General Imports

!! IMPORTANT !!
If you did NOT install opengrid with pip, 
make sure the path to the opengrid folder is added to your PYTHONPATH



In [ ]:

    
import os
import inspect
import sys
import pandas as pd
import charts

from opengrid.library import houseprint



In [ ]:

    
import matplotlib.pyplot as plt
%matplotlib inline
plt.rcParams['figure.figsize'] = 16,8

Houseprint



In [ ]:

    
hp = houseprint.Houseprint()
# for testing:
# hp = houseprint.Houseprint(spreadsheet='unit and integration test houseprint')



In [ ]:

    
hp



In [ ]:

    
hp.sites[:5]



In [ ]:

    
hp.get_devices()[:4]



In [ ]:

    
hp.get_sensors('water')[:3]

A Houseprint object can be saved as a pickle. It loses its tmpo session however (connections cannot be pickled)



In [ ]:

    
hp.save('new_houseprint.pkl')



In [ ]:

    
hp = houseprint.load_houseprint_from_file('new_houseprint.pkl')

TMPO

The houseprint, sites, devices and sensors all have a get_data method. In order to get these working for the fluksosensors, the houseprint creates a tmpo session.



In [ ]:

    
hp.init_tmpo()
hp._tmpos.debug = False
hp.sync_tmpos()

Lookup sites, devices, sensors based on key

These methods return a single object



In [ ]:

    
hp.find_site(1)



In [ ]:

    
hp.find_device('FL03001562')



In [ ]:

    
sensor = hp.find_sensor('d5a747b86224834f745f4c9775d70241')



In [ ]:

    
print(sensor.site)
print(sensor.unit)

Lookup sites, devices, sensors based on search criteria

These methods return a list with objects satisfying the criteria



In [ ]:

    
hp.search_sites(inhabitants=5)



In [ ]:

    
hp.search_sensors(type='electricity', direction='Import')

Get Data



In [ ]:

    
hp.sync_tmpos()



In [ ]:

    
head = pd.Timestamp('20151102')
tail = pd.Timestamp('20151103')
df = hp.get_data(sensortype='water', head=head, tail=tail, diff=True, resample='min', unit='l/min')
charts.plot(df, stock=True, show='inline')

Site



In [ ]:

    
site = hp.find_site(1)
site



In [ ]:

    
print(site.size)
print(site.inhabitants)
print(site.postcode)
print(site.construction_year)
print(site.k_level)
print(site.e_level)
print(site.epc_cert)



In [ ]:

    
site.devices



In [ ]:

    
site.get_sensors('electricity')



In [ ]:

    
head = pd.Timestamp('20150617')
tail = pd.Timestamp('20150628')
df=site.get_data(sensortype='electricity', head=head,tail=tail, diff=True, unit='kW')
charts.plot(df, stock=True, show='inline')

Device



In [ ]:

    
device = hp.find_device('FL03001552')
device



In [ ]:

    
device.key



In [ ]:

    
device.get_sensors('gas')



In [ ]:

    
head = pd.Timestamp('20151101')
tail = pd.Timestamp('20151104')
df = hp.get_data(sensortype='gas', head=head,tail=tail, diff=True, unit='kW')
charts.plot(df, stock=True, show='inline')

Sensor



In [ ]:

    
sensor = hp.find_sensor('53b1eb0479c83dee927fff10b0cb0fe6')
sensor



In [ ]:

    
sensor.key



In [ ]:

    
sensor.type



In [ ]:

    
sensor.description



In [ ]:

    
sensor.system



In [ ]:

    
sensor.unit



In [ ]:

    
head = pd.Timestamp('20150617')
tail = pd.Timestamp('20150618')
df=sensor.get_data(head,tail,diff=True, unit='W')
charts.plot(df, stock=True, show='inline')

Getting data for a selection of sensors



In [ ]:

    
sensors = hp.search_sensors(type='electricity', system='solar')
print(sensors)
df = hp.get_data(sensors=sensors, head=head, tail=tail, diff=True, unit='W')
charts.plot(df, stock=True, show='inline')

Dynamically loading data sensor per sensor

A call to hp.get_data() is lazy: it creates a big list of Data Series per sensor and concatenates them. This can take a while, specifically when you need many sensors and a large time span.

Often, you don't use the big DataFrame as a whole, you rather re-divide it by using a for loop and looking at each sensor individually.

By using hp.get_data_dynamic(), data is fetched from tmpo per sensor at a time, just when you need it.



In [ ]:

    
dyn_data = hp.get_data_dynamic(sensortype='electricity', head=head, tail=tail)



In [ ]:

    
ts = next(dyn_data)
df = pd.DataFrame(ts)
charts.plot(df, stock=True, show='inline')

You can run the cell above multiple times and eacht time the next sensor will be fetched



In [ ]: