library(TSA)
library(caTools)
library(dplyr)
library(forecast)
library(ggplot2)
library(reticulate)
library(tidyr)
theme_set(theme_bw())
use_virtualenv("../venv/")

# Load, transform data
p <- import("pandas")
tempdf <- p$read_pickle("../data/2016-18-weather.pkl") %>% 
    select(-record_no, -station) %>% 
    mutate(temp_date = as.Date(temp_date))
str(tempdf)

# Get a data frame of all the dates/times we want
wdts <- seq.POSIXt(min(tempdf$temp_timestamp), max(tempdf$temp_timestamp), by = "30 mins")
hhdt <- as.difftime("30", format = "%M")
wantdf <- data.frame(temp_timestamp = wdts, temp_date = as.Date(wdts)) %>% arrange(temp_timestamp)
str(wantdf)

# Recursively impute NAs from previous non-NA value; if first value is NA replaced by median but this shouldn't happen
# If all values are NA will hang, but this shouldn't happen either
fulltemp <- left_join(wantdf, tempdf, by = c("temp_timestamp", "temp_date"))
str(fulltemp)
fulltemp <- fill(fulltemp, tmax_c:rhmean)
sum(is.na(fulltemp$tmin_c))
fulltemp$runmin <- runmin(fulltemp$tmin_c, 48, endrule = "min", align = "right")
fulltemp$runmax <- runmax(fulltemp$tmax_c, 48, endrule = "max", align = "right")
str(fulltemp)

# Temperature plots
tplot <- ggplot(fulltemp, aes(x = temp_timestamp, y = tmin_c)) + geom_line() +
    geom_line(aes(y = tmax_c), color = "magenta") +
    geom_line(aes(y = runmin), color = "blue") +
    geom_line(aes(y = runmax), color = "red")

# tplot

# tplot + coord_cartesian(xlim = c(as.POSIXct("2016-12-01"), as.POSIXct("2017-03-01")))

# Create a harmonic (sine wave) model for minimum temperature
yharm <- harmonic(ts(1:nrow(fulltemp), frequency = floor(365.25 * 48)), 2) %>% as.data.frame()
#dharm <- harmonic(ts(1:nrow(fulltemp), frequency = floor(48)), 1)
hmod <- lm(fulltemp$runmin ~ ., data = yharm)
summary(hmod)

hmdf <- data.frame(x = fulltemp$temp_timestamp, y = fulltemp$runmin, f = fitted(hmod), r = resid(hmod))
tmplot <- ggplot(hmdf, aes(x = x, y = y)) + geom_line(aes(y = f), color = "blue", size = 2) + geom_point() +
    geom_point(aes(y = r), color = "darkgreen")

# tmplot

# tmplot + coord_cartesian(xlim = c(as.POSIXct("2017-05-01", tz = "UTC"), as.POSIXct("2017-06-01", tz = "UTC")))



maxhmod <- lm(fulltemp$runmax ~ ., data = yharm)
summary(maxhmod)
mhmdf <- data.frame(x = fulltemp$temp_timestamp, y = fulltemp$runmax, f = fitted(maxhmod), r = resid(maxhmod))

mtmplot <- ggplot(mhmdf, aes(x = x, y = y)) + geom_line(aes(y = f), color = "blue", size = 2) + geom_point() +
    geom_point(aes(y = r), color = "darkgreen")

# mtmplot

# mtmplot + coord_cartesian(xlim = c(as.POSIXct("2017-05-01", tz = "UTC"), as.POSIXct("2017-06-01", tz = "UTC")))

hmdf.comb <- left_join(hmdf, mhmdf, by = "x", suffix = c(".min", ".max"))

ctmplot <- ggplot(hmdf.comb, aes(x = x, y = f.min)) + geom_line(color = "blue", size = 2) +
    geom_line(aes(y = f.max), color = "red", size = 2) +
    geom_point(aes(y = y.max), color = "magenta") +
    geom_point(aes(y = y.min), color = "lightblue")

# ctmplot

# ctmplot + coord_cartesian(xlim = c(as.POSIXct("2017-05-01", tz = "UTC"), as.POSIXct("2017-06-01", tz = "UTC")))

write.csv(hmdf.comb, "../data/weatherharm.csv", row.names = FALSE)

saveRDS(hmod, "../models/weatherMinharmonic.rds")
saveRDS(maxhmod, "../models/weatherMaxharmonic.rds")