#!/usr/bin/env python
# coding: utf-8

# # EnergyPlus Output Data Analysis Example
#     
# Created by Clayton Miller (miller.clayton@arch.ethz.ch)
# 
# The goal of this notebook is to give a user a glimpse at the loading and manipulation of a .csv output of EnergyPlus
# 
# Execute the cells in this notebook one at a time and try to understand what each code snippet is doing. There will be text notation before many of the cells attempting to explain what is going on.

# First we load some libraries that we will use.

# In[ ]:


import pandas as pd
import datetime
from datetime import timedelta
import time


# In[ ]:


get_ipython().run_line_magic('matplotlib', 'inline')


# The following is a Python `fucntion` that I created ot read a .csv file, do a conversion and change the timestamp

# In[ ]:


def loadsimdata(file,pointname,ConvFactor):
    df = pd.read_csv(file)
    df['desiredpoint'] = df[pointname]*ConvFactor
    df.index = eplustimestamp(df)
    pointdf = df['desiredpoint']
    return pointdf


# ## Open an EnergyPlus Output CSV file
# 
# First let's open the .csv output file from an EnergyPlus run and visualize it.

# In[ ]:


SimulationData = pd.read_csv('EnergyplusSimulationData.csv')


# In[ ]:


SimulationData


# We can select a certain column and see what's inside -- there is quite a few columns in this output file. We can see using the `.info` function from pandas that there are 256 columns

# In[ ]:


SimulationData.info()


# In[ ]:


SimulationData['AIRNODE_ZONENODE_U1_S:System Node Setpoint Temp[C](Hourly) ']#.plot()


# In[ ]:


SimulationData['Date/Time'].tail()


# ### We need to convert 24:00:00 to 00:00:00 for it to play nice with Pandas

# In[ ]:


#Function to convert timestamps
def eplustimestamp(simdata):
    timestampdict={}
    for i,row in simdata.T.iteritems():
        timestamp = str(2013) + row['Date/Time']
        try:
            timestampdict[i] = datetime.datetime.strptime(timestamp,'%Y %m/%d  %H:%M:%S')
        except ValueError:
            tempts = timestamp.replace(' 24', ' 23')
            timestampdict[i] = datetime.datetime.strptime(tempts,'%Y %m/%d  %H:%M:%S')
            timestampdict[i] += timedelta(hours=1)
    timestampseries = pd.Series(timestampdict)
    return timestampseries


# In[ ]:


SimulationData.index = eplustimestamp(SimulationData)


# In[ ]:


SimulationData.info()


# In[ ]:


SimulationData['AIRNODE_ZONENODE_U1_S:System Node Setpoint Temp[C](Hourly) ']


# In[ ]:


ColumnsList = pd.Series(SimulationData.columns)
ColumnsList.head(100)


# ## Built-in String Query functions to find desired columns
# 
# We can use the string query functions in Pandas to find columns to select for visualization, etc without manually inputting them.
# 

# In[ ]:


ColumnsList.str.endswith("Zone Mean Air Temperature [C](Hourly)")


# In[ ]:


ColumnsList[(ColumnsList.str.endswith("Zone Mean Air Temperature [C](Hourly)"))]


# In[ ]:


ZoneTempPointList = list(ColumnsList[(ColumnsList.str.endswith("Zone Mean Air Temperature [C](Hourly)"))])
ZoneTempPointList


# In[ ]:


BasementZoneTemp = list(ColumnsList[(ColumnsList.str.endswith("Zone Mean Air Temperature [C](Hourly)"))&(ColumnsList.str.contains("U1"))])
GroundFloorZoneTemp = list(ColumnsList[(ColumnsList.str.endswith("Zone Mean Air Temperature [C](Hourly)"))&(ColumnsList.str.contains("00"))])
Floor1ZoneTemp = list(ColumnsList[(ColumnsList.str.endswith("Zone Mean Air Temperature [C](Hourly)"))&(ColumnsList.str.contains("01"))])
Floor2ZoneTemp = list(ColumnsList[(ColumnsList.str.endswith("Zone Mean Air Temperature [C](Hourly)"))&(ColumnsList.str.contains("02"))])
Floor3ZoneTemp = list(ColumnsList[(ColumnsList.str.endswith("Zone Mean Air Temperature [C](Hourly)"))&(ColumnsList.str.contains("03"))])
Floor4ZoneTemp = list(ColumnsList[(ColumnsList.str.endswith("Zone Mean Air Temperature [C](Hourly)"))&(ColumnsList.str.contains("04"))])


# In[ ]:


ZoneTemp = SimulationData[ZoneTempPointList]#.drop(['EMS:All Zones Total Heating Energy {J}(Hourly)'],axis=1)


# In[ ]:


ZoneTemp.info()


# ## Visualization
# 
# Now that we have the data in a Pandas Dataframe, let's do some analysis -- visualization, statistics, etc

# In[ ]:


ZoneTemp.plot(figsize=(25,15))


# In[ ]:


ZoneTemp[BasementZoneTemp].plot(figsize=(25,10))


# In[ ]:


ZoneTemp[GroundFloorZoneTemp].truncate(before='2013-03-10',after='2013-03-14').plot(figsize=(25,10))


# In[ ]:


ZoneTemp[Floor1ZoneTemp].plot(figsize=(25,10))


# In[ ]:


ZoneTemp[Floor2ZoneTemp].plot(figsize=(25,10))


# ## Zooming in using the `.truncate()` pandas method

# In[ ]:


SimulationData['Environment:Outdoor Dry Bulb [C](Hourly)'].truncate(before='2013-03-10',after='2013-03-14').plot(figsize=(25,10))


# ## Let's take a deeper look at the Floor 2 Zone Temperations

# In[ ]:


Floor2Temps = ZoneTemp[Floor2ZoneTemp]


# In[ ]:


Floor2Temps.info()


# In[ ]:


Floor2Temps.describe()


# ## Heatmaps
# 
# Heatmaps are a great way to visualize performance data

# In[ ]:


import pandas as pd
import matplotlib.pyplot as plt
import numpy as np
from matplotlib.backends.backend_pdf import PdfPages
import os
import matplotlib.dates as mdates
import datetime as dt


# In[ ]:


Floor1Energy = list(ColumnsList[(ColumnsList.str.endswith("Total Heating Energy {J}(Hourly)"))&(ColumnsList.str.contains("01"))])


# In[ ]:


Floor1Energy


# In[ ]:


df_hourly = SimulationData.resample('H').mean()

#Add the Date and time for pivoting
df_hourly['Date'] = df_hourly.index.map(lambda t: t.date())
df_hourly['Time'] = df_hourly.index.map(lambda t: t.time())


# In[ ]:


numberofplots = len(Floor1Energy)


# In[ ]:


pointcounter = 1
fig = plt.figure(figsize=(40, 4 * numberofplots))


for energypoint in Floor1Energy:
    
    print "Loading data from "+energypoint
    
    #Pivot
    df_pivot = pd.pivot_table(df_hourly, values=energypoint, index='Time', columns='Date')

    # Get the data
    x = mdates.drange(df_pivot.columns[0], df_pivot.columns[-1] + datetime.timedelta(days=1), dt.timedelta(days=1))
    y = np.linspace(1, 24, 24)

    # Plot
    ax = fig.add_subplot(numberofplots, 1, pointcounter)

    data = np.ma.masked_invalid(np.array(df_pivot))

    qmesh = ax.pcolormesh(x, y, data)
    cbar = fig.colorbar(qmesh, ax=ax)
    cbar.ax.tick_params(labelsize= 24)
    ax.axis('tight')

    try:
        plt.title(energypoint, fontsize=26)

    except IndexError:
        continue

    # Set up as dates
    ax.xaxis_date()
    fig.autofmt_xdate()
    fig.subplots_adjust(hspace=.5)
    
    pointcounter += 1



# In[ ]:





# In[ ]: