infoex-autowx.py

   1 #!/usr/bin/env python
   2 # -*- coding: utf-8 -*-
   3
   4 """
   5 InfoEx <-> NRCS/MesoWest Auto Wx implementation
   6 Alexander Vasarab
   7 Wylark Mountaineering LLC
   8
   9 This program fetches data from either an NRCS SNOTEL site or MesoWest
  10 weather station and pushes it to InfoEx using the new automated weather
  11 system implementation.
  12
  13 It is designed to be run hourly, and it asks for the last three hours
  14 of data of each desired type, and selects the most recent one. This
  15 lends some resiliency to the process and helps ensure that we have a
  16 value to send, but it can lead to somewhat inconsistent/untruthful
  17 data if e.g. the HS is from the last hour but the tempPres is from two
  18 hours ago because the instrumentation had a hiccup. It's worth
  19 considering if this is a bug or a feature.
  20
  21 For more information, see file: README
  22 For licensing, see file: LICENSE
  23 """
  24
  25 import configparser
  26 import csv
  27 import datetime
  28 import logging
  29 import os
  30 import sys
  31 import time
  32
  33 from ftplib import FTP
  34 from optparse import OptionParser
  35
  36 import requests
  37
  38 import zeep
  39 import zeep.cache
  40 import zeep.transports
  41
  42 __version__ = '2.0.0'
  43
  44 LOG = logging.getLogger(__name__)
  45 LOG.setLevel(logging.NOTSET)
  46
  47 def get_parser():
  48     """Return OptionParser for this program"""
  49     parser = OptionParser(version=__version__)
  50
  51     parser.add_option("--config",
  52         dest="config",
  53         metavar="FILE",
  54         help="location of config file")
  55
  56     parser.add_option("--log-level",
  57         dest="log_level",
  58         default=None,
  59         help="set the log level (debug, info, warning)")
  60
  61     parser.add_option("--dry-run",
  62         action="store_true",
  63         dest="dry_run",
  64         default=False,
  65         help="fetch data but don't upload to InfoEx")
  66
  67     return parser
  68
  69 def setup_config(config):
  70     """Setup config variable based on values specified in the ini file"""
  71     try:
  72         infoex = {
  73             'host': config['infoex']['host'],
  74             'uuid': config['infoex']['uuid'],
  75             'api_key': config['infoex']['api_key'],
  76             'csv_filename': config['infoex']['csv_filename'],
  77             'location_uuid': config['infoex']['location_uuid'],
  78             'wx_data': {}, # placeholder key, values to come later
  79         }
  80
  81         data = dict()
  82         data['provider'] = config['station']['type']
  83
  84         if data['provider'] not in ['nrcs', 'mesowest']:
  85             print("Please specify either nrcs or mesowest as the station type.")
  86             sys.exit(1)
  87
  88         if data['provider'] == 'nrcs':
  89             data['source'] = 'https://www.wcc.nrcs.usda.gov/awdbWebService/services?WSDL'
  90             data['station_id'] = config['station']['station_id']
  91
  92             try:
  93                 data['desired_data'] = config['station']['desired_data'].split(',')
  94             except:
  95                 # desired_data malformed or missing, setting default
  96                 data['desired_data'] = [
  97                                        'TOBS', # AIR TEMPERATURE OBSERVED (degF)
  98                                        'SNWD', # SNOW DEPTH (in)
  99                                        'PREC'  # PRECIPITATION ACCUMULATION (in)
 100                                        ]
 101
 102             # XXX: For NRCS, we're manually overriding units for now! Once
 103             #      unit conversion is supported for NRCS, REMOVE THIS!
 104             if 'units' not in data:
 105                 data['units'] = 'imperial'
 106
 107         if data['provider'] == 'mesowest':
 108             data['source'] = 'https://api.synopticdata.com/v2/stations/timeseries'
 109             data['station_id'] = config['station']['station_id']
 110             data['units'] = config['station']['units']
 111
 112             try:
 113                 data['desired_data'] = config['station']['desired_data']
 114             except:
 115                 # desired_data malformed or missing, setting default
 116                 data['desired_data'] = 'air_temp,snow_depth'
 117
 118             # construct full API URL (sans start/end time, added later)
 119             data['source'] = data['source'] + '?token=' + config['station']['token'] + '&within=60&units=' + data['units'] + '&stid=' + data['station_id'] + '&vars=' + data['desired_data']
 120
 121     except KeyError as e:
 122         LOG.critical("%s not defined in %s" % (e, options.config))
 123         exit(1)
 124     except Exception as exc:
 125         LOG.critical("Exception occurred in config parsing: '%s'" % (exc))
 126         exit(1)
 127
 128     # all sections/values present in config file, final sanity check
 129     try:
 130         for key in config.sections():
 131             for subkey in config[key]:
 132                 if not len(config[key][subkey]):
 133                     raise ValueError;
 134     except ValueError as exc:
 135         LOG.critical("Config value '%s.%s' is empty" % (key, subkey))
 136         exit(1)
 137
 138     return (infoex, data)
 139
 140 def setup_logging(log_level):
 141     """Setup our logging infrastructure"""
 142     try:
 143         from systemd.journal import JournalHandler
 144         LOG.addHandler(JournalHandler())
 145     except:
 146         ## fallback to syslog
 147         #import logging.handlers
 148         #LOG.addHandler(logging.handlers.SysLogHandler())
 149         # fallback to stdout
 150         handler = logging.StreamHandler(sys.stdout)
 151         LOG.addHandler(handler)
 152
 153     # ugly, but passable
 154     if log_level in [None, 'debug', 'info', 'warning']:
 155         if log_level == 'debug':
 156             LOG.setLevel(logging.DEBUG)
 157         elif log_level == 'info':
 158             LOG.setLevel(logging.INFO)
 159         elif log_level == 'warning':
 160             LOG.setLevel(logging.WARNING)
 161         else:
 162             LOG.setLevel(logging.NOTSET)
 163     else:
 164         return False
 165
 166     return True
 167
 168 def main():
 169     """Main routine: sort through args, decide what to do, then do it"""
 170     parser = get_parser()
 171     (options, args) = parser.parse_args()
 172
 173     config = configparser.ConfigParser(allow_no_value=False)
 174
 175     if not options.config:
 176         parser.print_help()
 177         print("\nPlease specify a configuration file via --config.")
 178         sys.exit(1)
 179
 180     config.read(options.config)
 181
 182     if not setup_logging(options.log_level):
 183         parser.print_help()
 184         print("\nPlease select an appropriate log level or remove the switch (--log-level).")
 185         sys.exit(1)
 186
 187     (infoex, data) = setup_config(config)
 188
 189     LOG.debug('Config parsed, starting up')
 190
 191     # create mappings
 192     (fmap, final_data) = setup_infoex_fields_mapping(infoex['location_uuid'])
 193     iemap = setup_infoex_counterparts_mapping(data['provider'])
 194
 195     # override units if user selected metric
 196     #
 197     # NOTE: to update this, use the fmap<->final_data mapping laid out above
 198     #
 199     # NOTE: this only 'works' with MesoWest for now, as the MesoWest API
 200     #       itself handles the unit conversion; in the future, we will also
 201     #       support NRCS unit conversion, but this must be done by this
 202     #       program.
 203     if data['units'] == 'metric':
 204         final_data[fmap['tempPresUnit']] = 'C'
 205         final_data[fmap['hsUnit']] = 'm'
 206         final_data[fmap['windSpeedUnit']] = 'm/s'
 207         final_data[fmap['windGustSpeedNumUnit']] = 'm/s'
 208
 209     # floor time to nearest hour
 210     dt = datetime.datetime.now()
 211     end_date = dt - datetime.timedelta(minutes=dt.minute % 60,
 212                                        seconds=dt.second,
 213                                        microseconds=dt.microsecond)
 214     begin_date = end_date - datetime.timedelta(hours=3)
 215
 216     # get the data
 217     LOG.debug("Getting %s data from %s to %s" % (str(data['desired_data']),
 218         str(begin_date), str(end_date)))
 219
 220     time_all_elements = time.time()
 221
 222     # NRCS-specific code
 223     if data['provider'] == 'nrcs':
 224         transport = zeep.transports.Transport(cache=zeep.cache.SqliteCache())
 225         client = zeep.Client(wsdl=data['source'], transport=transport)
 226
 227         for elementCd in data['desired_data']:
 228             time_element = time.time()
 229
 230             # get the last three hours of data for this elementCd
 231             tmp = client.service.getHourlyData(
 232                     stationTriplets=[data['station_id']],
 233                     elementCd=elementCd,
 234                     ordinal=1,
 235                     beginDate=begin_date,
 236                     endDate=end_date)
 237
 238             LOG.info("Time to get elementCd '%s': %.3f sec" % (elementCd,
 239                 time.time() - time_element))
 240
 241             values = tmp[0]['values']
 242
 243             # sort and isolate the most recent
 244             #
 245             # NOTE: we do this because sometimes there are gaps in hourly data
 246             #       in NRCS; yes, we may end up with slightly inaccurate data,
 247             #       so perhaps this decision will be re-evaluated in the future
 248             if values:
 249                 ordered = sorted(values, key=lambda t: t['dateTime'], reverse=True)
 250                 infoex['wx_data'][elementCd] = ordered[0]['value']
 251             else:
 252                 infoex['wx_data'][elementCd] = None
 253
 254     # MesoWest-specific code
 255     elif data['provider'] == 'mesowest':
 256         # massage begin/end date format
 257         begin_date_str = begin_date.strftime('%Y%m%d%H%M')
 258         end_date_str = end_date.strftime('%Y%m%d%H%M')
 259
 260         # construct final, completed API URL
 261         api_req_url = data['source'] + '&start=' + begin_date_str + '&end=' + end_date_str
 262         req = requests.get(api_req_url)
 263
 264         try:
 265             json = req.json()
 266         except ValueError:
 267             LOG.error("Bad JSON in MesoWest response")
 268             sys.exit(1)
 269
 270         try:
 271             observations = json['STATION'][0]['OBSERVATIONS']
 272         except ValueError:
 273             LOG.error("Bad JSON in MesoWest response")
 274             sys.exit(1)
 275
 276         pos = len(observations['date_time']) - 1
 277
 278         for elementCd in data['desired_data'].split(','):
 279             # sort and isolate the most recent, see note above in NRCS for how and
 280             # why this is done
 281             #
 282             # NOTE: Unlike in the NRCS case, the MesoWest API response contains all
 283             #       data (whereas with NRCS, we have to make a separate request for
 284             #       each element we want). This is nice for network efficiency but
 285             #       it means we have to handle this part differently for each.
 286             #
 287             # NOTE: Also unlike NRCS, MesoWest provides more granular data; NRCS
 288             #       provides hourly data, but MesoWest can often provide data every
 289             #       10 minutes -- though this provides more opportunity for
 290             #       irregularities
 291
 292             # we may not have the data at all
 293             key_name = elementCd + '_set_1'
 294             if key_name in observations:
 295                 if observations[key_name][pos]:
 296                     infoex['wx_data'][elementCd] = observations[key_name][pos]
 297                 else:
 298                     infoex['wx_data'][elementCd] = None
 299             else:
 300                 infoex['wx_data'][elementCd] = None
 301
 302     LOG.info("Time to get all data : %.3f sec" % (time.time() -
 303         time_all_elements))
 304
 305     LOG.debug("infoex[wx_data]: %s", str(infoex['wx_data']))
 306
 307     # Now we only need to add in what we want to change thanks to that
 308     # abomination of a variable declaration earlier
 309     final_data[fmap['Location UUID']] = infoex['location_uuid']
 310     final_data[fmap['obDate']] = end_date.strftime('%m/%d/%Y')
 311     final_data[fmap['obTime']] = end_date.strftime('%H:%M')
 312
 313     for elementCd in infoex['wx_data']:
 314         if elementCd not in iemap:
 315             LOG.warning("BAD KEY wx_data['%s']" % (elementCd))
 316             continue
 317
 318         # CONSIDER: Casting every value to Float() -- need to investigate if
 319         #           any possible elementCds we may want are any other data
 320         #           type than float.
 321         #
 322         #           Another possibility is to query the API with
 323         #           getStationElements and temporarily store the
 324         #           storedUnitCd. But that's pretty network-intensive and
 325         #           may not even be worth it if there's only e.g. one or two
 326         #           exceptions to any otherwise uniformly Float value set.
 327         final_data[fmap[iemap[elementCd]]] = infoex['wx_data'][elementCd]
 328
 329     LOG.debug("final_data: %s" % (str(final_data)))
 330
 331     if not write_local_csv(infoex['csv_filename'], final_data):
 332         LOG.warning('Could not write local CSV file: %s',
 333                     infoex['csv_filename'])
 334         return 1;
 335
 336     if not options.dry_run:
 337         upload_csv(infoex['csv_filename'], infoex)
 338
 339     LOG.debug('DONE')
 340     return 0
 341
 342 # Data structure operations
 343 def setup_infoex_fields_mapping(location_uuid):
 344     """
 345     Create a mapping of InfoEx fields to the local data's indexing scheme.
 346
 347     INFOEX FIELDS
 348
 349     This won't earn style points in Python, but here we establish a couple
 350     of helpful mappings variables. The reason this is helpful is that the
 351     end result is simply an ordered set, the CSV file. But we still may
 352     want to manipulate the values arbitrarily before writing that file.
 353
 354     Also note that the current Auto Wx InfoEx documentation shows these
 355     keys in a graphical table with the "index" beginning at 1, but here we
 356     sanely index beginning at 0.
 357     """
 358     fmap = {}                           ; final_data     = [None] * 29
 359     fmap['Location UUID'] = 0           ; final_data[0]  = location_uuid
 360     fmap['obDate'] = 1                  ; final_data[1]  = None
 361     fmap['obTime'] = 2                  ; final_data[2]  = None
 362     fmap['timeZone'] = 3                ; final_data[3]  = 'Pacific'
 363     fmap['tempMaxHour'] = 4             ; final_data[4]  = None
 364     fmap['tempMaxHourUnit'] = 5         ; final_data[5]  = 'F'
 365     fmap['tempMinHour'] = 6             ; final_data[6]  = None
 366     fmap['tempMinHourUnit'] = 7         ; final_data[7]  = 'F'
 367     fmap['tempPres'] = 8                ; final_data[8]  = None
 368     fmap['tempPresUnit'] = 9            ; final_data[9]  = 'F'
 369     fmap['precipitationGauge'] = 10     ; final_data[10] = None
 370     fmap['precipitationGaugeUnit'] = 11 ; final_data[11] = 'in'
 371     fmap['windSpeedNum'] = 12           ; final_data[12] = None
 372     fmap['windSpeedUnit'] = 13          ; final_data[13] = 'mph'
 373     fmap['windDirectionNum'] = 14       ; final_data[14] = None
 374     fmap['hS'] = 15                     ; final_data[15] = None
 375     fmap['hsUnit'] = 16                 ; final_data[16] = 'in'
 376     fmap['baro'] = 17                   ; final_data[17] = None
 377     fmap['baroUnit'] = 18               ; final_data[18] = 'inHg'
 378     fmap['rH'] = 19                     ; final_data[19] = None
 379     fmap['windGustSpeedNum'] = 20       ; final_data[20] = None
 380     fmap['windGustSpeedNumUnit'] = 21   ; final_data[21] = 'mph'
 381     fmap['windGustDirNum'] = 22         ; final_data[22] = None
 382     fmap['dewPoint'] = 23               ; final_data[23] = None
 383     fmap['dewPointUnit'] = 24           ; final_data[24] = 'F'
 384     fmap['hn24Auto'] = 25               ; final_data[25] = None
 385     fmap['hn24AutoUnit'] = 26           ; final_data[26] = 'in'
 386     fmap['hstAuto'] = 27                ; final_data[27] = None
 387     fmap['hstAutoUnit'] = 28            ; final_data[28] = 'in'
 388
 389     return (fmap, final_data)
 390
 391 def setup_infoex_counterparts_mapping(provider):
 392     """
 393     Create a mapping of the NRCS/MesoWest fields that this program supports to
 394     their InfoEx counterparts
 395     """
 396     iemap = {}
 397
 398     if provider == 'nrcs':
 399         iemap['PREC'] = 'precipitationGauge'
 400         iemap['TOBS'] = 'tempPres'
 401         iemap['SNWD'] = 'hS'
 402         iemap['PRES'] = 'baro'
 403         iemap['RHUM'] = 'rH'
 404         iemap['WSPD'] = 'windSpeedNum'
 405         iemap['WDIR'] = 'windDirectionNum'
 406         # unsupported by NRCS:
 407         # windGustSpeedNum
 408     elif provider == 'mesowest':
 409         iemap['precip_accum'] = 'precipitationGauge'
 410         iemap['air_temp'] = 'tempPres'
 411         iemap['snow_depth'] = 'hS'
 412         iemap['pressure'] = 'baro'
 413         iemap['relative_humidity'] = 'rH'
 414         iemap['wind_speed'] = 'windSpeedNum'
 415         iemap['wind_direction'] = 'windDirectionNum'
 416         iemap['wind_gust'] = 'windGustSpeedNum'
 417
 418     return iemap
 419
 420 # CSV operations
 421 def write_local_csv(path_to_file, data):
 422     """Write the specified CSV file to disk"""
 423     with open(path_to_file, 'w') as f:
 424         # The requirement is that empty values are represented in the CSV
 425         # file as "", csv.QUOTE_NONNUMERIC achieves that
 426         LOG.debug("writing CSV file '%s'" % (path_to_file))
 427         writer = csv.writer(f, quoting=csv.QUOTE_NONNUMERIC)
 428         writer.writerow(data)
 429         f.close()
 430     return True
 431
 432 def upload_csv(path_to_file, infoex_data):
 433     """Upload the specified CSV file to InfoEx FTP and remove the file"""
 434     with open(path_to_file, 'rb') as file_object:
 435         LOG.debug("uploading FTP file '%s'" % (infoex_data['host']))
 436         ftp = FTP(infoex_data['host'], infoex_data['uuid'],
 437                   infoex_data['api_key'])
 438         ftp.storlines('STOR ' + path_to_file, file_object)
 439         ftp.close()
 440         file_object.close()
 441     os.remove(path_to_file)
 442
 443 if __name__ == "__main__":
 444     sys.exit(main())