'''
Created on May 19, 2014
@author: Brett Paufler
Copyright Brett Paufler 5-20-14

2pm start 4:30start   2.5hours from start to something graphical
7:30 - 8:30            1 hour 5x charts
10-:30 11            .5 hr correlation
4 hours, done with oil

Eh, not quite, +1 more hour for scatter and a few more rolling correlations



'''



import pandas

import matplotlib.pyplot as plt



'''
Could work as a function by why bother?

vary these four variables and should work fairly well for other commodity types

commodityDataFile IS location of csv file
cN IS header name for commodity in that file 
cL IS the name desired on graphical output
normalization factor IS rough factor difference between commodity and corn


default values for OIL ARE
    commodityDataFile = r"rawCommodityData/RWTC-WTI-Crude-Oil-Spot-Price-Cushing-OK-FOB.csv"
    cN = "Oil"
    cL = "Price of Oil in Dollars"
    normalizationFactor = 5
    
#FOR REFERENCE, VALUES TO BELOW WORK FOR OIL   
commodityDataFile = r"rawCommodityData/RWTC-WTI-Crude-Oil-Spot-Price-Cushing-OK-FOB.csv"
cN = "Oil"
cL = "Price of Oil in Dollars"
normalizationFactor = 5   
'''
#Modify These Four to Transfer to Another Commodity
commodityDataFile = r"rawCommodityData/RWTC-WTI-Crude-Oil-Spot-Price-Cushing-OK-FOB.csv"
cN = "Oil"
cL = "Price of Oil in Dollars"
normalizationFactor = 5






#From Here down, be careful about modifying anything

cornDataFile = r"./cornOneDataFile/compositeCorn.txt"



cornData = pandas.read_csv(cornDataFile, index_col='Date', parse_dates=['Date'])
print cornData

otherData = pandas.read_csv(commodityDataFile, index_col='Date', parse_dates=['Date'])
print otherData



'''merging the Data frames
    cornData is left
    otherData is right
    how='inner' means the intersection of the
    
        left_index=True means left_index is used as value to merge the left table on
        right_index=True means right_index is used as the value to merge the right table on   
        
        #Old not used anymore
        on="Date", Date value in both dataFrames
'''
mergedData = cornData.merge(otherData, how='inner', left_index="Date", right_index=True)


del mergedData["Open"]
del mergedData["High"]
del mergedData["Low"]
del mergedData["Volume"]
del mergedData["Interest"]


mD = mergedData


#Prepping the DateYear Value (extra, stringify, slice)
mD["dateYear"] = mD.index
mD["dateYear"] = mD["dateYear"].apply(str)
mD["dateYear"] = mD["dateYear"].str.slice(0,4,1)

#much faster, in onde step
mD["futureYear"] = mD["Future"].str.replace("CZ", "")

#This is the Filter Logic
mD = mD[mD["dateYear"] == mD["futureYear"]]


#delete extraneous
del mD["dateYear"]
del mD["futureYear"]
del mD["Future"]


mD[cL] = mD[cN]
del mD[cN]

mD["Price of Corn in Dollars"] = mD["Settle"]
del mD["Settle"]

'''
mD is the raw dataFrame, complete and full at this point
index = date
Price of Corn in Dollars
Price of Whatever in Dollars
'''

print mD




#Making and Printing the Charts
plt.rcParams['figure.figsize'] = 20,10


fig = mD.plot()
outputName = "./chartsOut/Corn_vs_" + cN + "_Line.png"
print "Working On " + outputName
plt.savefig(outputName)

fig = mD.boxplot()
outputName = "./chartsOut/Corn_vs_" + cN + "_BoxPlot.png"
print "Working On " + outputName
plt.savefig(outputName)



corrData = pandas.rolling_corr(mD["Price of Corn in Dollars"], mD["Price of Oil in Dollars"], window=120)
corrThirty = pandas.rolling_corr(mD["Price of Corn in Dollars"], mD["Price of Oil in Dollars"], window=30)
corrYear = pandas.rolling_corr(mD["Price of Corn in Dollars"], mD["Price of Oil in Dollars"], window=360)
corrThreeYear = pandas.rolling_corr(mD["Price of Corn in Dollars"], mD["Price of Oil in Dollars"], window=1080)


factor = normalizationFactor
if factor:
    fS = "Price of Oil in %dx Dollars" % factor
    mD[fS] = mD[cL] * factor
    del mD[cL]
    
    
    mD["temp"] = mD["Price of Corn in Dollars"]
    del mD["Price of Corn in Dollars"]
    mD["Price of Corn in Dollars"] = mD["temp"]
    del mD["temp"]
    
    fig = mD.plot()
    outputName = "./chartsOut/Corn_vs_%s_%dx_Line.png" % (cN, factor)
    print "Working On " + outputName
    plt.savefig(outputName)

    fig = mD.boxplot()
    outputName = "./chartsOut/Corn_vs_%s_%dx_BoxPlot.png" % (cN, factor)
    print "Working On " + outputName
    plt.savefig(outputName)
      
    mD[cL] = mD[fS] / factor
    del mD[fS]
    
    print mD 
else:
    print "No Factor"

#This would often work better reversed...
mD["Ratio Price Corn / Price Oil"] = mD["Price of Corn in Dollars"] / mD[cL]
del mD["Price of Corn in Dollars"] 
del mD[cL]

fig = mD.plot()
outputName = "./chartsOut/Corn_vs_" + cN + "_Ratio_Line.png"
print "Working On " + outputName
plt.savefig(outputName)

fig = mD.boxplot()
outputName = "./chartsOut/Corn_vs_" + cN + "_Ratio_BoxPlot.png"
print "Working On " + outputName
plt.savefig(outputName)





print "Printing mD"
print mD


#preFormatting the corrData dataFrames
print "Printing corrData"
corrData = pandas.DataFrame(corrData)
corrData.columns = ["Correlation Oil to Corn",]
print corrData

print "Printing corrThirty"
corrThirty = pandas.DataFrame(corrThirty)
corrThirty.columns = ["Correlation Oil to Corn",]
print corrThirty

print "Printing corrYear"
corrYear = pandas.DataFrame(corrYear)
corrYear.columns = ["Correlation Oil to Corn",]
print corrYear

print "Printing corrYear"
corrThreeYear = pandas.DataFrame(corrThreeYear)
corrThreeYear.columns = ["Correlation Oil to Corn",]
print corrThreeYear



#line plot corr 360, year long
fig = corrThreeYear.plot()
outputName = "./chartsOut/Corn_vs_" + cN + "_1080_Rolling_Correlation.png"
print "Working On " + outputName
plt.savefig(outputName)


#line plot corr 360, year long
fig = corrYear.plot()
outputName = "./chartsOut/Corn_vs_" + cN + "_360_Rolling_Correlation.png"
print "Working On " + outputName
plt.savefig(outputName)

#line plot corr 30
fig = corrThirty.plot()
outputName = "./chartsOut/Corn_vs_" + cN + "_30_Rolling_Correlation.png"
print "Working On " + outputName
plt.savefig(outputName)

#line plot, corr 120
fig = corrData.plot()
outputName = "./chartsOut/Corn_vs_" + cN + "_120_Rolling_Correlation.png"
print "Working On " + outputName
plt.savefig(outputName)



#scatter plot corr 120
plt.scatter(corrData.index, corrData["Correlation Oil to Corn"] )
outputName = "./chartsOut/Corn_vs_" + cN + "_120_Rolling_Scatter.png"
print "Working On " + outputName
plt.savefig(outputName)





print corrThirty












print "End of cornToCommodity"