library(tidyverse)
library(dplyr)
library(lubridate)
c1_event_date <- as.Date("2013-07-01")
c2_event_date <- as.Date("2013-08-28")
c3_event_date <- as.Date("2015-07-02")
relative_week <- function(date, ref_date) {
  as.integer(as.numeric(difftime(date, ref_date, units = "days")) %/% 7)
}

core_csv <-"~/121325_work/121225_vd_data/extension_VisualEditor_2000-01-01_to_2016-12-31.csv"
core_df <- read.csv(core_csv, header = TRUE) 
known_affil_emails <- c("krinkle@fastmail.com", "roan.kattouw@gmail.com", 
                        "trevorparscal@gmail.com", "krinklemail@gmail.com", "moriel@gmail.com")
active_names<- c("Timo Tijhof", "Krinkle", "Roan Kattouw", "Catrope",
                 "Trevor Parscal", "Ed Sanders")
core_df <- core_df |>
  mutate(commit_date = ymd_hms(commit_date)) |>
  mutate(isAuthorWMF = case_when(
    author_name %in% active_names ~ "FIVE",
    grepl("@wikimedia\\.org", author_email, ignore.case = TRUE) ~ "TRUE",
    grepl("@wikimedia\\.de", author_email, ignore.case = TRUE) ~ "TRUE",
    grepl("l10n-bot@translatewiki\\.net", author_email, ignore.case = TRUE) ~ "localization",
    grepl("@gerrit\\.wikimedia\\.org", author_email, ignore.case = TRUE) ~ "Gerrit",
    TRUE ~ "FALSE"
  )) |>
  mutate(isVE = case_when(
    grepl("VisualEditor", message, ignore.case = TRUE) ~ TRUE,
    grepl(" VE ", message, ignore.case = TRUE) ~ TRUE,
    TRUE ~ FALSE
  ))

c1_core_weekly <- core_df |>
  mutate(week_index = relative_week(commit_date, c1_event_date)) |>
  group_by(week_index, isAuthorWMF)|>
  summarise(count = n(), .groups = 'drop')|>
  filter(week_index >= -9 & week_index < -4) |>
  mutate(source = 'c1')
c1summary <- c1_core_weekly |>
  group_by(isAuthorWMF)|>
  summarize(total = sum(count))


c2_core_weekly <- core_df |>
  mutate(week_index = relative_week(commit_date, c2_event_date)) |>
  group_by(week_index, isAuthorWMF)|>
  summarise(count = n(), .groups = 'drop')|>
  filter(week_index >= -104 & week_index <= 13) |>
  mutate(source = 'c2')

c3_core_weekly <- core_df |>
  mutate(week_index = relative_week(commit_date, c3_event_date)) |>
  group_by(week_index, isAuthorWMF)|>
  summarise(count = n(), .groups = 'drop')|>
  filter(week_index >= -83 & week_index <= 13) |>
  mutate(source = 'c3')
#collate and save
core_weekly <- rbind(c1_core_weekly, c2_core_weekly, c3_core_weekly)

c1summary <- c1_core_weekly |>
  group_by(isAuthorWMF)|>
  summarize(total = sum(count))

c2summary <- c2_core_weekly |>
  group_by(isAuthorWMF)|>
  summarize(total = sum(count))

c3summary <- c3_core_weekly |>
  group_by(isAuthorWMF)|>
  summarize(total = sum(count))