Add num_hrs_to_fetch var

[infoex-autowx.git] / infoex-autowx.py

diff --git a/infoex-autowx.py b/infoex-autowx.py
index c5ec48d04637ac2610ab7607f2620cf1cfe260b2..2b7880eaf9865acad39b83c6a56d8b5e1a55f92c 100755 (executable)
--- a/infoex-autowx.py
+++ b/infoex-autowx.py
@@ -29,21 +29,27 @@ import logging
 import os
 import sys
 import time
 import os
 import sys
 import time

+import urllib3
+import importlib.util

 
 from ftplib import FTP
 from argparse import ArgumentParser
 
 
 from ftplib import FTP
 from argparse import ArgumentParser
 

+import pytz
+

 import requests
 
 import zeep
 import zeep.cache
 import zeep.transports
 
 import requests
 
 import zeep
 import zeep.cache
 import zeep.transports
 

-__version__ = '2.0.2'
+__version__ = '3.2.4'

 
 LOG = logging.getLogger(__name__)
 LOG.setLevel(logging.NOTSET)
 
 
 LOG = logging.getLogger(__name__)
 LOG.setLevel(logging.NOTSET)
 

+urllib3.disable_warnings()
+

 def get_parser():
     """Return OptionParser for this program"""
     parser = ArgumentParser()
 def get_parser():
     """Return OptionParser for this program"""
     parser = ArgumentParser()


@@ -85,7 +91,7 @@ def setup_config(config):
         station = dict()
         station['provider'] = config['station']['type']
 
         station = dict()
         station['provider'] = config['station']['type']
 

-        if station['provider'] not in ['nrcs', 'mesowest']:
+        if station['provider'] not in ['nrcs', 'mesowest', 'python']:

             print("Please specify either nrcs or mesowest as the station type.")
             sys.exit(1)
 
             print("Please specify either nrcs or mesowest as the station type.")
             sys.exit(1)
 


@@ -93,11 +99,7 @@ def setup_config(config):
             station['source'] = 'https://www.wcc.nrcs.usda.gov/awdbWebService/services?WSDL'
             station['station_id'] = config['station']['station_id']
             station['desired_data'] = config['station']['desired_data'].split(',')
             station['source'] = 'https://www.wcc.nrcs.usda.gov/awdbWebService/services?WSDL'
             station['station_id'] = config['station']['station_id']
             station['desired_data'] = config['station']['desired_data'].split(',')

-
-            # XXX: For NRCS, we're manually overriding units for now! Once
-            #      unit conversion is supported for NRCS, REMOVE THIS!
-            if 'units' not in station:
-                station['units'] = 'imperial'
+            station['units'] = config['station']['units']

 
         if station['provider'] == 'mesowest':
             station['source'] = 'https://api.synopticdata.com/v2/stations/timeseries'
 
         if station['provider'] == 'mesowest':
             station['source'] = 'https://api.synopticdata.com/v2/stations/timeseries'


@@ -112,9 +114,59 @@ def setup_config(config):
                                 '&stid=' + station['station_id'] + \
                                 '&vars=' + station['desired_data']
 
                                 '&stid=' + station['station_id'] + \
                                 '&vars=' + station['desired_data']
 

+        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)
+
+        # By default, fetch three hours of data
+        #
+        # If user wants hn24 or wind averaging, then
+        # we need more.
+        station['num_hrs_to_fetch'] = 3
+
+        # HN24
+        if 'hn24' in config['station']:
+            if config['station']['hn24'] not in ['true', 'false']:
+                raise ValueError("hn24 must be either 'true' or 'false'")
+
+            if config['station']['hn24'] == "true":
+                station['hn24'] = True
+                station['num_hrs_to_fetch'] = 24
+            else:
+                station['hn24'] = False
+        else:
+            # default to False
+            station['hn24'] = False
+
+        # Wind mode
+        if 'wind_mode' in config['station']:
+            if config['station']['wind_mode'] not in ['normal', 'average']:
+                raise ValueError("wind_mode must be either 'normal' or 'average'")
+
+            station['wind_mode'] = config['station']['wind_mode']
+
+            if station['wind_mode'] == "average":
+                station['num_hrs_to_fetch'] = 24
+        else:
+            # default to False
+            station['wind_mode'] = "normal"
+

     except KeyError as err:
         LOG.critical("%s not defined in configuration file", err)
     except KeyError as err:
         LOG.critical("%s not defined in configuration file", err)

-        exit(1)
+        sys.exit(1)
+    except ValueError as err:
+        LOG.critical("%s", err)
+        sys.exit(1)

 
     # all sections/values present in config file, final sanity check
     try:
 
     # all sections/values present in config file, final sanity check
     try:


@@ -124,7 +176,7 @@ def setup_config(config):
                     raise ValueError
     except ValueError:
         LOG.critical("Config value '%s.%s' is empty", key, subkey)
                     raise ValueError
     except ValueError:
         LOG.critical("Config value '%s.%s' is empty", key, subkey)

-        exit(1)
+        sys.exit(1)

 
     return (infoex, station)
 
 
     return (infoex, station)
 


@@ -139,6 +191,11 @@ def setup_logging(log_level):
         #LOG.addHandler(logging.handlers.SysLogHandler())
         # fallback to stdout
         handler = logging.StreamHandler(sys.stdout)
         #LOG.addHandler(logging.handlers.SysLogHandler())
         # fallback to stdout
         handler = logging.StreamHandler(sys.stdout)

+        formatter = logging.Formatter('%(asctime)s.%(msecs)03d '
+                                      '%(levelname)s %(module)s - '
+                                      '%(funcName)s: %(message)s',
+                                      '%Y-%m-%d %H:%M:%S')
+        handler.setFormatter(formatter)

         LOG.addHandler(handler)
 
     # ugly, but passable
         LOG.addHandler(handler)
 
     # ugly, but passable


@@ -184,14 +241,17 @@ def main():
     iemap = setup_infoex_counterparts_mapping(station['provider'])
 
     # override units if user selected metric
     iemap = setup_infoex_counterparts_mapping(station['provider'])
 
     # override units if user selected metric

-    if station['units'] == 'metric':
+    if station['provider'] != 'python' and station['units'] == 'metric':

         final_data = switch_units_to_metric(final_data, fmap)
 
         final_data = switch_units_to_metric(final_data, fmap)
 

-    (begin_date, end_date) = setup_time_values()
+    (begin_date, end_date) = setup_time_values(station)

 
 

-    # get the data
-    LOG.debug("Getting %s data from %s to %s", str(station['desired_data']),
-              str(begin_date), str(end_date))
+    if station['provider'] == 'python':
+        LOG.debug("Getting custom data from external Python program")
+    else:
+        LOG.debug("Getting %s data from %s to %s (%s)",
+                  str(station['desired_data']),
+                  str(begin_date), str(end_date), end_date.tzinfo.zone)

 
     time_all_elements = time.time()
 
 
     time_all_elements = time.time()
 


@@ -201,23 +261,80 @@ def main():
     elif station['provider'] == 'mesowest':
         infoex['wx_data'] = get_mesowest_data(begin_date, end_date,
                                               station)
     elif station['provider'] == 'mesowest':
         infoex['wx_data'] = get_mesowest_data(begin_date, end_date,
                                               station)

+    elif station['provider'] == 'python':
+        try:
+            spec = importlib.util.spec_from_file_location('custom_wx',
+                                                          station['path'])
+            mod = importlib.util.module_from_spec(spec)
+            spec.loader.exec_module(mod)
+            mod.LOG = LOG
+
+            try:
+                infoex['wx_data'] = mod.get_custom_data()
+
+                if infoex['wx_data'] is None:
+                    infoex['wx_data'] = []
+            except Exception as exc:
+                LOG.error("Python program for custom Wx data failed in "
+                          "execution: %s", str(exc))
+                sys.exit(1)
+
+            LOG.info("Successfully executed external Python program")
+        except ImportError:
+            LOG.error("Please upgrade to Python 3.3 or later")
+            sys.exit(1)
+        except FileNotFoundError:
+            LOG.error("Specified Python program for custom Wx data "
+                      "was not found")
+            sys.exit(1)
+        except Exception as exc:
+            LOG.error("A problem was encountered when attempting to "
+                      "load your custom Wx program: %s", str(exc))
+            sys.exit(1)

 
     LOG.info("Time taken to get all data : %.3f sec", time.time() -
              time_all_elements)
 
     LOG.debug("infoex[wx_data]: %s", str(infoex['wx_data']))
 
 
     LOG.info("Time taken to get all data : %.3f sec", time.time() -
              time_all_elements)
 
     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']
     # 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')
+    final_data[fmap['timeZone']] = station['tz'].zone

 
     for element_cd in infoex['wx_data']:
         if element_cd not in iemap:
             LOG.warning("BAD KEY wx_data['%s']", element_cd)
             continue
 
 
     for element_cd in infoex['wx_data']:
         if element_cd not in iemap:
             LOG.warning("BAD KEY wx_data['%s']", element_cd)
             continue
 

+        if infoex['wx_data'][element_cd] is None:
+            continue
+
+        # do the conversion before the rounding
+        if station['provider'] == 'nrcs' and station['units'] == 'metric':
+            infoex['wx_data'][element_cd] = convert_nrcs_units_to_metric(element_cd, infoex['wx_data'][element_cd])
+
+        # Massage precision of certain values to fit InfoEx's
+        # expectations
+        #
+        # 0 decimal places: relative humidity, wind speed, wind
+        #                   direction, wind gust, snow depth
+        # 1 decimal place:  air temp, baro
+        # Avoid transforming None values
+        if element_cd in ['wind_speed', 'WSPD', 'wind_direction',
+                          'RHUM', 'relative_humidity', 'WDIR',
+                          'wind_gust', 'SNWD', 'snow_depth']:
+            infoex['wx_data'][element_cd] = round(infoex['wx_data'][element_cd])
+        elif element_cd in ['TOBS', 'air_temp', 'PRES', 'pressure']:
+            infoex['wx_data'][element_cd] = round(infoex['wx_data'][element_cd], 1)
+        elif element_cd in ['PREC', 'precip_accum']:
+            infoex['wx_data'][element_cd] = round(infoex['wx_data'][element_cd], 2)
+

         # CONSIDER: Casting every value to Float() -- need to investigate if
         #           any possible elementCds we may want are any other data
         #           type than float.
         # CONSIDER: Casting every value to Float() -- need to investigate if
         #           any possible elementCds we may want are any other data
         #           type than float.


@@ -231,13 +348,14 @@ def main():
 
     LOG.debug("final_data: %s", str(final_data))
 
 
     LOG.debug("final_data: %s", str(final_data))
 

-    if not write_local_csv(infoex['csv_filename'], final_data):
-        LOG.warning('Could not write local CSV file: %s',
-                    infoex['csv_filename'])
-        return 1
+    if infoex['wx_data']:
+        if not write_local_csv(infoex['csv_filename'], final_data):
+            LOG.warning('Could not write local CSV file: %s',
+                        infoex['csv_filename'])
+            return 1

 
 

-    if not options.dry_run:
-        upload_csv(infoex['csv_filename'], infoex)
+        if not options.dry_run:
+            upload_csv(infoex['csv_filename'], infoex)

 
     LOG.debug('DONE')
     return 0
 
     LOG.debug('DONE')
     return 0


@@ -302,6 +420,8 @@ def setup_infoex_counterparts_mapping(provider):
     if provider == 'nrcs':
         iemap['PREC'] = 'precipitationGauge'
         iemap['TOBS'] = 'tempPres'
     if provider == 'nrcs':
         iemap['PREC'] = 'precipitationGauge'
         iemap['TOBS'] = 'tempPres'

+        iemap['TMAX'] = 'tempMaxHour'
+        iemap['TMIN'] = 'tempMinHour'

         iemap['SNWD'] = 'hS'
         iemap['PRES'] = 'baro'
         iemap['RHUM'] = 'rH'
         iemap['SNWD'] = 'hS'
         iemap['PRES'] = 'baro'
         iemap['RHUM'] = 'rH'


@@ -312,12 +432,26 @@ def setup_infoex_counterparts_mapping(provider):
     elif provider == 'mesowest':
         iemap['precip_accum'] = 'precipitationGauge'
         iemap['air_temp'] = 'tempPres'
     elif provider == 'mesowest':
         iemap['precip_accum'] = 'precipitationGauge'
         iemap['air_temp'] = 'tempPres'

+        iemap['air_temp_high_24_hour'] = 'tempMaxHour'
+        iemap['air_temp_low_24_hour'] = 'tempMinHour'

         iemap['snow_depth'] = 'hS'
         iemap['pressure'] = 'baro'
         iemap['relative_humidity'] = 'rH'
         iemap['wind_speed'] = 'windSpeedNum'
         iemap['wind_direction'] = 'windDirectionNum'
         iemap['wind_gust'] = 'windGustSpeedNum'
         iemap['snow_depth'] = 'hS'
         iemap['pressure'] = 'baro'
         iemap['relative_humidity'] = 'rH'
         iemap['wind_speed'] = 'windSpeedNum'
         iemap['wind_direction'] = 'windDirectionNum'
         iemap['wind_gust'] = 'windGustSpeedNum'

+    elif provider == 'python':
+        # we expect Python programs to use the InfoEx data type names
+        iemap['precipitationGauge'] = 'precipitationGauge'
+        iemap['tempPres'] = 'tempPres'
+        iemap['tempMaxHour'] = 'tempMaxHour'
+        iemap['tempMinHour'] = 'tempMinHour'
+        iemap['hS'] = 'hS'
+        iemap['baro'] = 'baro'
+        iemap['rH'] = 'rH'
+        iemap['windSpeedNum'] = 'windSpeedNum'
+        iemap['windDirectionNum'] = 'windDirectionNum'
+        iemap['windGustSpeedNum'] = 'windGustSpeedNum'

 
     return iemap
 
 
     return iemap
 


@@ -325,9 +459,14 @@ def setup_infoex_counterparts_mapping(provider):
 def get_nrcs_data(begin, end, station):
     """get the data we're after from the NRCS WSDL"""
     transport = zeep.transports.Transport(cache=zeep.cache.SqliteCache())
 def get_nrcs_data(begin, end, station):
     """get the data we're after from the NRCS WSDL"""
     transport = zeep.transports.Transport(cache=zeep.cache.SqliteCache())

+    transport.session.verify = False

     client = zeep.Client(wsdl=station['source'], transport=transport)
     remote_data = {}
 
     client = zeep.Client(wsdl=station['source'], transport=transport)
     remote_data = {}
 

+    # massage begin/end date format
+    begin_date_str = begin.strftime('%Y-%m-%d %H:%M:00')
+    end_date_str = end.strftime('%Y-%m-%d %H:%M:00')
+

     for element_cd in station['desired_data']:
         time_element = time.time()
 
     for element_cd in station['desired_data']:
         time_element = time.time()
 


@@ -336,8 +475,8 @@ def get_nrcs_data(begin, end, station):
             stationTriplets=[station['station_id']],
             elementCd=element_cd,
             ordinal=1,
             stationTriplets=[station['station_id']],
             elementCd=element_cd,
             ordinal=1,

-            beginDate=begin,
-            endDate=end)
+            beginDate=begin_date_str,
+            endDate=end_date_str)

 
         LOG.info("Time to get NRCS elementCd '%s': %.3f sec", element_cd,
                  time.time() - time_element)
 
         LOG.info("Time to get NRCS elementCd '%s': %.3f sec", element_cd,
                  time.time() - time_element)


@@ -367,7 +506,12 @@ def get_mesowest_data(begin, end, station):
 
     # construct final, completed API URL
     api_req_url = station['source'] + '&start=' + begin_date_str + '&end=' + end_date_str
 
     # construct final, completed API URL
     api_req_url = station['source'] + '&start=' + begin_date_str + '&end=' + end_date_str

-    req = requests.get(api_req_url)
+
+    try:
+        req = requests.get(api_req_url)
+    except requests.exceptions.ConnectionError:
+        LOG.error("Could not connect to '%s'", api_req_url)
+        sys.exit(1)

 
     try:
         json = req.json()
 
     try:
         json = req.json()


@@ -377,8 +521,19 @@ def get_mesowest_data(begin, end, station):
 
     try:
         observations = json['STATION'][0]['OBSERVATIONS']
 
     try:
         observations = json['STATION'][0]['OBSERVATIONS']

-    except ValueError:
-        LOG.error("Bad JSON in MesoWest response")
+    except KeyError as exc:
+        LOG.error("Unexpected JSON in MesoWest response: '%s'", exc)
+        sys.exit(1)
+    except IndexError as exc:
+        LOG.error("Unexpected JSON in MesoWest response: '%s'", exc)
+        try:
+            LOG.error("Detailed MesoWest response: '%s'",
+                      json['SUMMARY']['RESPONSE_MESSAGE'])
+        except KeyError:
+            pass
+        sys.exit(1)
+    except ValueError as exc:
+        LOG.error("Bad JSON in MesoWest response: '%s'", exc)

         sys.exit(1)
 
     pos = len(observations['date_time']) - 1
         sys.exit(1)
 
     pos = len(observations['date_time']) - 1


@@ -399,9 +554,20 @@ def get_mesowest_data(begin, end, station):
 
         # we may not have the data at all
         key_name = element_cd + '_set_1'
 
         # we may not have the data at all
         key_name = element_cd + '_set_1'

+

         if key_name in observations:
             if observations[key_name][pos]:
                 remote_data[element_cd] = observations[key_name][pos]
         if key_name in observations:
             if observations[key_name][pos]:
                 remote_data[element_cd] = observations[key_name][pos]

+
+                # mesowest by default provides wind_speed in m/s, but
+                # we specify 'english' units in the request; either way,
+                # we want mph
+                if element_cd in ('wind_speed', 'wind_gust'):
+                    remote_data[element_cd] = kn_to_mph(remote_data[element_cd])
+
+                # mesowest provides HS in mm, not cm; we want cm
+                if element_cd == 'snow_depth' and station['units'] == 'metric':
+                    remote_data[element_cd] = mm_to_cm(remote_data[element_cd])

             else:
                 remote_data[element_cd] = None
         else:
             else:
                 remote_data[element_cd] = None
         else:


@@ -414,18 +580,29 @@ def switch_units_to_metric(data_map, mapping):
 
     # NOTE: to update this, use the fmap<->final_data mapping laid out
     #       in setup_infoex_fields_mapping ()
 
     # NOTE: to update this, use the fmap<->final_data mapping laid out
     #       in setup_infoex_fields_mapping ()

-    #
-    # NOTE: this only 'works' with MesoWest for now, as the MesoWest API
-    #       itself handles the unit conversion; in the future, we will also
-    #       support NRCS unit conversion, but this must be done by this
-    #       program.
+    data_map[mapping['tempMaxHourUnit']] = 'C'
+    data_map[mapping['tempMinHourUnit']] = 'C'

     data_map[mapping['tempPresUnit']] = 'C'
     data_map[mapping['tempPresUnit']] = 'C'

-    data_map[mapping['hsUnit']] = 'm'
+    data_map[mapping['precipitationGaugeUnit']] = 'mm'
+    data_map[mapping['hsUnit']] = 'cm'

     data_map[mapping['windSpeedUnit']] = 'm/s'
     data_map[mapping['windGustSpeedNumUnit']] = 'm/s'
     data_map[mapping['windSpeedUnit']] = 'm/s'
     data_map[mapping['windGustSpeedNumUnit']] = 'm/s'

+    data_map[mapping['dewPointUnit']] = 'C'
+    data_map[mapping['hn24AutoUnit']] = 'cm'
+    data_map[mapping['hstAutoUnit']] = 'cm'

 
     return data_map
 
 
     return data_map
 

+def convert_nrcs_units_to_metric(element_cd, value):
+    """convert NRCS values from English to metric"""
+    if element_cd == 'TOBS':
+        value = f_to_c(value)
+    elif element_cd == 'SNWD':
+        value = in_to_cm(value)
+    elif element_cd == 'PREC':
+        value = in_to_mm(value)
+    return value
+

 # CSV operations
 def write_local_csv(path_to_file, data):
     """Write the specified CSV file to disk"""
 # CSV operations
 def write_local_csv(path_to_file, data):
     """Write the specified CSV file to disk"""


@@ -450,15 +627,47 @@ def upload_csv(path_to_file, infoex_data):
     os.remove(path_to_file)
 
 # other miscellaneous routines
     os.remove(path_to_file)
 
 # other miscellaneous routines

-def setup_time_values():
+def setup_time_values(station):

     """establish time bounds of data request(s)"""
     """establish time bounds of data request(s)"""

+
+    # default timezone to UTC (for MesoWest)
+    tz = pytz.utc
+
+    # but for NRCS, use the config-specified timezone
+    if station['provider'] == 'nrcs':
+        tz = station['tz']
+

     # floor time to nearest hour
     # floor time to nearest hour

-    date_time = datetime.datetime.now()
+    date_time = datetime.datetime.now(tz=tz)

     end_date = date_time - datetime.timedelta(minutes=date_time.minute % 60,
                                               seconds=date_time.second,
                                               microseconds=date_time.microsecond)
     end_date = date_time - datetime.timedelta(minutes=date_time.minute % 60,
                                               seconds=date_time.second,
                                               microseconds=date_time.microsecond)

-    begin_date = end_date - datetime.timedelta(hours=3)
+    begin_date = end_date - datetime.timedelta(hours=station['num_hrs_to_fetch'])

     return (begin_date, end_date)
 
     return (begin_date, end_date)
 

+def f_to_c(f):
+    """convert Fahrenheit to Celsius"""
+    return (float(f) - 32) * 5.0/9.0
+
+def in_to_cm(inches):
+    """convert inches to centimetrs"""
+    return float(inches) * 2.54
+
+def in_to_mm(inches):
+    """convert inches to millimeters"""
+    return (float(inches) * 2.54) * 10.0
+
+def ms_to_mph(ms):
+    """convert meters per second to miles per hour"""
+    return ms * 2.236936
+
+def kn_to_mph(kn):
+    """convert knots to miles per hour"""
+    return kn * 1.150779
+
+def mm_to_cm(mm):
+    """convert millimeters to centimetrs"""
+    return mm / 10.0
+

 if __name__ == "__main__":
     sys.exit(main())
 if __name__ == "__main__":
     sys.exit(main())