class PhilDBSink(BufferedSink):
"""
A buffered sink using the PhilDB timeseries database.
"""
def __init__(self, dbfile, *args, **kwargs):
super(PhilDBSink, self).__init__(*args, **kwargs)
try:
create(dbfile)
except AlreadyExistsError:
pass # Database already exists, so no creation required.
self.db = PhilDB(dbfile)
self.last_known_freq = None
try:
self.db.add_source('SENSOR', 'Data from hardware sensor')
except DuplicateError:
pass # DuplicateError means the source already existed
def write_buffer(self, param_name, series):
"""Write buffer of data to database."""
if len(series) == 0:
return
try:
self.db.add_measurand(param_name, param_name, param_name)
except DuplicateError:
pass # DuplicateError means the measurand already existed
try:
self.db.add_timeseries(param_name)
except DuplicateError:
pass # DuplicateError means the timeseries already existed
freq = series.index.inferred_freq
# need to handle special case where only one value being written
# unable to calculate the frequency so we use the last known
# value which in general should always be the same.
if len(series) == 1:
freq = self.last_known_freq
elif freq is not None:
self.last_known_freq = freq
if freq is None:
raise ValueError('Unable to determine sensor frequency')
try:
self.db.add_timeseries_instance(param_name,
freq,
'None',
measurand=param_name,
source='SENSOR')
except DuplicateError:
pass # DuplicateError - the timeseries instance already existed
self.db.write(param_name,
freq,
series,
measurand=param_name,
source='SENSOR')
def test_update_and_append(self):
db = PhilDB(self.test_tsdb)
db.write('410730',
'D',
pd.Series(index=[
datetime(2014, 1, 2),
datetime(2014, 1, 3),
datetime(2014, 1, 4),
datetime(2014, 1, 5),
datetime(2014, 1, 6)
],
data=[2.5, 3.0, 4.0, 5.0, 6.0]),
measurand='Q',
source='DATA_SOURCE')
data = db.read('410730', 'D', measurand='Q', source='DATA_SOURCE')
self.assertEqual(1.0, data.values[0])
self.assertEqual(2.5, data.values[1])
self.assertEqual(3.0, data.values[2])
self.assertEqual(4.0, data.values[3])
self.assertEqual(5.0, data.values[4])
self.assertEqual(6.0, data.values[5])
self.assertEqual(datetime(2014, 1, 1), data.index[0].to_pydatetime())
self.assertEqual(datetime(2014, 1, 2), data.index[1].to_pydatetime())
self.assertEqual(datetime(2014, 1, 3), data.index[2].to_pydatetime())
self.assertEqual(datetime(2014, 1, 4), data.index[3].to_pydatetime())
self.assertEqual(datetime(2014, 1, 5), data.index[4].to_pydatetime())
self.assertEqual(datetime(2014, 1, 6), data.index[5].to_pydatetime())
def test_new_write(self):
db = PhilDB(self.test_tsdb)
db.add_timeseries('410731')
db.add_timeseries_instance('410731',
'D',
'Foo',
measurand='Q',
source='DATA_SOURCE')
db.write('410731',
'D',
pd.Series(index=[
datetime(2014, 1, 1),
datetime(2014, 1, 2),
datetime(2014, 1, 3)
],
data=[1.0, 2.0, 3.0]),
measurand='Q',
source='DATA_SOURCE')
results = db.read('410731', 'D', measurand='Q', source='DATA_SOURCE')
self.assertEqual(results.index[0].year, 2014)
self.assertEqual(results.index[0].month, 1)
self.assertEqual(results.index[0].day, 1)
self.assertEqual(results.index[1].day, 2)
self.assertEqual(results.index[2].day, 3)
self.assertEqual(results.values[0], 1.0)
self.assertEqual(results.values[1], 2.0)
self.assertEqual(results.values[2], 3.0)
def test_update_and_append(self):
db = PhilDB(self.test_tsdb)
db.write('410730', 'D', pd.Series(index = [datetime(2014,1,2), datetime(2014,1,3), datetime(2014,1,4), datetime(2014,1,5), datetime(2014,1,6)], data = [2.5, 3.0, 4.0, 5.0, 6.0]), measurand = 'Q', source = 'DATA_SOURCE')
data = db.read('410730', 'D', measurand = 'Q', source = 'DATA_SOURCE')
self.assertEqual(1.0, data.values[0])
self.assertEqual(2.5, data.values[1])
self.assertEqual(3.0, data.values[2])
self.assertEqual(4.0, data.values[3])
self.assertEqual(5.0, data.values[4])
self.assertEqual(6.0, data.values[5])
self.assertEqual(datetime(2014,1,1), data.index[0].to_pydatetime())
self.assertEqual(datetime(2014,1,2), data.index[1].to_pydatetime())
self.assertEqual(datetime(2014,1,3), data.index[2].to_pydatetime())
self.assertEqual(datetime(2014,1,4), data.index[3].to_pydatetime())
self.assertEqual(datetime(2014,1,5), data.index[4].to_pydatetime())
self.assertEqual(datetime(2014,1,6), data.index[5].to_pydatetime())
def test_new_write(self):
db = PhilDB(self.test_tsdb)
db.add_timeseries('410731')
db.add_timeseries_instance('410731', 'D', 'Foo', measurand = 'Q', source = 'DATA_SOURCE')
db.write('410731', 'D', pd.Series(index = [datetime(2014,1,1), datetime(2014,1,2), datetime(2014,1,3)], data = [1.0, 2.0, 3.0]), measurand = 'Q', source = 'DATA_SOURCE')
results = db.read('410731', 'D', measurand = 'Q', source = 'DATA_SOURCE')
self.assertEqual(results.index[0].year, 2014)
self.assertEqual(results.index[0].month, 1)
self.assertEqual(results.index[0].day, 1)
self.assertEqual(results.index[1].day, 2)
self.assertEqual(results.index[2].day, 3)
self.assertEqual(results.values[0], 1.0)
self.assertEqual(results.values[1], 2.0)
self.assertEqual(results.values[2], 3.0)
def test_log_write(self):
db = PhilDB(self.test_tsdb)
db.add_timeseries('410731')
db.add_timeseries_instance('410731',
'D',
'Foo',
measurand='Q',
source='DATA_SOURCE')
dates = [
datetime(2014, 1, 1),
datetime(2014, 1, 2),
datetime(2014, 1, 3)
]
db.write('410731',
'D',
pd.Series(index=dates, data=[1.0, 2.0, 3.0]),
measurand='Q',
source='DATA_SOURCE')
db.write('410731',
'D',
pd.Series(index=dates, data=[1.0, 2.5, 3.0]),
measurand='Q',
source='DATA_SOURCE')
db.write('410731',
'D',
pd.Series(index=[datetime(2014, 1, 4)], data=[4.0]),
measurand='Q',
source='DATA_SOURCE')
results = db.read('410731', 'D')
self.assertEqual(results.values[0], 1.0)
self.assertEqual(results.values[1], 2.5)
self.assertEqual(results.values[2], 3.0)
self.assertEqual(results.values[3], 4.0)
with tables.open_file(db.get_file_path('410731', 'D', ftype='hdf5'),
'r') as hdf5_file:
log_grp = hdf5_file.get_node('/data')
self.assertEqual(log_grp.log[0][0], 1388534400)
self.assertEqual(log_grp.log[0][1], 1.0)
self.assertEqual(log_grp.log[0][2], 0)
self.assertEqual(log_grp.log[1][0], 1388620800)
self.assertEqual(log_grp.log[1][1], 2.0)
self.assertEqual(log_grp.log[2][0], 1388707200)
self.assertEqual(log_grp.log[2][1], 3.0)
self.assertEqual(log_grp.log[3][0], 1388620800)
self.assertEqual(log_grp.log[3][1], 2.5)
self.assertEqual(log_grp.log[4][0], 1388793600)
self.assertEqual(log_grp.log[4][1], 4.0)
def write_phildb(file_list, results_file, first_run = False):
if first_run:
create('hrs_phildb')
db = PhilDB('hrs_phildb')
if first_run:
db.add_measurand('Q', 'STREAMFLOW', 'Streamflow')
db.add_source('BOM_HRS', 'Bureau of Meteorology; Hydrological Reference Stations dataset.')
write_times = []
for filename in file_list:
print("Processing file: ", filename, '...')
station_id = os.path.basename(filename).split('_')[0]
print("Using station ID: ", station_id, '...')
streamflow = pd.read_csv(filename, parse_dates=True, index_col=0, header = None)
if first_run:
db.add_timeseries(station_id)
db.add_timeseries_instance(station_id, freq, '', measurand = 'Q', source = 'BOM_HRS')
start = time.time()
db.write(station_id, freq, streamflow, measurand = 'Q', source = 'BOM_HRS')
write_times.append(time.time() - start)
np.savetxt(results_file, np.array(write_times))
def test_log_write(self):
db = PhilDB(self.test_tsdb)
db.add_timeseries('410731')
db.add_timeseries_instance('410731', 'D', 'Foo', measurand = 'Q', source = 'DATA_SOURCE')
dates = [datetime(2014,1,1), datetime(2014,1,2), datetime(2014,1,3)]
db.write('410731', 'D', pd.Series(index = dates, data = [1.0, 2.0, 3.0]), measurand = 'Q', source = 'DATA_SOURCE')
db.write('410731', 'D', pd.Series(index = dates, data = [1.0, 2.5, 3.0]), measurand = 'Q', source = 'DATA_SOURCE')
db.write('410731', 'D', pd.Series(index = [datetime(2014,1,4)], data = [4.0]), measurand = 'Q', source = 'DATA_SOURCE')
results = db.read('410731', 'D')
self.assertEqual(results.values[0], 1.0)
self.assertEqual(results.values[1], 2.5)
self.assertEqual(results.values[2], 3.0)
self.assertEqual(results.values[3], 4.0)
with tables.open_file(db.get_file_path('410731', 'D', ftype='hdf5'), 'r') as hdf5_file:
log_grp = hdf5_file.get_node('/data')
self.assertEqual(log_grp.log[0][0], 1388534400)
self.assertEqual(log_grp.log[0][1], 1.0)
self.assertEqual(log_grp.log[0][2], 0)
self.assertEqual(log_grp.log[1][0], 1388620800)
self.assertEqual(log_grp.log[1][1], 2.0)
self.assertEqual(log_grp.log[2][0], 1388707200)
self.assertEqual(log_grp.log[2][1], 3.0)
self.assertEqual(log_grp.log[3][0], 1388620800)
self.assertEqual(log_grp.log[3][1], 2.5)
self.assertEqual(log_grp.log[4][0], 1388793600)
self.assertEqual(log_grp.log[4][1], 4.0)
db = PhilDB(test_tsdb_path)
db.add_measurand('Q', 'STREAMFLOW', 'Streamflow')
db.add_source('DATA_SOURCE', '')
db.add_timeseries('410730')
db.add_timeseries_instance('410730',
'D',
'',
measurand='Q',
source='DATA_SOURCE')
db.write('410730',
'D',
pd.Series(index=[
datetime.date(2014, 1, 1),
datetime.date(2014, 1, 2),
datetime.date(2014, 1, 3)
],
data=[1, 2, 3]),
source='DATA_SOURCE',
measurand='Q')
db.add_timeseries('123456')
db.add_timeseries_instance('123456',
'D',
'',
measurand='Q',
source='DATA_SOURCE')
db.write('123456',
'D',
pd.Series(index=[
datetime.date(2014, 1, 1),
data.append(ob[measurand])
dates.reverse()
data.reverse()
station_id = station_json['observations']['header'][0]['ID']
return station_id, pd.Series(data, dates)
measurand = 'air_temp'
source = 'BOM_OBS'
freq = '30min'
station_id, data = parse(json.load(open(sys.argv[2])), measurand)
db.add_measurand(measurand, measurand, 'Air Temperature')
db.add_source('BOM_OBS', 'Australian Bureau of Meteorology Observations')
db.add_timeseries(station_id)
db.add_timeseries_instance(station_id, freq, 'None', measurand = measurand, source = source)
db.write(station_id, freq, data, measurand = measurand, source = source)
for i in range(3, len(sys.argv)):
print("Processing file: ", sys.argv[i], '...')
try:
x = parse(json.load(open(sys.argv[i])), measurand)
db.write(station_id, freq, x, measurand = measurand, source = source)
except ValueError as e:
print('Could not parse: {0}'.format(sys.argv[i]))
station_id = station_json['observations']['header'][0]['ID']
return station_id, pd.Series(data, dates)
measurand = 'air_temp'
source = 'BOM_OBS'
freq = '30min'
station_id, data = parse(json.load(open(sys.argv[2])), measurand)
db.add_measurand(measurand, measurand, 'Air Temperature')
db.add_source('BOM_OBS', 'Australian Bureau of Meteorology Observations')
db.add_timeseries(station_id)
db.add_timeseries_instance(station_id,
freq,
'None',
measurand=measurand,
source=source)
db.write(station_id, freq, data, measurand=measurand, source=source)
for i in range(3, len(sys.argv)):
print("Processing file: ", sys.argv[i], '...')
try:
x = parse(json.load(open(sys.argv[i])), measurand)
db.write(station_id, freq, x, measurand=measurand, source=source)
except ValueError as e:
print('Could not parse: {0}'.format(sys.argv[i]))
shutil.rmtree(test_tsdb_path)
except OSError as e:
if e.errno != 2: # Code 2: No such file or directory.
raise
create(test_tsdb_path)
db = PhilDB(test_tsdb_path)
db.add_measurand('Q', 'STREAMFLOW', 'Streamflow')
db.add_source('DATA_SOURCE', '')
db.add_timeseries('410730')
db.add_timeseries_instance('410730', 'D', '', measurand = 'Q', source = 'DATA_SOURCE')
db.write('410730', 'D', pd.Series( index = [datetime.date(2014, 1, 1),
datetime.date(2014, 1, 2),
datetime.date(2014, 1, 3)],
data = [1,2,3]),
source = 'DATA_SOURCE',
measurand = 'Q'
)
db.add_timeseries('123456')
db.add_timeseries_instance('123456', 'D', '', measurand = 'Q', source = 'DATA_SOURCE')
db.write('123456', 'D', pd.Series(index = [datetime.date(2014, 1, 1),
datetime.date(2014, 1, 2),
datetime.date(2014, 1, 3)],
data = [1,2,3]),
source = 'DATA_SOURCE',
measurand = 'Q'
)
import os
import sys
import datetime
import pandas as pd
from phildb.database import PhilDB
print("Writing to PhilDB({0})".format(sys.argv[1]))
db = PhilDB(sys.argv[1])
db.add_measurand('maxT', 'MAXIMUM_TEMPERATURE', 'Maximum Temperature')
db.add_measurand('minT', 'MINIMUM_TEMPERATURE', 'Minimum Temperature')
db.add_source('BOM_ACORN_SAT', 'Bureau of Meteorology; Hydrological Reference Stations dataset.')
freq = 'D'
for i in range(2, len(sys.argv)):
print("Processing file: ", sys.argv[i], '...')
station_id = "{0:06d}".format(int(os.path.basename(sys.argv[i])))
print("Using station ID: ", station_id, '...')
db.add_timeseries(station_id)
for variable in ['minT', 'maxT']:
input_file = 'data/acorn.sat.{0}.{1}.daily.txt'.format(variable, station_id)
df = pd.read_csv(input_file, parse_dates=[0], index_col=0, header=None, skiprows=1, sep=r"\s+", na_values='99999.9', names=['Date',variable])
db.add_timeseries_instance(station_id, freq, 'ACORN-SAT', measurand = variable, source = 'BOM_ACORN_SAT')
db.write(station_id, freq, df[variable], measurand = variable, source = 'BOM_ACORN_SAT')
