In [1]:

    

from paramiko import SSHClient, SSHConfig
import os
import sys
sys.path.append('../../')
import datetime
import calendar
from xml.dom import minidom
from lxml import etree


    

In [2]:

    

from disaggregator import GreenButtonDatasetAdapter as gbda


    

    




/Users/sabina/anaconda/lib/python2.7/site-packages/pandas/io/excel.py:626: UserWarning: Installed openpyxl is not supported at this time. Use >=1.6.1 and <2.0.0.
  .format(openpyxl_compat.start_ver, openpyxl_compat.stop_ver))

In [3]:

    

from disaggregator import appliance


    

In [4]:

    

reload(appliance)


    

    
Out[4]:





<module 'disaggregator.appliance' from '../../disaggregator/appliance.pyc'>

In [5]:

    

import get_files


    

    




['identityfile', 'forwardx11', 'hostname', 'user', 'forwardagent']

In [6]:

    

reload(get_files)


    

    




['identityfile', 'forwardx11', 'hostname', 'user', 'forwardagent']






    
Out[6]:





<module 'get_files' from 'get_files.pyc'>

In [ ]:

    

f = get_files.get()


    

In [95]:

    

fi = f[0].read()


    

In [96]:

    

fi


    

    
Out[96]:





''

In [51]:

    

file_strings=[t+'\n      </IntervalBlock>\n    </content>\n  </entry>\n</feed>' for t in file_strings]


    

In [42]:

    

xml_string = file_strings[1]
xmldoc = minidom.parseString(xml_string)


    

In [43]:

    

trace = gbda.get_trace_from_xml(file_strings[1])


    

In [53]:

    

traces = [gbda.get_trace_from_xml(t) for t in file_strings]


    

In [54]:

    

traces[0].metadata


    

    
Out[54]:





{'source': 'GreenButtonXML'}

In [101]:

    

trace.series.index.month


    

    
Out[101]:





array([6, 6, 6, ..., 7, 7, 7])

In [60]:

    

def get_month_info(year,month):
    day_one = calendar.monthrange(year,month)[0]
    day_last = calendar.monthrange(year,month)[1]
    range_start = datetime.datetime(year,month,day_one,0,0)
    range_end = datetime.datetime(year,month,day_last,0,0) 
    good_traces = [] 
    for t in traces:
        rng = [dt.to_datetime() for dt in t.series.index]
        if range_start in rng and range_end in rng:
            good_traces.append(t)
    return good_traces


    

In [61]:

    

gt = get_month_info(2014,6)