#!/usr/bin/env python # -*- coding: utf-8 -*- """ InfoEx <-> NRCS/MesoWest Auto Wx implementation Alexander Vasarab Wylark Mountaineering LLC This program fetches data from either an NRCS SNOTEL site or MesoWest weather station and pushes it to InfoEx using the new automated weather system implementation. It is designed to be run hourly, and it asks for the last three hours of data of each desired type, and selects the most recent one. This lends some resiliency to the process and helps ensure that we have a value to send, but it can lead to somewhat inconsistent/untruthful data if e.g. the HS is from the last hour but the tempPres is from two hours ago because the instrumentation had a hiccup. It's worth considering if this is a bug or a feature. For more information, see file: README For licensing, see file: LICENSE """ import configparser import csv import datetime import logging import os import sys import time from ftplib import FTP from argparse import ArgumentParser import requests import zeep import zeep.cache import zeep.transports __version__ = '2.1.0' LOG = logging.getLogger(__name__) LOG.setLevel(logging.NOTSET) def get_parser(): """Return OptionParser for this program""" parser = ArgumentParser() parser.add_argument("--version", action="version", version=__version__) parser.add_argument("--config", dest="config", metavar="FILE", help="location of config file") parser.add_argument("--log-level", dest="log_level", default=None, help="set the log level (debug, info, warning)") parser.add_argument("--dry-run", action="store_true", dest="dry_run", default=False, help="fetch data but don't upload to InfoEx") return parser def setup_config(config): """Setup config variable based on values specified in the ini file""" try: infoex = { 'host': config['infoex']['host'], 'uuid': config['infoex']['uuid'], 'api_key': config['infoex']['api_key'], 'csv_filename': config['infoex']['csv_filename'], 'location_uuid': config['infoex']['location_uuid'], 'wx_data': {}, # placeholder key, values to come later } station = dict() station['provider'] = config['station']['type'] if station['provider'] not in ['nrcs', 'mesowest']: print("Please specify either nrcs or mesowest as the station type.") sys.exit(1) if station['provider'] == 'nrcs': 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' if station['provider'] == 'mesowest': station['source'] = 'https://api.synopticdata.com/v2/stations/timeseries' station['station_id'] = config['station']['station_id'] station['units'] = config['station']['units'] station['desired_data'] = config['station']['desired_data'] # construct full API URL (sans start/end time, added later) station['source'] = station['source'] + '?token=' + \ config['station']['token'] + \ '&within=60&units=' + station['units'] + \ '&stid=' + station['station_id'] + \ '&vars=' + station['desired_data'] except KeyError as err: LOG.critical("%s not defined in configuration file", err) exit(1) # all sections/values present in config file, final sanity check try: for key in config.sections(): for subkey in config[key]: if not config[key][subkey]: raise ValueError except ValueError: LOG.critical("Config value '%s.%s' is empty", key, subkey) exit(1) return (infoex, station) def setup_logging(log_level): """Setup our logging infrastructure""" try: from systemd.journal import JournalHandler LOG.addHandler(JournalHandler()) except ImportError: ## fallback to syslog #import logging.handlers #LOG.addHandler(logging.handlers.SysLogHandler()) # fallback to stdout handler = logging.StreamHandler(sys.stdout) LOG.addHandler(handler) # ugly, but passable if log_level in [None, 'debug', 'info', 'warning']: if log_level == 'debug': LOG.setLevel(logging.DEBUG) elif log_level == 'info': LOG.setLevel(logging.INFO) elif log_level == 'warning': LOG.setLevel(logging.WARNING) else: LOG.setLevel(logging.NOTSET) else: return False return True def main(): """Main routine: sort through args, decide what to do, then do it""" parser = get_parser() options = parser.parse_args() config = configparser.ConfigParser(allow_no_value=False) if not options.config: parser.print_help() print("\nPlease specify a configuration file via --config.") sys.exit(1) config.read(options.config) if not setup_logging(options.log_level): parser.print_help() print("\nPlease select an appropriate log level or remove the switch (--log-level).") sys.exit(1) (infoex, station) = setup_config(config) LOG.debug('Config parsed, starting up') # create mappings (fmap, final_data) = setup_infoex_fields_mapping(infoex['location_uuid']) iemap = setup_infoex_counterparts_mapping(station['provider']) # override units if user selected metric if station['units'] == 'metric': final_data = switch_units_to_metric(final_data, fmap) (begin_date, end_date) = setup_time_values() # get the data LOG.debug("Getting %s data from %s to %s", str(station['desired_data']), str(begin_date), str(end_date)) time_all_elements = time.time() # get the data if station['provider'] == 'nrcs': infoex['wx_data'] = get_nrcs_data(begin_date, end_date, station) elif station['provider'] == 'mesowest': infoex['wx_data'] = get_mesowest_data(begin_date, end_date, station) 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'])) # 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') for element_cd in infoex['wx_data']: if element_cd not in iemap: LOG.warning("BAD KEY wx_data['%s']", element_cd) continue # Massage precision of certain values to fit InfoEx's # expectations # # 0 decimal places: wind speed, wind direction, wind gust, snow depth # 1 decimal place: air temp, baro # Avoid transforming None values if infoex['wx_data'][element_cd] is None: continue elif element_cd in ['wind_speed', 'WSPD', 'wind_direction', '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) # CONSIDER: Casting every value to Float() -- need to investigate if # any possible elementCds we may want are any other data # type than float. # # Another possibility is to query the API with # getStationElements and temporarily store the # storedUnitCd. But that's pretty network-intensive and # may not even be worth it if there's only e.g. one or two # exceptions to any otherwise uniformly Float value set. final_data[fmap[iemap[element_cd]]] = infoex['wx_data'][element_cd] 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 not options.dry_run: upload_csv(infoex['csv_filename'], infoex) LOG.debug('DONE') return 0 # data structure operations def setup_infoex_fields_mapping(location_uuid): """ Create a mapping of InfoEx fields to the local data's indexing scheme. INFOEX FIELDS This won't earn style points in Python, but here we establish a couple of helpful mappings variables. The reason this is helpful is that the end result is simply an ordered set, the CSV file. But we still may want to manipulate the values arbitrarily before writing that file. Also note that the current Auto Wx InfoEx documentation shows these keys in a graphical table with the "index" beginning at 1, but here we sanely index beginning at 0. """ # pylint: disable=too-many-statements,multiple-statements,bad-whitespace fmap = {} ; final_data = [None] * 29 fmap['Location UUID'] = 0 ; final_data[0] = location_uuid fmap['obDate'] = 1 ; final_data[1] = None fmap['obTime'] = 2 ; final_data[2] = None fmap['timeZone'] = 3 ; final_data[3] = 'Pacific' fmap['tempMaxHour'] = 4 ; final_data[4] = None fmap['tempMaxHourUnit'] = 5 ; final_data[5] = 'F' fmap['tempMinHour'] = 6 ; final_data[6] = None fmap['tempMinHourUnit'] = 7 ; final_data[7] = 'F' fmap['tempPres'] = 8 ; final_data[8] = None fmap['tempPresUnit'] = 9 ; final_data[9] = 'F' fmap['precipitationGauge'] = 10 ; final_data[10] = None fmap['precipitationGaugeUnit'] = 11 ; final_data[11] = 'in' fmap['windSpeedNum'] = 12 ; final_data[12] = None fmap['windSpeedUnit'] = 13 ; final_data[13] = 'mph' fmap['windDirectionNum'] = 14 ; final_data[14] = None fmap['hS'] = 15 ; final_data[15] = None fmap['hsUnit'] = 16 ; final_data[16] = 'in' fmap['baro'] = 17 ; final_data[17] = None fmap['baroUnit'] = 18 ; final_data[18] = 'inHg' fmap['rH'] = 19 ; final_data[19] = None fmap['windGustSpeedNum'] = 20 ; final_data[20] = None fmap['windGustSpeedNumUnit'] = 21 ; final_data[21] = 'mph' fmap['windGustDirNum'] = 22 ; final_data[22] = None fmap['dewPoint'] = 23 ; final_data[23] = None fmap['dewPointUnit'] = 24 ; final_data[24] = 'F' fmap['hn24Auto'] = 25 ; final_data[25] = None fmap['hn24AutoUnit'] = 26 ; final_data[26] = 'in' fmap['hstAuto'] = 27 ; final_data[27] = None fmap['hstAutoUnit'] = 28 ; final_data[28] = 'in' return (fmap, final_data) def setup_infoex_counterparts_mapping(provider): """ Create a mapping of the NRCS/MesoWest fields that this program supports to their InfoEx counterparts """ iemap = {} if provider == 'nrcs': iemap['PREC'] = 'precipitationGauge' iemap['TOBS'] = 'tempPres' iemap['SNWD'] = 'hS' iemap['PRES'] = 'baro' iemap['RHUM'] = 'rH' iemap['WSPD'] = 'windSpeedNum' iemap['WDIR'] = 'windDirectionNum' # unsupported by NRCS: # windGustSpeedNum elif provider == 'mesowest': iemap['precip_accum'] = 'precipitationGauge' iemap['air_temp'] = 'tempPres' iemap['snow_depth'] = 'hS' iemap['pressure'] = 'baro' iemap['relative_humidity'] = 'rH' iemap['wind_speed'] = 'windSpeedNum' iemap['wind_direction'] = 'windDirectionNum' iemap['wind_gust'] = 'windGustSpeedNum' return iemap # provider-specific operations 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()) client = zeep.Client(wsdl=station['source'], transport=transport) remote_data = {} for element_cd in station['desired_data']: time_element = time.time() # get the last three hours of data for this elementCd/element_cd tmp = client.service.getHourlyData( stationTriplets=[station['station_id']], elementCd=element_cd, ordinal=1, beginDate=begin, endDate=end) LOG.info("Time to get NRCS elementCd '%s': %.3f sec", element_cd, time.time() - time_element) values = tmp[0]['values'] # sort and isolate the most recent # # NOTE: we do this because sometimes there are gaps in hourly data # in NRCS; yes, we may end up with slightly inaccurate data, # so perhaps this decision will be re-evaluated in the future if values: ordered = sorted(values, key=lambda t: t['dateTime'], reverse=True) remote_data[element_cd] = ordered[0]['value'] else: remote_data[element_cd] = None return remote_data def get_mesowest_data(begin, end, station): """get the data we're after from the MesoWest/Synoptic API""" remote_data = {} # massage begin/end date format begin_date_str = begin.strftime('%Y%m%d%H%M') end_date_str = end.strftime('%Y%m%d%H%M') # 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: json = req.json() except ValueError: LOG.error("Bad JSON in MesoWest response") sys.exit(1) try: observations = json['STATION'][0]['OBSERVATIONS'] except ValueError: LOG.error("Bad JSON in MesoWest response") sys.exit(1) pos = len(observations['date_time']) - 1 for element_cd in station['desired_data'].split(','): # sort and isolate the most recent, see note above in NRCS for how and # why this is done # # NOTE: Unlike in the NRCS case, the MesoWest API response contains all # data (whereas with NRCS, we have to make a separate request for # each element we want). This is nice for network efficiency but # it means we have to handle this part differently for each. # # NOTE: Also unlike NRCS, MesoWest provides more granular data; NRCS # provides hourly data, but MesoWest can often provide data every # 10 minutes -- though this provides more opportunity for # irregularities # 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] else: remote_data[element_cd] = None else: remote_data[element_cd] = None return remote_data def switch_units_to_metric(data_map, mapping): """replace units with metric counterparts""" # 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['tempPresUnit']] = 'C' data_map[mapping['hsUnit']] = 'm' data_map[mapping['windSpeedUnit']] = 'm/s' data_map[mapping['windGustSpeedNumUnit']] = 'm/s' return data_map # CSV operations def write_local_csv(path_to_file, data): """Write the specified CSV file to disk""" with open(path_to_file, 'w') as file_object: # The requirement is that empty values are represented in the CSV # file as "", csv.QUOTE_NONNUMERIC achieves that LOG.debug("writing CSV file '%s'", path_to_file) writer = csv.writer(file_object, quoting=csv.QUOTE_NONNUMERIC) writer.writerow(data) file_object.close() return True def upload_csv(path_to_file, infoex_data): """Upload the specified CSV file to InfoEx FTP and remove the file""" with open(path_to_file, 'rb') as file_object: LOG.debug("uploading FTP file '%s'", infoex_data['host']) ftp = FTP(infoex_data['host'], infoex_data['uuid'], infoex_data['api_key']) ftp.storlines('STOR ' + path_to_file, file_object) ftp.close() file_object.close() os.remove(path_to_file) # other miscellaneous routines def setup_time_values(): """establish time bounds of data request(s)""" # floor time to nearest hour date_time = datetime.datetime.now() 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) return (begin_date, end_date) if __name__ == "__main__": sys.exit(main())