sys.exit(1)
+
     except KeyError as err:
         LOG.critical("%s not defined in configuration file", err)
         sys.exit(1)
@@ -196,14 +209,14 @@ def main():
                       "file")
             sys.exit(1)
 
-    (begin_date, end_date) = setup_time_values(station['provider'])
+    (begin_date, end_date) = setup_time_values(station)
 
     if station['provider'] == 'python':
         LOG.debug("Getting custom data from external Python program")
     else:
-        LOG.debug("Getting %s data from %s to %s",
+        LOG.debug("Getting %s data from %s to %s (%s)",
                   str(station['desired_data']),
-                  str(begin_date), str(end_date))
+                  str(begin_date), str(end_date), end_date.tzinfo.zone)
 
     time_all_elements = time.time()
 
@@ -251,11 +264,14 @@ def main():
 
     LOG.debug("infoex[wx_data]: %s", str(infoex['wx_data']))
 
+    # timezone massaging
+    final_end_date = end_date.astimezone(station['tz'])
+
     # Now we only need to add in what we want to change thanks to that
     # abomination of a variable declaration earlier
     final_data[fmap['Location UUID']] = infoex['location_uuid']
-    final_data[fmap['obDate']] = end_date.strftime('%m/%d/%Y')
-    final_data[fmap['obTime']] = end_date.strftime('%H:%M')
+    final_data[fmap['obDate']] = final_end_date.strftime('%m/%d/%Y')
+    final_data[fmap['obTime']] = final_end_date.strftime('%H:%M')
 
     for element_cd in infoex['wx_data']:
         if element_cd not in iemap:
@@ -537,14 +553,15 @@ def upload_csv(path_to_file, infoex_data):
     os.remove(path_to_file)
 
 # other miscellaneous routines
-def setup_time_values(provider):
+def setup_time_values(station):
     """establish time bounds of data request(s)"""
 
     # default timezone to UTC (for MesoWest)
-    tz = datetime.timezone.utc
+    tz = pytz.utc
 
-    if provider == 'nrcs':
-        tz = None
+    # but for NRCS, use the config-specified timezone
+    if station['provider'] == 'nrcs':
+        tz = station['tz']
 
     # floor time to nearest hour
     date_time = datetime.datetime.now(tz=tz)
diff --git a/requirements.txt b/requirements.txt
index 2aec4ac..cb9dc30 100644
--- a/requirements.txt
+++ b/requirements.txt
@@ -1,2 +1,3 @@
+pytz>=2020.1
 requests>=2.0.0
 zeep>=3.4.0
-- 
2.30.2