library(tidyverse)
library(fixest)
library(janitor)
###########
CROP_LEGAL <- read_csv("Data/Created/Crop_Legal_Int.csv") %>% clean_names() %>% select(-fid,-fid_2,-div,-district,-master_id,-master_id_2,-polygon_id,-shape_area,-int_area,-comments,-district_2,-path,-comments_2,-shape_leng_2,-shape_area_2,-master_id_2,-gwzone_2,-comments_2,-shape_leng_2,shape_area_2,-path)
CROP_LEGAL  %>% select(farm_unit_2) %>% group_by(farm_unit_2) %>% summarize(n())
CROP_LEGAL 
colnames(CROP_LEGAL )
CROP_LEGAL
CROP_LEGAL %>% select(layer) %>% gsub(" Parcels","")
CROP_LEGAL %>% filter(name=='3-200004-0') 
/home/alex/Git/Publications/CREP_Policy_Map/Data/Processed
CROP_LEGAL %>% group_by(parcel_id) %>% filter(calc_area==max(calc_area)) %>% ungroup
read_csv("Data/Input_Data/CROPS.csv") 

#function "i" needed for a factor
crop  <- read_csv("./Data/Input_Data/CROPS.csv",col_types = cols(PARCEL_ID=col_character(),SW_WDID1=col_character(),SW_WDID2=col_character(),SW_WDID3=col_character(),SW_WDID4=col_character(),SW_WDID5=col_character(),GW_ID1=col_character(),GW_ID2=col_character(),GW_ID3=col_character(),GW_ID4=col_character(),GW_ID5=col_character(),GW_ID6=col_character(),GW_ID7=col_character(),GW_ID6=col_character(),GW_CLASS6=col_character(),GW_ID7=col_character(),GW_CLASS7=col_character(),GW_CLASS1=col_character(),GW_CLASS2 =col_character(),GW_CLASS3 =col_character(),GW_CLASS4 =col_character(),GW_CLASS5 =col_character(),GW_CLASS6 =col_character(),GW_CLASS7 =col_character() )) %>% select(PARCEL_ID,Year=CAL_YEAR,CROP_TYPE,IRR_TY=IRRIG_TYPE,CROP_AREA,GW_ID1,GW_ID2,GW_ID3,GW_ID4,GW_ID5,GW_ID6,GW_ID7,SW_WDID1,SW_WDID2,SW_WDID3,SW_WDID4,SW_WDID5)

crop$IRR_TY  <- ifelse(crop$IRR_TY=='FLOOD','F','S')

crop$CROP_TYPE <- ifelse(crop$CROP_TYPE=="NEW_ALFALFA","ALFALFA",crop$CROP_TYPE)
crop$CROP_TYPE <- ifelse(crop$CROP_TYPE=="BARLEY","SMALL_GRAINS",crop$CROP_TYPE)
#blue grass, cover crop, sorghum grain, vegetables, wheat, fall wheat, spring wheat, and corn.
crop$CROP_TYPE <- ifelse(crop$CROP_TYPE=="COVER_CROP","OTHER",crop$CROP_TYPE)
crop$CROP_TYPE <- ifelse(crop$CROP_TYPE=="VEGETABLES","OTHER",crop$CROP_TYPE)
crop$CROP_TYPE <- ifelse(crop$CROP_TYPE=="SORGHUM_GRAIN","OTHER",crop$CROP_TYPE)
crop$CROP_TYPE <- ifelse(crop$CROP_TYPE=="WHEAT","OTHER",crop$CROP_TYPE)
crop$CROP_TYPE <- ifelse(crop$CROP_TYPE=="WHEAT_FALL","OTHER",crop$CROP_TYPE)
crop$CROP_TYPE <- ifelse(crop$CROP_TYPE=="WHEAT_SPRING","OTHER",crop$CROP_TYPE)
crop$CROP_TYPE <- ifelse(crop$CROP_TYPE=="CORN","OTHER",crop$CROP_TYPE)
crop$CROP_TYPE <- ifelse(crop$CROP_TYPE=="","OTHER",crop$CROP_TYPE)
crop$POT  <- ifelse(crop$CROP_TYPE=="POTATOES",crop$CROP_AREA,0)
crop$SG<- ifelse(crop$CROP_TYPE=="SMALL_GRAINS",crop$CROP_AREA,0)
crop$ALF  <- ifelse(crop$CROP_TYPE=="ALFALFA",crop$CROP_AREA,0)
crop$OT  <- ifelse(crop$CROP_TYPE=="OTHER",crop$CROP_AREA,0)
crop$GP  <- ifelse(crop$CROP_TYPE=="GRASS_PASTURE",crop$CROP_AREA,0)

SW_LIST  <- crop %>% group_by(SW_WDID1) %>% filter(!is.na(SW_WDID1)) %>% summarize(num=n()) %>% arrange(desc(num)) %>% slice(1:10) %>% pull(SW_WDID1)
SW_COL <- paste0("SW_",SW_LIST)
SW_DUM  <- function(X){
return(ifelse(crop$SW_WDID1 %in% X | crop$SW_WDID2 %in% X | crop$SW_WDID3 %in% X | crop$SW_WDID3 %in% X | crop$SW_WDID5 %in% X,1,0)) 
}
crop$SW_2000812  <-      SW_DUM(SW_LIST[1])  
crop$SW_2000623  <-      SW_DUM(SW_LIST[2])  
crop$SW_2000631  <-      SW_DUM(SW_LIST[3])  
crop$SW_2000753  <-      SW_DUM(SW_LIST[4])  
crop$SW_2200593  <-      SW_DUM(SW_LIST[5])  
crop$SW_3500570  <-      SW_DUM(SW_LIST[6])  
crop$SW_2000829  <-      SW_DUM(SW_LIST[7])  
crop$SW_2000798 <-      SW_DUM(SW_LIST[8])  
crop$SW_2100521  <-      SW_DUM(SW_LIST[9])  
crop$SW_2200541  <-      SW_DUM(SW_LIST[10])  



wells  <- read_rds("./Data/Input_Data/Well_Pumping.rds")
mt  <- crop %>% select(PARCEL_ID,Year=Year,wdid=GW_ID1,all_of(SW_COL)) %>% distinct
m1  <- wells %>% inner_join(mt,multiple="all")
mt  <- crop %>% select(PARCEL_ID,Year=Year,wdid=GW_ID2,all_of(SW_COL)) %>% distinct
m2  <- wells %>% inner_join(mt,multiple="all")
mt  <- crop %>% select(PARCEL_ID,Year=Year,wdid=GW_ID3,all_of(SW_COL)) %>% distinct
m3  <- wells %>% inner_join(mt,multiple="all")
mt  <- crop %>% select(PARCEL_ID,Year=Year,wdid=GW_ID4,all_of(SW_COL)) %>% distinct
m4  <- wells %>% inner_join(mt,multiple="all")
mt  <- crop %>% select(PARCEL_ID,Year=Year,wdid=GW_ID5,all_of(SW_COL)) %>% distinct
m5  <- wells %>% inner_join(mt,multiple="all")
mt  <- crop %>% select(PARCEL_ID,Year=Year,wdid=GW_ID6,all_of(SW_COL)) %>% distinct
m6  <- wells %>% inner_join(mt,multiple="all")
mt  <- crop %>% select(PARCEL_ID,Year=Year,wdid=GW_ID7,all_of(SW_COL)) %>% distinct
m7  <- wells %>% inner_join(mt,multiple="all")
m  <- rbind(m1,m2,m3,m4,m5,m6,m7) %>% rename(GW_wdid=wdid)

crop  <- crop %>% select(PARCEL_ID,Year,CROP_TYPE,CROP_AREA,IRR_TY,POT,SG,ALF,GP,OT)
#Crep

CREP_PERM_2014  <- c(2013956,2005950,2005951,2005955,2014091,2006322,2006321,2014092, 2006327,2006328,2006325,2006326,2006332,2006331,2006684,2006685,2006686,2014274,2014107,2005512,2005448)
CREP_PERM_2015  <- c(2005098,2008177,2008178,2013955,2006005,2006656,2005171,2006655,2006323,2006324,2014088,2008223,2008224,2008225,2014054,2705248)
CREP_PERM_2016  <- c(2005476,2005537,2005538,2014266,2005769,2005770,2005771,2014270,2005766,2005767,2005768,2014267,2014268,2008439,2008440,2208441,2009992,2009197,2014045,2014046,2006003,2006004,2006653,2006654,2014311,2005952,2005953,2005954,2005121,2008772)
CREP_PERM_2018  <- c(2014309,2006334,2014080,2006333,2006337,2006338,2014081,2014082)
CREP_PERM_2020  <- c(2010499,2010500,2013906,2014178,2006528,2006529,2014189,2009110,2008239,2705472)
CREP_PERM_2021  <- c(2008146,2008147,2013953,2005665,2005664,2005666,2005662,2005663,2010786,2010787,2008389,2014023)
CREP_PERM_2014  <- cbind(CREP_PERM_2014,rep(2014,length(CREP_PERM_2014)),rep(1,length(CREP_PERM_2014)))
CREP_PERM_2015  <- cbind(CREP_PERM_2015,rep(2015,length(CREP_PERM_2015)),rep(1,length(CREP_PERM_2015)))
CREP_PERM_2016  <- cbind(CREP_PERM_2016,rep(2016,length(CREP_PERM_2016)),rep(1,length(CREP_PERM_2016)))  
CREP_PERM_2018  <- cbind(CREP_PERM_2018,rep(2018,length(CREP_PERM_2018)),rep(1,length(CREP_PERM_2018)))
CREP_PERM_2020  <- cbind(CREP_PERM_2020,rep(2020,length(CREP_PERM_2020)),rep(1,length(CREP_PERM_2020)))
CREP_PERM_2021  <- cbind(CREP_PERM_2021,rep(2021,length(CREP_PERM_2021)),rep(1,length(CREP_PERM_2021)))

CREP_PERM  <- rbind(CREP_PERM_2014,CREP_PERM_2015,CREP_PERM_2016,CREP_PERM_2018,CREP_PERM_2020,CREP_PERM_2021)
colnames(CREP_PERM) <- c("GW_wdid","CREP_Year","PERM_CREP") 



CREP_TEMP_2014 <- c(2005642,2005643,2014474,2006153,2013962,2006478,2008677,2008678,2012887,2005857,2008391,2705126,2705519,2706148,2705342,2706196,2705341,2706195)
CREP_TEMP_2015 <- c(2008155,2008156,2008129,2008130,2014244,2005921,2005941,2006283,2006525,2006335,2006336,2014086,2014087,2705318,2012537,2014288,2706253,2705317,2705186,2705328,2705054)
CREP_TEMP_2016 <- c(2005133,2005533,2005886,2005868,2005595,2013377,2013618,2706014,2706246)
CREP_TEMP_2017 <- c(2009113,2705344,2705067,2705068,2705523,2705069,2705070)
CREP_TEMP_2018 <- c(2005774,2005775,2705293,2705290,2705224,2705225,2705225,2705224,2705197,2705359,2705246,2706237,2705006,2705790,2705184,2705185,2705356,2705327)
CREP_TEMP_2019 <- c(2006376,2006375,2005127,2005168,2013784,2705259,2705021,2705020,2706194)
CREP_TEMP_2020 <- c(2705347,2706258)
CREP_TEMP_2022  <- c(2006678,2006679,2005923,2005116)
CREP_TEMP_2023  <- c(2009084,2009082,2013887,2009940,2013577,2005038,2010590,2014490,2705247,2705080)

CREP_TEMP_2014  <- cbind(CREP_TEMP_2014,rep(2014,length(CREP_TEMP_2014)),rep(0,length(CREP_TEMP_2014)))
CREP_TEMP_2015  <- cbind(CREP_TEMP_2015,rep(2015,length(CREP_TEMP_2015)),rep(0,length(CREP_TEMP_2015)))
CREP_TEMP_2016  <- cbind(CREP_TEMP_2016,rep(2016,length(CREP_TEMP_2016)),rep(0,length(CREP_TEMP_2016)))
CREP_TEMP_2017  <- cbind(CREP_TEMP_2017,rep(2017,length(CREP_TEMP_2017)),rep(0,length(CREP_TEMP_2017)))
CREP_TEMP_2018  <- cbind(CREP_TEMP_2018,rep(2018,length(CREP_TEMP_2018)),rep(0,length(CREP_TEMP_2018)))
CREP_TEMP_2019  <- cbind(CREP_TEMP_2019,rep(2019,length(CREP_TEMP_2019)),rep(0,length(CREP_TEMP_2019)))
CREP_TEMP_2020  <- cbind(CREP_TEMP_2020,rep(2020,length(CREP_TEMP_2020)),rep(0,length(CREP_TEMP_2020)))
#CREP_TEMP_2022  <- cbind(CREP_TEMP_2022,rep(2022,length(CREP_TEMP_2022)),rep(0,length(CREP_TEMP_2022)))
#CREP_TEMP_2023  <- cbind(CREP_TEMP_2023,rep(2023,length(CREP_TEMP_2023)),rep(0,length(CREP_TEMP_2023)))
CREP_TEMP  <- rbind(CREP_TEMP_2014,CREP_TEMP_2015,CREP_TEMP_2016,CREP_TEMP_2017,CREP_TEMP_2018,CREP_TEMP_2019,CREP_TEMP_2020)
colnames(CREP_TEMP) <- c("GW_wdid","CREP_Year","PERM_CREP") 
IN_CREP <- rbind(CREP_TEMP,CREP_PERM) %>% as_tibble %>% mutate(GW_wdid=as.character(GW_wdid))

FALL_2020  <- c(2013693,2005731,2013884,2706159,2705471,2705395,2705517)
FALL_2020 <- cbind(FALL_2020,rep(2020,length(FALL_2020)),rep(1,length(FALL_2020)))
FALL_2021 <- c(2008226,2014273,2008591,2009461,2009462,2009114,2010568,2005035,2014173)
FALL_2021 <- cbind(FALL_2021,rep(2021,length(FALL_2021)),rep(1,length(FALL_2021)))
FALL_PROG  <- rbind(FALL_2020,FALL_2021)  %>% as_tibble
colnames(FALL_PROG)  <- c("GW_wdid","FALL_PROG_YEAR","FALL_PROG")

#Legal Parcel
LP  <- read_csv("./Data/Input_Data/LEGAL.csv") %>% rename(County=Layer) %>% select(-fid)
LP$County <- ifelse(is.na(LP$County),"Saguache",LP$County)
LP$County<- ifelse(LP$County=="Saguache",'SG',LP$County)
LP$County<- ifelse(LP$County=="Saguache Parcels",'SG',LP$County)
LP$County<- ifelse(LP$County=="Alamosa Parcels",'AL',LP$County)
LP$County <- ifelse(LP$County=="Rio Grande Parcels",'RG',LP$County)
LP$County <- ifelse(LP$County=="Conejos Parcels",'CN',LP$County)
#Crop to Legal Parcel
CL  <- read_csv("./Data/Input_Data/CROP_LEGAL.csv",col_types=cols(PARCEL_ID=col_character())) %>% select(Name,PARCEL_ID,Year=CAL_YEAR) 


SL <- read_csv("Data/Input_Data/SBD_LEGAL.csv") 
SL$SBD1  <-  ifelse(SL$SBD==1,1,0)
SL$SBD2  <-  ifelse(SL$SBD==2,1,0)
SL$SBD3  <-  ifelse(SL$SBD==3,1,0)
SL$SBD4  <-  ifelse(SL$SBD==4,1,0)
SL$SBD5  <-  ifelse(SL$SBD==5,1,0)
SL$SBD6  <-  ifelse(SL$SBD==6,1,0)
SL  <- SL %>% select(-SBD) %>% group_by(Name) %>% mutate(SBD1=max(SBD1),SBD2=max(SBD2),SBD3=max(SBD3),SBD4=max(SBD4),SBD5=max(SBD5),SBD6=max(SBD6)) %>% distinct
SL  <- LP %>% select(Name) %>% distinct %>% left_join(SL) %>% 	mutate(across(everything(), ~replace_na(.x, 0)))
LP  <- LP %>% left_join(SL)

CS  <- read_csv("Data/Input_Data/SBD_CROP.csv",col_types=cols(PARCEL_ID=col_character())) %>%  rename(Year=CAL_YEAR) %>% distinct
CS$SBD1  <-  ifelse(CS$SBD==1,1,0)
CS$SBD2  <-  ifelse(CS$SBD==2,1,0)
CS$SBD3  <-  ifelse(CS$SBD==3,1,0)
CS$SBD4  <-  ifelse(CS$SBD==4,1,0)
CS$SBD5  <-  ifelse(CS$SBD==5,1,0)
CS$SBD6  <-  ifelse(CS$SBD==6,1,0)
CS  <- CS %>% select(-SBD) %>% group_by(PARCEL_ID,Year) %>% mutate(SBD1=max(SBD1),SBD2=max(SBD2),SBD3=max(SBD3),SBD4=max(SBD4),SBD5=max(SBD5),SBD6=max(SBD6)) %>% distinct
CS  <- crop %>% select(PARCEL_ID,Year) %>% left_join(CS) %>%	mutate(across(everything(), ~replace_na(.x, 0)))
m  <- m %>% group_by(GW_wdid) %>% mutate(SW_2000812=max(SW_2000812),SW_2000623=max(SW_2000623),SW_2000631=max(SW_2000631),SW_2000753=max(SW_2000753),SW_2200593=max(SW_2200593),SW_3500570=max(SW_3500570),SW_2000829=max(SW_2000829),SW_2000798=max(SW_2000798),SW_2100521=max(SW_2100521),SW_2200541=max(SW_2200541)) %>% ungroup
WSW <- m %>% select(-Year,-AF,-PARCEL_ID) %>% unique
GW_CREP  <- m %>% left_join(CL,multiple="all") %>% left_join(IN_CREP) 
GW_CREP$IN_CREP  <- ifelse(is.na(GW_CREP$CREP_Year),0,1)

#Well data set
well_df  <- GW_CREP %>% left_join(m)
well_df <- well_df %>% full_join(crop)
well_df  <- well_df %>% left_join(CS)
#well_df %>% filter(SBD1==0,SBD2==0,SBD3==0,SBD4==0,SBD5==0,SBD6==0)
well_df[which(is.na(well_df$GW_wdid)),1]  <- "NONE"
well_df$AF<- ifelse(is.na(well_df$AF),0,well_df$AF)
#One point is -0.125 checked Hydrobase to confirm values. Either bad entry or augmentation
well_df[well_df$AF<0,]$AF  <- 0
#well_df$IN_CREP<- ifelse(is.na(well_df$IN_CREP),0,well_df$IN_CREP)
well_df$IN_CREP  <- well_df$GW_wdid %in% (IN_CREP %>% pull(GW_wdid))

well_df  <- well_df %>% group_by_at(c(c("GW_wdid","Year","AF"),SW_COL)) %>% summarize(IN_CREP=max(IN_CREP),CREP_Year=min(CREP_Year,na.rm=TRUE),CROP_AREA=sum(CROP_AREA),POT=sum(POT),ALF=sum(ALF),GP=sum(GP),SG=sum(SG),OT=sum(OT),SBD1=max(SBD1),SBD2=max(SBD2),SBD3=max(SBD3),SBD4=max(SBD4),SBD5=max(SBD5),SBD6=max(SBD6)) 
well_df$CREP_Year[is.infinite(well_df$CREP_Year)]  <- NA
temp  <- well_df %>% group_by(GW_wdid) %>% mutate(ST_YEAR=min(Year))
well_df  <- well_df %>% select(GW_wdid,IN_CREP,CREP_Year,SBD1,Year,AF,CROP_AREA,POT,ALF,SG,GP,OT,all_of(SW_COL),SBD2,SBD3,SBD4,SBD5,SBD6)

LIST2009 <- well_df %>% ungroup %>% filter(Year==2009) %>% pull(GW_wdid) %>% unique
MISSING2009  <- well_df[which(!(well_df$GW_wdid %in% LIST2009)),] %>% ungroup
MISSING2009  <- MISSING2009 %>%   select(GW_wdid,IN_CREP,CREP_Year,SBD1,SBD2,SBD3,SBD4,SBD5,SBD6) %>% unique %>% mutate(Year=2009,AF=0,CROP_AREA=0,POT=0,ALF=0,SG=0,GP=0,OT=0) %>% left_join(WSW)
well_df  <- rbind(MISSING2009,well_df)

for(x in 2009:2021){
	S_YEAR  <- x;
	MISS <- well_df %>% group_by(GW_wdid,IN_CREP,CREP_Year,SBD1,SBD2,SBD3,SBD4,SBD5,SBD6) %>% filter(min(Year)<=S_YEAR,!any(Year==S_YEAR+1)) %>% select(GW_wdid,IN_CREP,CREP_Year,SBD1,SBD2,SBD3,SBD4,SBD5,SBD6) %>% ungroup %>% unique %>% mutate(Year=S_YEAR+1,AF=0,CROP_AREA=0,POT=0,ALF=0,SG=0,GP=0,OT=0) %>% left_join(WSW)

	well_df  <- rbind(well_df,MISS)
	print("done")
}
well_df  <- well_df %>% group_by(GW_wdid,Year) %>%  mutate(AF=sum(AF),IN_CREP=max(IN_CREP),CREP_Year=min(CREP_Year),CROP_AREA=sum(CROP_AREA),BOTH=ifelse(n()>1,1,0),POT=sum(POT),ALF=sum(ALF),SG=sum(SG),GP=sum(GP),OT=sum(OT),SBD1=max(SBD1),SW_2000812=max(SW_2000812),SW_2000623=max(SW_2000623),SW_2000631=max(SW_2000631),SW_2000753=max(SW_2000753),SW_2200593=max(SW_2200593),SW_3500570=max(SW_3500570),SW_2000829=max(SW_2000829),SW_2000798=max(SW_2000798),SW_2100521=max(SW_2100521),SW_2200541=max(SW_2200541),SBD1=max(SBD1),SBD2=max(SBD2),SBD3=max(SBD3),SBD4=max(SBD4),SBD5=max(SBD5),SBD6=max(SBD6)) %>% unique %>% ungroup

well_df$Post_11  <- ifelse(well_df$Year>=2011,1,0)
well_df  <- well_df %>% mutate(POST=ifelse(is.na(CREP_Year),0,ifelse(Year >=CREP_Year,1,0))) %>% select(GW_wdid,Year,POST,everything())
well_df  <- well_df %>% group_by(GW_wdid) %>% mutate(M_AF=max(AF),M_CROP=max(CROP_AREA)) %>% ungroup
well_df  <- well_df %>% filter(M_AF!=0)
	#only 2 2005 entries
well_df  <- well_df %>% filter(Year>2005)

##############Well Distances
AF_WELL  <- well_df %>% group_by(GW_wdid) %>% summarize(AF=max(AF)) %>% arrange(GW_wdid)
well_list  <- AF_WELL %>% filter(AF>0) %>% pull(GW_wdid)
WD <- read_csv("Data/Input_Data/Wells_Distance_Matrix.csv",col_types = cols(ID=col_character())) %>% rename(GW_wdid=ID)
##Keep only well records with pumping data, then arrange by well number for consistency
NM_SAVE  <- WD[WD$GW_wdid %in% t(well_list),1]
###DISTANCE MATRIX
WD  <- 	WD[WD$GW_wdid %in% well_list,colnames(WD) %in% well_list] 
WD_FULL <- cbind(NM_SAVE,WD) %>% as_tibble

##Data Clean
well_df$Yearf  <- as.factor(well_df$Year)
well_df$GW_wdid  <- as.factor(well_df$GW_wdid)
well_df$LAF  <- log(well_df$AF+0.00001)
well_df  <- well_df %>% rename(CREP_YEAR=CREP_Year)
well_df  <- well_df %>% filter(Year!=2022)
well_df  <- well_df %>%  mutate(CREP_rel_year=Year-CREP_YEAR,CREP_rel_year=replace_na(CREP_rel_year,Inf))
###Logit model of Fallowing
well_df$CREP_PRE <- ifelse(well_df$CREP_rel_year %in% c(-1,2),1,0)
well_df$CREP_PRE1 <- ifelse(well_df$CREP_rel_year %in% c(-1),1,0)
well_df$CREP_PRE2 <- ifelse(well_df$CREP_rel_year %in% c(-2),1,0)
well_df$CREP_PRE3 <- ifelse(well_df$CREP_rel_year %in% c(-3),1,0)
well_df$SUNAB_CREP  <- ifelse(well_df$IN_CREP==0,100000,well_df$CREP_YEAR)
well_df <- well_df %>% left_join(IN_CREP %>% group_by(GW_wdid) %>% summarize(PERM_CREP=max(PERM_CREP),TEMP_CREP=as.numeric(min(PERM_CREP)==0))) %>% mutate(PERM_CREP=ifelse(is.na(PERM_CREP),0,PERM_CREP),TEMP_CREP=ifelse(is.na(TEMP_CREP),0,TEMP_CREP))
PERM_CREP  <- IN_CREP %>% filter(PERM_CREP==1) %>% group_by(GW_wdid) %>% summarize(PERM_YEAR=min(CREP_Year))
TEMP_CREP  <- IN_CREP %>% filter(PERM_CREP==0) %>% group_by(GW_wdid) %>% summarize(TEMP_YEAR=min(CREP_Year))
well_df  <- well_df %>% left_join(TEMP_CREP ) %>% mutate(ifelse(is.na(TEMP_CREP),0,TEMP_CREP))
well_df  <- well_df %>% left_join(PERM_CREP ) %>% mutate(ifelse(is.na(PERM_CREP),0,PERM_CREP))
well_df  <- well_df %>% mutate(CREP_YEAR=ifelse(is.na(CREP_YEAR),Inf,CREP_YEAR),TEMP_YEAR=ifelse(is.na(TEMP_YEAR),Inf,TEMP_YEAR),PERM_YEAR=ifelse(is.na(PERM_YEAR),Inf,PERM_YEAR))
well_df$pumping_fee  <- ifelse(well_df$Year==2011,45,NA)
well_df$pumping_fee  <- ifelse(well_df$Year>2011 & well_df$Year<2018,75,well_df$pumping_fee)
well_df$pumping_fee  <- ifelse(well_df$Year==2018 | well_df$Year==2019,90,well_df$pumping_fee)
well_df$pumping_fee  <- ifelse(well_df$Year>2019 ,150,well_df$pumping_fee)
well_df$Post_16  <- ifelse(well_df$Year>=2016,1,0)
well_df$Post_11  <- ifelse(well_df$Year>=2011,1,0)

#df_half$SUNAB  <- ifelse(df_half$CLOSE_CREP==0,100000,df_half$CREP_CLOSE_START_YEAR)
well_df$FALLOW  <- ifelse(well_df$AF==0,1,0)
well_df$BETWEEN<- ifelse(well_df$IN_CREP==1 & well_df$Year<well_df$CREP_YEAR,1,0)
well_df <- 	well_df %>% left_join(FALL_PROG %>% mutate(GW_wdid=as.character(GW_wdid))) %>% rename(IN_FALL_PROG=FALL_PROG)%>% mutate(IN_FALL_PROG=ifelse(is.na(IN_FALL_PROG),0,1),FALL_PROG=ifelse(IN_FALL_PROG==1 & Year>=FALL_PROG_YEAR,1,0 ))
well_df$SBD1_Post <- ifelse(well_df$Post_11 & well_df$SBD1==1,1,0)
well_df$CREP_Post <- ifelse(well_df$Post_11 & well_df$IN_CREP==1,1,0)
well_df$SBD_ref_11  <- ifelse(well_df$SBD1==1,well_df$Year,Inf)
well_df$CREP_ref_11  <- ifelse(well_df$IN_CREP==1 & well_df$POST!=1,well_df$Year,Inf)
well_df$IND_CREP11  <- i(well_df$CREP_ref_11,ref=c(2009,2010,Inf))
well_df$IND_SBD11  <- i(well_df$SBD_ref_11,ref=c(2009,2010,Inf))
well_df  <- well_df %>% select(-CREP_ref_11,-SBD_ref_11)
well_df$FALL_PROG_YEAR  <- ifelse(is.na(well_df$FALL_PROG_YEAR),Inf,well_df$FALL_PROG_YEAR)

source("Scripts/Distance_Functions.r")
GET_DIS_DATA_SET  <- function(DIS,WELL_DF=well_df){
	DIS_DF  <- WELL_DF %>% left_join(GET_CREP_DAT(DIS) )
	DIS_DF$CLOSE_CREP  <- ifelse(DIS_DF$CREP_CLOSE_YEAR<100000,1,0)
	DIS_DF$CLOSE_BETWEEN <- ifelse(DIS_DF$Post_11 & DIS_DF$CLOSE_CREP==1 & DIS_DF$CREP_CLOSE_YEAR<DIS_DF$Year ,1,0)
	DIS_DF$POST_CLOSE <- ifelse(DIS_DF$Post_11 & DIS_DF$CLOSE_CREP==1 & DIS_DF$CREP_CLOSE_YEAR>=DIS_DF$Year ,1,0)
	#All Fallow plots
		DIS_DF  <- DIS_DF %>% left_join(DF_CLOSE_FALL(DIS))
	#Data from 4 year Fallow program of SBD1
		DIS_DF <- DIS_DF %>% left_join(GET_FALLOW_PROG_DAT(DIS))
		DIS_DF <- DIS_DF %>% left_join(GET_FALLOW_PROG_DAT(DIS))
		DIS_DF$FALL_PROG_CLOSE_YEAR  <- ifelse(is.na(DIS_DF$FALL_PROG_CLOSE_YEAR),Inf,DIS_DF$FALL_PROG_CLOSE_YEAR)
		DIS_DF$FALL_PROG_CLOSE_POST  <- ifelse(DIS_DF$FALL_PROG==1 & DIS_DF$FALL_PROG_CLOSE_YEAR>=DIS_DF$Year,1,0)
	return(DIS_DF)
}
df_quart  <- GET_DIS_DATA_SET(5280/4) 
df_half  <- GET_DIS_DATA_SET(5280/2) 
df_mile  <- GET_DIS_DATA_SET(5280) 
df_two  <- GET_DIS_DATA_SET(2*5280) 
#########################ADDING IN PROBIT MODEL OF ENTERING CREP DUE TO FEE
PRE_DF  <- df_half %>% group_by(GW_wdid,IN_CREP,CLOSE_CREP)%>% filter(Year<2011,SBD1==1)  %>% summarize(M_AF=max(AF),PRE_AF=mean(AF),PRE_CROP=mean(POT+ALF+SG+OT),PRE_POT=mean(POT)/PRE_CROP,PRE_ALF=mean(ALF)/PRE_CROP,PRE_SG=mean(SG)/PRE_CROP,PRE_OT=mean(OT)/PRE_CROP) %>% filter(PRE_CROP>0)
POST_DF  <- df_half %>% group_by(GW_wdid,IN_CREP,CLOSE_CREP)%>% filter(Year>=2011,Year<2014,SBD1==1)  %>% summarize(POST_AF=mean(AF),POST_CROP=mean(POT+ALF+SG+OT,na.rm=TRUE),POST_POT=mean(POT,na.rm=TRUE)/POST_CROP,POST_ALF=mean(ALF,na.rm=TRUE)/POST_CROP,POST_SG=mean(SG,na.rm=TRUE)/POST_CROP,POST_OT=mean(OT,na.rm=TRUE)/POST_CROP)
SW_DF  <- df_half[,c(1,19:28)] %>% unique
logit_df <- PRE_DF %>% left_join(POST_DF) %>% mutate(DIFF_AF=POST_AF-PRE_AF,DIFF_CROP=PRE_CROP-POST_CROP) %>% left_join(SW_DF)