library(tidyverse)
library(dplyr)
library(lubridate)
library(rdd)

readme_df_filepath <- "/mmfs1/gscratch/comdata/users/mjilg/govdoc-cr-data/final_data/metadata/final_0207_README_weekly_count_data.csv"
df = read.csv(readme_df_filepath, header = TRUE)

#filtered_df <- df |>
#  filter(!project_id == "letsencrypt_letsencrypt")

#output_filepath <-"/mmfs1/gscratch/comdata/users/mjilg/govdoc-cr-data/final_data/README_weekly_count_data.csv"
#write.csv(filtered_df, output_filepath, row.names = FALSE)

#EDA 
var(df$commit_count) #  207.7579
mean(df$commit_count) # 4.31725
median(df$commit_count) # 0
mean(df$age) # 5074.828
mean(df$age_at_commit) # 533.7488
median(df$age) # 4876
median(df$age_at_commit) # 0 days

# scale and log-transform
df$scaled_age <- scale(df$age)
df$scaled_age_at_commit <- scale(df$age_at_commit)
df$log1p_count <- log1p(df$commit_count)


#getting IK Bandwidth
get_optimal_bandwidth <- function(df){
  bw <- tryCatch({
    IKbandwidth(df$relative_week, df$log1p_count, cutpoint = 0, verbose = FALSE, kernel = "triangular")
  }, error = function(e) {
    NA
  })
}

mean_optimal_bandwidth <- df %>%
  group_by(project_id) %>%
  summarise(optimal_bandwidth = get_optimal_bandwidth(cur_data())) %>%
  drop_na(optimal_bandwidth) |>
  summarise(mean_optimal_bandwidth = mean(optimal_bandwidth))
  
#Mean Optimal Bandwidth: 5.471781