---
title: "Compare scenarios Transport"
output:
  html_document:
    df_print: paged
---

```{r, echo=FALSE, message=FALSE, warning=FALSE}
require(ggplot2)
require(moinput)
require(data.table)
require(dplyr)
require(remind)
require(gdxdt)
require(gdx)
require(rmndt)
require(magclass)
require(quitte)
require(ggpubr)
require(gridExtra)
#require(edgeTrpLib)
require(devtools)
load_all("../../../../../../edgetrplib/")

```


```{r, echo=FALSE, warning=FALSE}
iso_plot = "DEU"
output_folder = "EDGE-T/"

cols <- c("NG" = "#d11141",
          "Liquids" = "#8c8c8c",
          "Hybrid Liquids" = "#ffc425",
          "Hybrid Electric" = "#f37735",
          "BEV" = "#00b159",
          "FCEV" = "#00aedb")

datapath <- function(fname){
  file.path("./EDGE-T/", fname)
}

mapspath <- function(fname, scenariopath=""){
    file.path("../../modules/35_transport/edge_esm/input", fname)
}


## Load mappings
REMIND2ISO_MAPPING <- fread("../../config/regionmappingH12.csv")[, .(iso = CountryCode, region = RegionCode)]
EDGE2teESmap <- fread(mapspath("mapping_EDGE_REMIND_transport_categories.csv"))

## load input data from last EDGE run
demand_km <- readRDS(datapath(fname = "demandF_plot_pkm.RDS")) ## detailed energy services demand, million km
demand_ej <- readRDS(datapath(fname = "demandF_plot_EJ.RDS")) ## detailed final energy demand, EJ
vintcomp <- readRDS(datapath(fname = "vintcomp.RDS"))
newcomp <- readRDS(datapath(fname = "newcomp.RDS"))
shares <- readRDS(datapath(fname = "shares.RDS"))
pref <- readRDS(datapath(fname = "pref_output.RDS"))
EF_shares <- readRDS(datapath(fname = "EF_shares.RDS"))
annual_sales <- readRDS(datapath(fname = "annual_sales.RDS"))
mj_km_data <- readRDS(datapath(fname = "mj_km_data.RDS"))
stations <- readRDS(datapath(fname = "stations.RDS"))

name_mif = list.files(pattern = "REMIND_generic", full.names = F)
name_mif = name_mif[!grepl("withoutPlu", name_mif)]
miffile <- as.data.table(read.quitte(name_mif))
```

# LDVs vintages

```{r, echo=FALSE, warning=FALSE}

plotVint = function(vintcomp, newcomp, sharesVS1){
  vintcomp = vintcomp[,.(totdem, iso, subsector_L1, year, technology,vehicle_type, sector, sharetech_vint)]
  newcomp = newcomp[,.(iso, subsector_L1, year, technology,vehicle_type, sector, sharetech_new)]

  allfleet = merge(newcomp, vintcomp, all =TRUE, by = c("iso", "sector", "subsector_L1", "vehicle_type", "technology",  "year"))
  allfleet = merge(allfleet, sharesVS1[,.(shareVS1 = share, iso, year, vehicle_type, subsector_L1)], all.x=TRUE, by = c("iso", "year", "vehicle_type", "subsector_L1"))
  allfleet[,vintdem:=totdem*sharetech_vint*shareVS1]
  allfleet[,newdem:=totdem*sharetech_new*shareVS1]
  allfleet=melt(allfleet, id.vars = c("iso", "sector", "subsector_L1", "vehicle_type", "technology",
                                      "year"), measure.vars = c("vintdem", "newdem"))
  allfleet[,alpha:=ifelse(variable == "vintdem", 0, 1)]

  load_factor = 2
  annual_mileage = 15000
  allfleet = allfleet[,.(value = sum(value/load_factor/annual_mileage)), by = c("iso", "technology", "variable", "year")]

  allfleet = merge(allfleet, REMIND2ISO_MAPPING, by = "iso")
  allfleet = allfleet[,.(value = sum(value)), by = c("region", "technology", "variable", "year")]
  allfleet[,alphaval := ifelse(variable =="vintdem", 1,0)]


  p = ggplot()+
  geom_bar(data = allfleet[year %in% c(2015, 2030, 2050, 2100)],
  aes(x=as.character(year),y=value, group=interaction(variable, technology),
  fill = technology), alpha = 0.5, position="stack", stat = "identity", width = 0.5)+
  geom_bar(data = allfleet[year %in% c(2015, 2030, 2050, 2100)],
  aes(x=as.character(year),y=value, group=interaction(variable, technology),
  fill = technology, alpha = factor(alphaval)), position="stack", stat = "identity", width = 0.5, color = "black", size = 0.05)+
  guides(fill = guide_legend(reverse=TRUE))+
  theme_minimal()+
    facet_wrap(~region, nrow = 4)+
  theme(axis.text.x = element_text(angle = 90, hjust = 1),
        axis.text = element_text(size=7),
        title = element_text(size=8),
        legend.text = element_text(size=8))+
      scale_x_discrete(breaks = c(2015, 2030, 2050, 2100))+
  scale_alpha_discrete(breaks = c(1,0), name = "Status", labels = c("Vintages","New additions")) +
  guides(linetype=FALSE,
         fill=guide_legend(reverse=FALSE, title="Transport mode"))+
  scale_fill_manual(values = cols)+
  labs(y = "LDV fleet [million Veh]", x="")

  return(p)
  }


p = plotVint(vintcomp, newcomp, shares$VS1_shares)

p

```

# Inconvenience cost trend

```{r, echo=FALSE, warning=FALSE}

plotinconv = function(inco_tech, iso_plot, vehicle_type){
  p=ggplot()+
  geom_bar(data = inco_tech[iso == iso_plot & subsector_L1 == "trn_pass_road_LDV_4W" & vehicle_type == vehicle_type & year<=2100 & year>=2010], aes(x = as.character(year), y = value, group = logit_type, fill = logit_type), position = position_stack(), stat = "identity")+
  facet_grid(~technology)+
  theme_minimal()+
  # scale_fill_manual(values = cols)+
  expand_limits(y = c(0,0.8))+
  scale_x_discrete(breaks = c(2015,2050,2100))+
  theme(axis.text.x = element_text(angle = 90, vjust = +0.1),
        strip.background = element_rect(color = "grey"))+
  labs(x = "", y = "Inconvenience cost [$/pkm]", title = paste0("Example of ", iso_plot))

return(p)
}

plotinconv(inco_tech = pref$FV_final_pref[subsector_L1 == "trn_pass_road_LDV_4W"], iso_plot = "DEU", vehicle_type = "Large Car and SUV")
plotinconv(inco_tech = pref$FV_final_pref[subsector_L1 == "trn_pass_road_LDV_4W"], iso_plot = "USA", vehicle_type = "Large Car")
plotinconv(inco_tech = pref$FV_final_pref[subsector_L1 == "trn_pass_road_LDV_4W"], iso_plot = "JPN", vehicle_type = "Large Car")
plotinconv(inco_tech = pref$FV_final_pref[subsector_L1 == "trn_pass_road_LDV_4W"], iso_plot = "CHN", vehicle_type = "Large Car")
plotinconv(inco_tech = pref$FV_final_pref[subsector_L1 == "trn_pass_road_LDV_4W"], iso_plot = "IND", vehicle_type = "Large Car")

```

# Share of stations

```{r, echo=FALSE, warning=FALSE}

stationsplot = function(stations, iso_plot){
  p = ggplot()+
    geom_line(data = stations[iso == iso_plot], aes(x= year, y = fracst))+
    facet_grid(~technology)+
    theme_minimal()+
    labs(title = paste0("Stations for each fuel, ", iso_plot))

  return(p)
}

stationsplot(stations, iso_plot)
```


# Endogenous intensity for Liquids

```{r, echo=FALSE, message=FALSE, warning=FALSE}
## Choice of the energy intensity (of the new sales)
intcompplotf = function(EF_shares, FV_shares, VS1_shares){
  EF_shares = EF_shares[,c("iso", "year", "technology", "vehicle_type", "subsector_L1", "subsector_L2", "subsector_L3", "sector", "share","type")]
  setnames(EF_shares, old="share", new = "shareINT")
  FV_shares = FV_shares[iso == iso_plot & subsector_L1 == "trn_pass_road_LDV_4W" & technology == "Liquids"]
  setnames(FV_shares, old="share", new = "shareF")
  VS1_shares = VS1_shares[iso == iso_plot & subsector_L1 == "trn_pass_road_LDV_4W"]
  shares_LDV = merge(FV_shares, EF_shares, all = FALSE, by = c("iso", "year", "technology", "vehicle_type", "subsector_L1"))

  shares_LDV[, shareIF := shareF*shareINT]
  shares_LDV <- shares_LDV[,.(shareIF=sum(shareIF)),by=c("iso","technology","type","vehicle_type","subsector_L1", "year")]

  shares_LDV = merge(shares_LDV, VS1_shares, all = TRUE, by = c("iso", "year", "vehicle_type", "subsector_L1"))
  shares_LDV[, shareIS1 := shareIF*share]
  shares_LDV <- shares_LDV[,.(shareIS1=sum(shareIS1)),by=c("iso","type", "technology","subsector_L1","year")]
  p = ggplot()+
    geom_bar(data = shares_LDV[year<=2100  & year>=2025], aes(x=year,y=shareIS1, group = technology, fill = technology), alpha = 0.5, position = position_fill(), stat = "identity")+
    geom_bar(data = shares_LDV[year<=2100  & year>=2025], aes(x=year,y=shareIS1, group = technology, fill = technology, alpha = type), position = position_fill(), stat = "identity")+
    theme_minimal()+
    expand_limits(y = c(0,1))+
    scale_fill_manual("technology", values = cols)+
    scale_alpha_discrete("Type")+
    labs(y = "Share [%]", title = paste0("Energy intensity new sales of Liquids, example for ", iso_plot))

  return(p)

}

intcompplotf(EF_shares, shares$FV_shares, shares$VS1_shares)
```

# Sales of LDVs

```{r, echo=FALSE, warning=FALSE}
salesplot= function(shares_LDV, newcomp, sharesVS1){
  ## I need the total demand for each region to get the average composition in Europe (sales are on a country level)
  ## First I calculate the total demand for new sales using the shares on FV level (in newcomp) and on VS1 level
  newcomp = merge(newcomp, sharesVS1[,.(shareVS1 = share, iso, year, vehicle_type, subsector_L1)], all.x=TRUE, by = c("iso", "year", "vehicle_type", "subsector_L1"))
  newcomp[, newdem := totdem*sharetech_new*shareVS1]
  newcomp = newcomp[,.(value = sum(newdem)), by = c("iso", "year", "subsector_L1")]
  ## I have to interpolate in time the sales nto to loose the sales composition annual values
  newcomp=approx_dt(dt=newcomp, unique(shares_LDV$year),
                    xcol= "year",
                    ycol = "value",
                    idxcols=c("iso","subsector_L1"),
                    extrapolate=T)

  setnames(newcomp, new = "newdem", old = "value")

  ## I calculate the sales composition (disrespective to the vehicle type)
  shares_LDV = unique(shares_LDV[,c("iso","year", "technology", "shareFS1")])
  shares_LDV <- shares_LDV[,.(shareFS1=sum(shareFS1)),by=c("iso","technology","year")]

  ## I calculate the weighted regional sales (depending on the total volume of sales per country in each region)
  shares_LDV = merge(shares_LDV, newcomp)
  shares_LDV = merge(shares_LDV, REMIND2ISO_MAPPING, by = "iso")
  shares_LDV[, demfuel := shareFS1*newdem, by = c("year", "iso", "technology")]
  shares_LDV = shares_LDV[, .(demfuel = sum(demfuel)), by = c("year", "region", "technology")]
  shares_LDV[, shareFS1 := demfuel/sum(demfuel), by = c("year", "region")]

  ## plot features
  shares_LDV[, technology := factor(technology, levels = c("BEV", "Hybrid Electric", "FCEV", "Hybrid Liquids", "Liquids", "NG"))]


    plot = ggplot()+
    geom_bar(data = shares_LDV, aes(x=as.numeric(as.character(year)),y=shareFS1, group = technology, fill = technology), position = position_stack(), stat = "identity")+
    theme_minimal()+
    facet_grid(~ region )+
    scale_fill_manual("Technology", values = cols)+
    expand_limits(y = c(0,1))+
    scale_x_continuous(breaks = c(2015,2030,2050, 2100))+
#+
    theme(
axis.text.x = element_text(angle = 90, vjust=0.5, hjust=1, size = 12),
        axis.text.y = element_text(size=8),
          axis.line = element_line(size = 0.5, colour = "grey"),
            axis.title = element_text(size = 12),
          title = element_text(size = 12),
         # legend.text = element_text(12),
          #legend.title = element_text(size = 12),
          strip.text = element_text(size=12),
          strip.background = element_rect(color = "grey")
)+
    labs(x = "", y = "[%]", title = "Market share of new LDV sales")
  return(plot)
}

salesplot(annual_sales, newcomp, shares$VS1_shares)

```

# Final energy demand

```{r, echo=FALSE, warning=FALSE}
demandEJplotf = function(demandEJ){
  ## EDGE results
  demandEJ <- demandEJ[, c("sector", "subsector_L3", "subsector_L2", "subsector_L1", "vehicle_type", "technology", "iso", "year", "demand_EJ")]

  ## attribute aggregated mode and vehicle names for plotting purposes, and aggregate
  demandEJ[, aggr_mode := ifelse(subsector_L2 == "trn_pass_road_LDV", "LDV", NA)]
  demandEJ[, aggr_mode := ifelse(subsector_L3 %in% c("Passenger Rail", "HSR", "International Aviation", "Domestic Aviation"), "Pass non LDV", aggr_mode)]
  demandEJ[, aggr_mode := ifelse(subsector_L2 %in% c("trn_pass_road_bus", "Bus"), "Pass non LDV", aggr_mode)]
  demandEJ[, aggr_mode := ifelse(is.na(aggr_mode), "Freight", aggr_mode)]
  demandEJ[, veh := ifelse(grepl("Large|SUV|Midsize|Multipurpose Vehicle|Van|3W Rural", vehicle_type), "Large Cars", NA)]
  demandEJ[, veh := ifelse(grepl("Subcompact|Compact|Mini|Three-Wheeler", vehicle_type), "Small Cars", veh)]
  demandEJ[, veh := ifelse(grepl("Motorcycle|Moped|Scooter", vehicle_type), "Motorbikes", veh)]
  demandEJ[, veh := ifelse(grepl("bus|Bus", vehicle_type), "Bus", veh)]
  demandEJ[, veh := ifelse(grepl("Truck", vehicle_type) & vehicle_type != "Light Truck and SUV", "Truck", veh)]
  demandEJ[, veh := ifelse(grepl("Freight Rail_tmp_vehicletype", vehicle_type), "Freight Rail", veh)]
  demandEJ[, veh := ifelse(grepl("Passenger Rail|HSR", vehicle_type), "Passenger Rail", veh)]
  demandEJ[, veh := ifelse(subsector_L3 == "Domestic Ship", "Domestic Shipping", veh)]
  demandEJ[, veh := ifelse(subsector_L3 == "International Ship", "International Shipping", veh)]
  demandEJ[, veh := ifelse(subsector_L3 == "Domestic Aviation", subsector_L3, veh)]
  demandEJ[, veh := ifelse(subsector_L3 == "International Aviation", subsector_L3, veh)]
  demandEJ[, veh := ifelse(is.na(veh), vehicle_type, veh)]
  demandEJ = demandEJ[,.(demand_EJ = sum(demand_EJ)), by = c("iso", "year", "aggr_mode", "veh")]

  demandEJ[, vehicle_type_plot := factor(veh, levels = c("LDV","Freight Rail", "Truck","Domestic Shipping", "International Shipping",
                                                                "Motorbikes", "Small Cars", "Large Cars", "Van",
                                                                "Domestic Aviation", "International Aviation", "Bus", "Passenger Rail",
                                                                "Freight", "Freight (Inland)", "Pass non LDV", "Pass non LDV (Domestic)"))]

  legend_ord <- c("Freight Rail", "Truck", "International Shipping","Domestic Shipping",
                  "Motorbikes", "Small Cars", "Large Cars", "Van",
                  "International Aviation", "Domestic Aviation","Bus", "Passenger Rail",
                  "Freight", "LDV", "Pass non LDV", "Freight (Inland)", "Pass non LDV (Domestic)")

  demandEJ = merge(demandEJ, REMIND2ISO_MAPPING, by = "iso")
  demandEJ = demandEJ[,.(demand_EJ= sum(demand_EJ)), by = c("region", "year", "vehicle_type_plot", "aggr_mode")]

  p=ggplot()+
    geom_area(data = demandEJ[year > 2010], aes(x=year, y=demand_EJ, group = interaction(vehicle_type_plot,aggr_mode), fill = vehicle_type_plot), color = "black", position= position_stack())+
    facet_wrap(~region, nrow = 4)
    labs(x = "", y = "Final Energy demand [EJ]")+
    theme_minimal()+
    # scale_fill_manual("Vehicle Type",values = cols, breaks=legend_ord)+
    theme(axis.text.x = element_text(size = 8),
          axis.text.y = element_text(size=8),
          axis.title = element_text(size = 9),
          title = element_text(size = 9),
          legend.text = element_text(size = 9),
          legend.title = element_text(size =9),
          strip.text = element_text(size=9))


  return(p)
}

## Final Energy demand
demandEJplotf(demand_ej)
```

# LDVs final energy demand

```{r, echo=FALSE, warning=FALSE}
## demand EJ for LDV, divided by fuel type

demandEJLDVplotf <- function(demandEJ){
  demandEJ = demandEJ[subsector_L1 == "trn_pass_road_LDV_4W",]
  demandEJ <- demandEJ[, c("sector", "subsector_L3", "subsector_L2", "subsector_L1", "vehicle_type", "technology", "iso", "year", "demand_EJ")]

  demandEJ = merge(demandEJ, REMIND2ISO_MAPPING, by = "iso")
  demandEJ[technology == "Hybrid Liquids", technology := "Liquids"]
  demandEJ[technology == "FCEV", technology := "Hydrogen"]
  demandEJ[technology == "BEV", technology := "Electricity"]
  demandEJ = demandEJ[, .(demand_EJ = sum(demand_EJ)), by = c("region", "year", "technology")]

    p = ggplot()+
    geom_area(data = demandEJ[year > 2010], aes(x=year, y=demand_EJ, group = technology, fill = technology), color="black",position= position_stack())+
    labs(x = "", y = "Final energy demand for LDVs [EJ]")+
    facet_wrap(~region, nrow = 4)
    theme_minimal()+
    # scale_fill_manual("Vehicle Type",values = cols, breaks=legend_ord)+
    theme(axis.text.x = element_text(size = 7),
          axis.text.y = element_text(size=7),
          axis.title = element_text(size = 8),
          title = element_text(size = 8),
          legend.text = element_text(size = 8),
          legend.title = element_text(size = 8),
          strip.text = element_text(size=8))

    return(p)

}


demandEJLDVplotf(demand_ej)
```

# Energy services demand

```{r, echo=FALSE, warning=FALSE}
demandpkmplotf = function(demandpkm){
  ## REMIND-EDGE results
  demandpkm <- demandpkm[,c("sector","subsector_L3","subsector_L2",
                            "subsector_L1","vehicle_type","technology", "iso","year","demand_F")]
  demandpkm[,demand_F:=demand_F   ## in millionkm
            *1e-6                      ## in trillion km
            ]
  ## attribute aggregated mode and vehicle names for plotting purposes, and aggregate
  demandpkm[, aggr_mode := ifelse(subsector_L1 %in% c("Three-Wheeler", "trn_pass_road_LDV_4W"), "LDV", NA)]
  demandpkm[, aggr_mode := ifelse(sector %in% c("trn_freight", "trn_shipping_intl"), "Freight", aggr_mode)]
  demandpkm[, aggr_mode := ifelse(sector %in% c("trn_aviation_intl"), "Pass. non LDV", aggr_mode)]
  demandpkm[, aggr_mode := ifelse(subsector_L2 %in% c("trn_pass_road_bus", "HSR_tmp_subsector_L2", "Passenger Rail_tmp_subsector_L2", "Cycle_tmp_subsector_L2", "Walk_tmp_subsector_L2", "Domestic Aviation_tmp_subsector_L2", "Bus") | subsector_L1 %in% c("trn_pass_road_LDV_2W"), "Pass. non LDV", aggr_mode)]
  demandpkm[, veh := ifelse(grepl("Large|SUV|Midsize|Multipurpose Vehicle|Van|3W Rural", vehicle_type), "Large Cars", NA)]
  demandpkm[, veh := ifelse(grepl("Subcompact|Compact|Mini|Three-Wheeler_tmp_vehicletype", vehicle_type), "Small Cars", veh)]
  demandpkm[, veh := ifelse(grepl("Motorcycle|Moped|Scooter", vehicle_type), "Motorbikes", veh)]
  demandpkm[, veh := ifelse(grepl("bus|Bus", vehicle_type), "Bus", veh)]
  demandpkm[, veh := ifelse(grepl("Truck", vehicle_type) & vehicle_type != "Light Truck and SUV", "Truck", veh)]
  demandpkm[, veh := ifelse(subsector_L3 == "Domestic Aviation", "Domestic Aviation", veh)]
  demandpkm[, veh := ifelse(subsector_L3 == "International Aviation", "International Aviation", veh)]
  demandpkm[, veh := ifelse(grepl("Freight Rail", vehicle_type), "Freight Rail", veh)]
  demandpkm[, veh := ifelse(grepl("Passenger Rail|HSR", vehicle_type), "Passenger Rail", veh)]
  demandpkm[, veh := ifelse(grepl("Ship", vehicle_type), "Shipping", veh)]
  demandpkm[, veh := ifelse(grepl("Cycle|Walk", subsector_L3), "Non motorized", veh)]
  demandpkm[, veh := ifelse(is.na(veh), vehicle_type, veh)]

  demandpkm = demandpkm[,.(demand_F = sum(demand_F)), by = c("iso", "year", "aggr_mode", "veh")]
  setnames(demandpkm, old = "veh", new = "vehicle_type")

  demandpkm[, vehicle_type_plot := factor(vehicle_type, levels = c("LDV","Truck",
                                                                "Freight Rail",
                                                                "Motorbikes", "Small Cars", "Large Cars", "Van",
                                                                "Domestic Aviation", "International Aviation","Bus", "Passenger Rail",
                                                                "Freight", "Non motorized", "Shipping"))]


  demandpkm[, mode := ifelse(vehicle_type %in% c("Freight", "Freight Rail", "Truck", "Shipping"),"freight", "pass")]

  demandpkm = merge(demandpkm, REMIND2ISO_MAPPING, by = "iso")
  demandpkm = demandpkm[, .(demand_F = sum(demand_F)), by = c("region", "year", "vehicle_type_plot", "aggr_mode", "mode")]

    demandpkm = demandpkm[order(aggr_mode)]

  p = ggplot()+
    geom_area(data = demandpkm[mode =="pass"& year > 2010], aes(x=year, y=demand_F, group = interaction(vehicle_type_plot,aggr_mode), fill = vehicle_type_plot), color="black",position= position_stack())+
    labs(x = "", y = "Energy Services demand [trillion pkm]")+
    facet_wrap(~region, nrow = 4)
    theme_minimal()+
    # scale_fill_manual("Vehicle Type",values = cols, breaks=legend_ord)+
    theme(axis.text.x = element_text(size = 7),
          axis.text.y = element_text(size=7),
          axis.title = element_text(size = 8),
          title = element_text(size = 8),
          legend.text = element_text(size = 8),
          legend.title = element_text(size = 8),
          strip.text = element_text(size=8))


  return(p)

}

## energy services demand
demandpkmplotf(demand_km)
```

# CO2 intensity of new sales

```{r, echo=FALSE, warning=FALSE}

CO2km_intensity_newsalesplotf = function(annual_sales, mj_km_data, sharesVS1, shares_source_liquids){
  shares_source_liquids[, technology := ifelse(variable %in% c("FE|Transport|Liquids|Oil", "FE|Transport|Liquids|Coal"), "Oil", "Biodiesel")]
  shares_source_liquids = shares_source_liquids[,.(value = sum(value)), by = c("model","scenario","region", "period", "unit","technology")]
  shares_source_liquids = shares_source_liquids[region != "World"]
  shares_source_liquids[, region:=as.character(region)]
  shares_source_liquids[, year := period]
  shares_source_liquids[, period:=NULL]
  gdp <- getRMNDGDP(scenario = "SSP2", usecache = T)
  shares_source_liquids <- disaggregate_dt(shares_source_liquids, REMIND2ISO_MAPPING,
                                           valuecol="value",
                                           datacols=c("model","scenario", "unit","technology"),
                                           weights=gdp)

  shares_source_liquids[, shareliq := value/sum(value),by=c("iso", "year")]

  # ## CO2 content
  # CO2_petrol = 3.1 ## gCO2/gFUEL
  # CO2_biodiesel = 2.7 ## TODO this number is made up!
  # CO2_cng = 2.7 ## gCO2/gFUEL

  ## TODO of CO2 content of biodiesel is made up! gCO2/gFUEL
  emi_fuel = data.table(technology = c("Oil", "Biodiesel", "NG"), ei_gF_MJ = c(20, 20, 20), emi_cGO2_gF = c(3.1, 3.1, 2.7))

  emi_liquids = merge(shares_source_liquids, emi_fuel, all.x = TRUE, by = "technology")
  emi_liquids = emi_liquids[, .(ei_gF_MJ = sum(shareliq*ei_gF_MJ), emi_cGO2_gF = sum(shareliq*emi_cGO2_gF)), by = c("iso", "year")][, technology := "Liquids"]
  emi_NG = cbind(emi_fuel[technology == "NG"], unique(shares_source_liquids[,c("year", "iso")]))

  emi_fuel = rbind(emi_NG, emi_liquids)
  emi_fuel[, gCO2_MJ := ei_gF_MJ*emi_cGO2_gF]

  emi_fuel = merge(mj_km_data[subsector_L1 == "trn_pass_road_LDV_4W"], emi_fuel, all.x = TRUE, by = c("iso", "year", "technology"))
  emi_fuel[is.na(gCO2_MJ) & !technology %in% c("Liquids", "NG"), gCO2_MJ := 0]
  emi_fuel[, gCO2_km := MJ_km * gCO2_MJ]

  totalemi = merge(emi_fuel, annual_sales, all.y = TRUE, by = c("iso", "year", "technology", "vehicle_type", "subsector_L1"), all.x = TRUE)
  totalemi = totalemi[!is.na(share) & !is.na(gCO2_km)]
  totalemi[, gCO2_km_ave := gCO2_km*share]


  ##totalemi = merge(totalemi, demand_ej_plot)


  totalemi = totalemi[,.(gCO2_km_ave = sum(gCO2_km_ave)), by = c("year", "iso", "vehicle_type")]

  totalemi = merge(totalemi, sharesVS1, all.x = TRUE, by = c("iso", "year", "vehicle_type"))
  totalemi = totalemi[year %in% unique(sharesVS1$year)]

  totalemi = totalemi[,.(gCO2_km_ave = sum(gCO2_km_ave*share)), by = c("iso", "year", "subsector_L1")]


  totalemi = merge(totalemi, REMIND2ISO_MAPPING, by="iso")
  totalemi = merge(totalemi, gdp, all.x=TRUE, by = c("iso", "year"))
  totalemi[, share := weight/sum(weight), by = c("year", "region")]
  totalemi = totalemi[,.(gCO2_km_ave = sum(gCO2_km_ave*share)), by = c("year", "region")]

  p = ggplot()+
    geom_line(data = totalemi, aes(x = year, y = gCO2_km_ave))+
    labs(title = "gCO2/km average", y = "Average gCO2/km LDVs new additions")+
    facet_wrap(~region, nrow = 4)+
    theme_minimal()

  return(p)
}

shares_source_liquids = miffile[variable %in% c("FE|Transport|Liquids|Biomass", "FE|Transport|Liquids|Coal", "FE|Transport|Liquids|Oil"),]
CO2km_intensity_newsalesplotf(annual_sales, mj_km_data, sharesVS1 = shares$VS1_shares, shares_source_liquids)
```