In [53]:

    

import pandas as pd


    

In [54]:

    

import numpy as np
import matplotlib.pyplot as plt
%matplotlib inline


    

In [55]:

    

from influxdb import DataFrameClient


    

In [58]:

    

# see https://influxdb-python.readthedocs.io/en/latest/examples.html#tutorials-pandas
"""Instantiate the connection to the InfluxDB client."""

host, port = 'localhost', 8086
user, password = 'root', 'root'
dbname = 'weatherdata'
protocol = 'line'

CLIENT = DataFrameClient(host, port, user, password, dbname)


    

In [57]:

    

#print("Create database: " + dbname)
#CLIENT.create_database(dbname)

# /!\ delete /!\
# CLIENT.drop_database( dbname  )


    

In [6]:

    

coords = (45.1973288, 5.7139923)


    

In [7]:

    

firstday = '2017-11-05'
lastday = '2017-11-10'

daterange = pd.date_range(start=firstday, end=lastday,  freq='D', normalize=True)
daterange_iso = daterange.strftime('%Y-%m-%dT%H:%M:%SZ')


    

In [8]:

    

from darksky import forecast

# """ You get 1,000 API calls for free each day. """

# Load the API key for darksky
with open('darksky_key.txt') as f:
    KEY = f.read()

EXCLUDE = ['currently', 'minutely', 'daily', 'flags']  # from the query


    

In [9]:

    

def query2dataframe( coords,  day ):
    # query
    data = forecast(KEY, *coords, units='si', lang='fr', time=day, exclude=EXCLUDE)
    
    print('get data for %s' % day)
    
    # build the DataFrame from records (dict)
    records = data['hourly']['data']  
    weatherdataDF = pd.DataFrame.from_records(records, index='time')

    weatherdataDF.index = pd.to_datetime(weatherdataDF.index, unit='s')
    
    # debug: force les valeurs en flottant  ( plutôt que Int à cause des NaN ) 
    cols_int = list( weatherdataDF.select_dtypes(include=['int']).columns )
    for c in cols_int:
        weatherdataDF[c] = weatherdataDF[c].dropna().astype(float)
        
    return weatherdataDF


    

In [10]:

    

def dataFrame2influx( measurement_name, DF ):
    # Ajoute colone par colones pour retirer les NaN...
    for c in DF.columns:
        CLIENT.write_points(DF[[c]].dropna(), measurement_name)
        
def add1day( measurement_name, coords, day ):
    weatherdataDF = query2dataframe( coords,  day )
    dataFrame2influx( measurement_name, weatherdataDF )


    

In [11]:

    

measurement_name = 'darksky'


    

In [12]:

    

for d in daterange_iso:
    add1day( measurement_name, coords, d )


    

    




get data for 2017-11-05T00:00:00Z
get data for 2017-11-06T00:00:00Z
get data for 2017-11-07T00:00:00Z
get data for 2017-11-08T00:00:00Z
get data for 2017-11-09T00:00:00Z
get data for 2017-11-10T00:00:00Z

In [370]:

    

weatherdataDF = query2dataframe( coords,  '2017-09-02T00:00:00Z' )


    

    




get data for 2017-09-02T00:00:00Z

In [59]:

    

import emoncmsfeed as getfeeds


    

In [60]:

    

dataframefreq = '15min'
feeds =  { 'T_int': 3 }


    

In [61]:

    

measurement_name = 'openenergymonitor'


    

In [40]:

    

for date in daterange:
    startday = date
    lastday = date + pd.to_timedelta('1d')
    
    startday_iso = startday.strftime('%Y-%m-%dT%H:%M:%SZ')
    lastday_iso = lastday.strftime('%Y-%m-%dT%H:%M:%SZ')
    
    data = getfeeds.builddataframe( feeds, dataframefreq , startdate=startday, enddate=lastday )
    
    # cas jour sans donnée
    if len( data ) == 0 :
        print('no data %s'%daterange_iso[i])
        continue
    
    # Ajoute colone par colones pour retirer les NaN...
    for c in data.columns:
        CLIENT.write_points(data[[c]].dropna(), measurement_name)


    

In [66]:

    

startday = pd.to_datetime('2017-11-10T00:00:00Z', format='%Y-%m-%dT%H:%M:%SZ')
lastday = pd.Timestamp("today")


    

In [63]:

    

data = getfeeds.builddataframe( feeds, dataframefreq , startdate=startday, enddate=lastday )


    

In [64]:

    

# Ajoute colone par colones pour retirer les NaN...
for c in data.columns:
    CLIENT.write_points(data[[c]].dropna(), measurement_name)


    

In [65]:

    

data.plot(figsize=(12, 5))


    

    
Out[65]:





<matplotlib.axes._subplots.AxesSubplot at 0x7f7d8ab93ac8>






    

In [18]:

    

import sunradiation as sun


    

In [19]:

    

query = 'SELECT cloudCover FROM darksky'


    

In [20]:

    

cloudcover = CLIENT.query(query)
cloudcoverDF = cloudcover['darksky']


    

In [21]:

    

cloudcoverDF = cloudcoverDF.resample('15min').interpolate()


    

In [22]:

    

sundata = sun.buildmultidayDF( coords, cloudcoverDF.index, cloudCover = cloudcoverDF['cloudCover']  )


    

    




/home/etienne/Projets/thermique_appart/py3/lib/python3.5/site-packages/pysolar/time.py:105: UserWarning: I don't know about leap seconds after 2015
  (leap_seconds_base_year + len(leap_seconds_adjustments) - 1)

In [23]:

    

measurement_name = 'sundata'

# Ajoute colone par colones pour retirer les NaN...
for c in sundata.columns:
    CLIENT.write_points(sundata[[c]].dropna(), measurement_name)


    

In [24]:

    

# surface (m2),  sigma (deg),  azimuth (deg)
windows = { 'bastille':(1.2*0.8,    37, 50),
            'cuisine': (0.3*0.72 *2, 90, 50 ),
            'chambre': (0.3*0.72 *2, 90, 180+50),
            'vercors': (0.6*0.8 * 2, 37, 180+50) }


    

In [25]:

    

sunFlux_dict = {}

for k, values in windows.items():
    sunFlux_dict['flux_'+k] = values[0] * sun.projectDF( values[1], values[2], sundata )

sunFlux_DF = pd.DataFrame( sunFlux_dict )

# Somme
sunFlux_DF['flux_tot'] = sunFlux_DF.sum(axis=1)


    

In [26]:

    

measurement_name = 'sundata'

# Ajoute colone par colones pour retirer les NaN...
for c in sunFlux_DF.columns:
    CLIENT.write_points(sunFlux_DF[[c]].dropna(), measurement_name)