library(tidyverse)

df_path <-"/mmfs1/home/mjilg/git/mw-lifecycle-analysis/p2/071425_master_discussion_data.csv"
combined_df <- read.csv(df_path, header = TRUE) |> 
  filter(comment_type == "task_description")


combined_count_df <- combined_df |>
  group_by(AuthorWMFAffil, week_index, source) %>%
  summarise(unique_count = n_distinct(id),
            closed_relevance_count = sum(closed_relevance == TRUE, na.rm = TRUE),
            closed_relevance_resolved_count = sum(closed_relevance == TRUE & status == "resolved", na.rm = TRUE),
            closed_relevance_resolved_prop = closed_relevance_resolved_count / unique_count,
            closed_relevance_prop = closed_relevance_count / unique_count,
            .groups = "drop")

library(scales)
library(ggplot2)


affiliationColors <-
  setNames( c('#5da2d8', '#c7756a')
            ,c("False", "True"))

new_tasks <- combined_count_df |>
  ggplot(aes(x=week_index, 
             y=closed_relevance_resolved_prop, 
             fill=factor(AuthorWMFAffil))) +
  geom_col(position='dodge2') +
  labs(x = "Relative Week", y = "% of New Tasks Resolved by Deployment's End", fill="Task Author") +
  geom_vline(data = combined_df |> filter(source == "c1"), 
             aes(xintercept = -29), 
             linetype = "dotted", color = "black", linewidth = 0.5) +
  geom_vline(data = combined_df |> filter(source == "c1"), 
             aes(xintercept = -9), 
             linetype = "dotted", color = "black", linewidth = 0.5) +
  geom_vline(data = combined_df |> filter(source == "c1"), 
             aes(xintercept = -4), 
             linetype = "3313", color = "black", linewidth = 0.5) +
  geom_vline(data = combined_df |> filter(source == "c2"), 
             aes(xintercept = -99), 
             linetype = "dotted", color = "black", linewidth = 0.5) +
  geom_vline(data = combined_df |> filter(source == "c2"), 
             aes(xintercept = -4),   
             linetype = "3313", color = "black", linewidth = 0.5) +
  geom_vline(data = combined_df |> filter(source == "c3"), 
             aes(xintercept = -97), 
             linetype = "dotted", color = "black", linewidth = 0.5) +
  geom_vline(data = combined_df |> filter(source == "c3"), 
             aes(xintercept = -3), 
             linetype = "3313", color = "black", linewidth = 0.5) +
  geom_vline(xintercept = 0, linetype = "dashed", color = "black", linewidth = 0.5) + # Add vertical line at week 0
  scale_fill_manual(values = affiliationColors,
                    labels = c("False" = "Unaffiliated",
                               "True" = "WMF-affiliated")) +
  ggtitle("% of Tasks Resolved Within Three Months of Wide-release") +
  theme_bw()+
  theme(
    legend.position = "top",
    plot.title = element_text(size = 18, face = "bold"), # Increase title font size
    axis.title.x = element_text(size = 16),             # Increase x-axis title font size
    axis.title.y = element_text(size = 16),             # Increase y-axis title font size
    axis.text.x = element_text(size = 16),              # Increase x-axis text font size
    axis.text.y = element_text(size = 16),              # Increase y-axis text font size
    legend.text = element_text(size = 16),              # Increase legend text font size
    legend.title = element_text(size = 16),
    strip.text = element_text(size = 14)# Increase legend title font size
  ) +
  facet_wrap(~source, nrow = 3, scales = "free_y", labeller = labeller(source = c(
    "c1" = "VisualEditor (2013)",
    "c2" = "HTTPS-as-default (2013)",
    "c3" = "HTTP-deprecation (2015)"
  )))
new_tasks
#geom_text(data = data.frame(source = "c1", relative_week = -42.5, lengthened_commit_count = 0.95), 
#aes(x = relative_week, y = lengthened_commit_count, label = "Opt-In Testing Deployment"), 
#inherit.aes = FALSE, color = "black", size = 4) +
#geom_text(data = data.frame(source = "c2", relative_week = 7, lengthened_commit_count = 0.95), 
#         aes(x = relative_week, y = lengthened_commit_count, label = "Wide Deployment"), 
#          inherit.aes = FALSE, color = "black", size = 4) +
#  geom_text(data = data.frame(source = "c3", relative_week = -16, lengthened_commit_count = 0.95), 
#            aes(x = relative_week, y = lengthened_commit_count, label = "Wide Deployment Announcement"), 
#            inherit.aes = FALSE, color = "black", size = 4) +

ggsave(filename = "073125-fossy-tasks-resolved.png", plot = new_tasks, width = 15, height = 9, dpi = 800)