#* index_dates_vs_age:
#*   attr:
#*     fillcolor: '2'
#*   desc: Summary visualization patient index dates vs their age.
#*   ext: R
#*   inputs:
#*   - person_all_facts
#* 

library(dplyr)
library(ggplot2)
library(tidyr)
library(hexbin)

index_dates_vs_age <- function( person_all_facts) {
    epoch_stats_by_person <- SparkR::collect(SparkR::select(person_all_facts, c("pre_first_date", "cci_score_quan", "age_at_index", "cohort", "index_date"))) %>%
        mutate(cci_score_quan = as.integer(cci_score_quan))
    print(str(epoch_stats_by_person))
    
    #epoch_stats_by_person <- epoch_stats_by_person %>% sample_frac(0.2)
    
    p <- ggplot(epoch_stats_by_person) +
        stat_summary_hex(aes(x = index_date, y = age_at_index, z = cci_score_quan), fun = length, bins = 60) +
        #scale_x_continuous(trans = "log10") + 
        #scale_y_continuous(trans = "log10") +
        #scale_fill_continuous(trans = "log10") +
        #annotation_logticks(sides = "bl") +
        facet_wrap(~ cohort) 
        #coord_equal()

    plot(p)

    # p2 <- ggplot(epoch_stats_by_person) +
    #     stat_summary_2d(aes(x = pre_epoch_effective_days, y = post_epoch_effective_days, z = increase_effective_days), fun = length, bins = 60) +
    #     scale_x_continuous(trans = "log10") + 
    #     scale_y_continuous(trans = "log10") +
    #     scale_fill_continuous(trans = "log10") +
    #     annotation_logticks(sides = "bl")
    # plot(p2)

    return(NULL)
}