import os
import sys
import pandas as pd
from pandas import Series
from datetime import datetime
pd.options.mode.chained_assignment = None  # default='warn'
pd.set_option('precision', 1)

def lat_lon_to_float(v):
#	Convert strings from NHC file (Lat Lon with N/S and W/E to floats)
	if (v[-1] == 'S') or (v[-1] =='W'):
		multiplier = -1
	else:
		multiplier = 1
	return float(v[:-1]) * multiplier

def convert_1d_to_2d(l, cols):
    return [l[i:i + cols] for i in range(0, len(l), cols)]

SummaryList = []
RegionalACE =[]
RegionalACE.append(['Year','MDRACE', 'GMXACE','EXTACE'])
data = []
data2= []
with open('atl_hurdat.txt','r') as f:
	for line in f.readlines():
		if line.startswith('AL'):
			storm_id =line.split(',')
			storm_id.pop()
#			storm_id_str=','.join([str(elem) for elem in storm_id])
			storm_numberFull = storm_id[0].strip()
			storm_name   = storm_id[1].strip()
			basin = storm_numberFull[0:2]
			storm_number = int(storm_numberFull[2:4])
			storm_year   = int(storm_numberFull[4:])
		else:
			location_line = line.split(',')
			location_line.pop()
#			location_line_str=','.join([str(elem) for elem in location_line])
			#print(location_line)
			dt = datetime.strptime(location_line[0]+ location_line[1], '%Y%m%d %H%M')
			dy = datetime.strptime(location_line[0], '%Y%m%d')
			jdy = dy.strftime("%j")
			jdy=int(jdy)
			hr = int(location_line[1][1:3])
			mns = int(location_line[1][3:5])
			lfall = location_line[2].strip()
			storm_status = location_line[3].strip()
			storm_lat = lat_lon_to_float(location_line[4].strip())
			storm_lon = lat_lon_to_float(location_line[5].strip())
			max_wind  = int(location_line[6].strip())
			data.append([basin, storm_number, storm_year, storm_name, storm_status, storm_lat, storm_lon, lfall, dt, dy, jdy, hr, mns, max_wind])
#			data2.append([storm_id_str,location_line_str])

#bigdf=pd.DataFrame(data2,columns=['id','locline'])
#s=bigdf['id']
#p=s.str.split(pat=",",expand=True)
#s=bigdf['locline']
#t=s.str.split(pat=",",expand=True)
#newBigdf=pd.concat([p,t],axis=1,join="inner")
#print(newBigdf)
#newBigdf.to_csv('./atl_hurdat.csv')

df = pd.DataFrame(data, columns=['Basin','StormNumber', 'StormYear', 'StormName', 'StormStatus', 'Lat', 'Lon', 'LandFall','Date_Time', 'Day', 'JulDay', 'Hour', 'Minutes','MaxWind'])
df['MaxWSqd']=df['MaxWind']*df['MaxWind']
df.drop(df[df['StormYear']<1950].index, inplace=True)
#dfPost1950=df
for status in ['TD','EX', 'LO','DB', 'SD', 'WV']:
	df.drop(df[(df['StormStatus'] == status)].index,inplace=True)

print(df)
Totallen=len(df['StormYear'].unique())
# Section to modify to pick off specific dat ranges
df.drop(df[df['Day'].dt.month < 8].index,inplace=True)
df.drop(df[df['Day'].dt.month > 8].index,inplace=True)
print(df)

years= df['StormYear'].unique()
for yr in years:
#	print("\n"+str(yr))
	dfYear=df[df['StormYear']==int(yr)]
#	print(dfYear.drop_duplicates(subset='StormName'))
	SSCount=0
	MHCount=0
	MDRACE=0
	GMXACE=0
	EXTACE=0
	MDRCount=0
	GMXCount=0
	EXTCount=0
	for num in dfYear['StormNumber'].unique():
		dfByNum=dfYear[dfYear['StormNumber']==num]
		try:
			if(dfByNum['StormStatus'].value_counts().SS == dfByNum.shape[0]):
				SSCount=SSCount+1
		except:
			pass
		if(len(dfByNum[dfByNum['MaxWind'] > 95])>0):
			MHCount=MHCount+1
		Lat=dfByNum['Lat'].iloc[0]
		Lon=dfByNum['Lon'].iloc[0]
		# Now start filtering for ACE calculations
		# None of the off rows for landsfalls and such
		dfByNum.drop(dfByNum[dfByNum['Minutes'] != 0].index,inplace=True)
		# Only grab rows of 0,6,12,18
		dfByNum.drop(dfByNum[dfByNum['Hour'].mod(6) != 0].index,inplace=True)
		StormACE=dfByNum.MaxWSqd.sum()
		StormACE=round((float(StormACE)/10000),2)

		if ((0 < Lat < 21) and (-90 < Lon < -20)):
			FormReg = "MDR"
			MDRCount = MDRCount + 1
			MDRACE = MDRACE + StormACE
		elif ((Lat >= 21) and (Lon < -81) or (Lat >= 18) and (-100 < Lon < -90)):
			FormReg = "GMX"
			GMXCount = GMXCount + 1
			GMXACE = GMXACE + StormACE
		else:
			FormReg = "EXT"
			EXTCount = EXTCount + 1
			EXTACE = EXTACE  + StormACE
#		print(dfByNum['StormName'].iloc[0], str(Lat), str(Lon), FormReg, str(round(StormACE,2)), round(MDRACE,2), round(GMXACE,2), round(EXTACE,2))

	RegionalACE.append([str(yr),MDRACE,GMXACE,EXTACE])


	dfHU=dfYear[dfYear['StormStatus']=='HU']
	dfHU.drop_duplicates(subset='StormName',inplace=True)

#	print("SubTropStorms  : "+str(SSCount))
#	print("Named Storms   : "+str(len(dfYear['StormNumber'].unique())))
#	print("Hurricane      : "+str(dfHU.shape[0]))
#	print("Major Hurricane: "+str(MHCount))
	NamedStorms = len(dfYear['StormNumber'].unique())
	HU=dfHU.shape[0]

	# Now start filtering for ACE calculations
	# None of the off rows for landsfalls and such
	dfYear.drop(dfYear[dfYear['Minutes'] != 0].index,inplace=True)
	# Only grab rows of 0,6,12,18
	dfYear.drop(dfYear[dfYear['Hour'].mod(6) != 0].index,inplace=True)
	#print(dfYear)
	ACEwSS=round(dfYear['MaxWSqd'].sum()/10000,1)
#	print("ACE wSS    : "+str(ACEwSS))


#	print(str(yr),round(MDRACE,2), round(GMXACE,2), round(EXTACE,2),ACEwSS)
	SummaryList.append([yr,NamedStorms,HU, MHCount,ACEwSS])

#for vals in RegionalACE:
#	print(vals)
print(RegionalACE)
#print(convert_1d_to_2d(RegionalACE,1))
SummaryDF=pd.DataFrame(SummaryList,columns=['YEAR','NamedTS', 'HU', 'MH','ACE'])
print(SummaryDF)
SummaryDF = pd.concat([SummaryDF,SummaryDF.sum().to_frame().T], ignore_index=True)
#print(SummaryDF)
#os.system('more ./ATL_1950_Present_Stats.csv')

SummaryDF.at[len(SummaryDF)-1,'YEAR']=0000.0
SummaryDF.at[len(SummaryDF)-1,'NamedTS']=(SummaryDF.at[len(SummaryDF)-1,'NamedTS'])/Totallen
SummaryDF.at[len(SummaryDF)-1,'HU']=(SummaryDF.at[len(SummaryDF)-1,'HU'])/Totallen
SummaryDF.at[len(SummaryDF)-1,'MH']=(SummaryDF.at[len(SummaryDF)-1,'MH'])/Totallen
SummaryDF.at[len(SummaryDF)-1,'ACE']=(SummaryDF.at[len(SummaryDF)-1,'ACE'])/Totallen
print(SummaryDF)
SummaryDF.to_csv("./ATL_Aug_Stats.csv", index=False, float_format='%.1f')