library(dplyr)
library(lubridate)
library(rdd)
library(stringr)

readme_count_data_filepath <- "/mmfs1/gscratch/comdata/users/mjilg/govdoc-cr-data/final_data/README_weekly_count_data.csv"
readme_count_df = read.csv(readme_count_data_filepath, header = TRUE) 

readme_topic_dist_filepath <- "text_analysis/020125_README_file_topic_distributions.csv"
readme_topics_df = read.csv(readme_topic_dist_filepath, header = TRUE) 

window_num <- 5
readme_count_df <- readme_count_df |>
  filter(week_index >= (- window_num) & week_index <= (window_num)) |>
  mutate(scaled_age = scale(age)) |>
  mutate(scaled_age_at_commit = scale(age_at_commit))|>
  mutate(log1p_count = log1p(commit_count))

summed_data <- readme_count_df |>
  filter(before_after == 1) |>
  group_by(project_id) |>
  summarise_at(vars(commit_count), list(summed_count=sum))

readme_topics_df <- readme_topics_df |>
  mutate(project_id = sapply(str_split(filename, "_hullabaloo_"), `[`, 1)) |> 
  mutate(project_id = ifelse(filename=="jaraco_keyrings.alt_hullabaloo_README.rst", "jaraco_keyrings.alt", project_id)) |>
  mutate(project_id = ifelse(filename=="_vcr_vcr_README.md", "vcr_vcr", project_id)) 
  
filtered_topics <- readme_topics_df |>
  filter(project_id %in% summed_data$project_id)

merged_df <- inner_join(summed_data, readme_topics_df, by="project_id")

multiple_mappings <- merged_df %>%
  group_by(project_id) %>%
  filter(n() > 1) %>%
  ungroup()

merged_df$logged_commits <- log1p(merged_df$summed_count)