updating with search for relevant commits

121325_work/relevant_commit_attribution.R

@@ -1,6 +1,7 @@
 library(tidyverse)
 library(dplyr)
 library(lubridate)
+library(stringr)
 c1_event_date <- as.Date("2013-07-01")
 c2_event_date <- as.Date("2013-08-28")
 c3_event_date <- as.Date("2015-07-02")
@@ -30,6 +31,7 @@ active_names<- c("Timo Tijhof", "Krinkle", "Roan Kattouw", "Catrope",
                  "Trevor Parscal", "Ed Sanders", "Moriel Schottlender", "Gabriel Wicke", "C. Scott Ananian")
 core_df <- core_df |>
   mutate(commit_date = ymd_hms(commit_date)) |>
+  mutate(is_http = ifelse(contains_http_but_not_url(message), TRUE, FALSE))|>
   mutate(isAuthorWMF = case_when(
     author_name %in% active_names ~ "NAMES",
     grepl("@wikimedia\\.org", author_email, ignore.case = TRUE) ~ "TRUE",
@@ -54,8 +56,28 @@ c1_relevant <- core_df |>
   mutate(week_index = relative_week(commit_date, c1_event_date)) |>
   filter(week_index >= -33 & week_index <= 13) |>
   filter(author_email != "jenkins-bot@gerrit.wikimedia.org") |>
-  filter(isVE == TRUE)
+  filter(isVE == TRUE)|>
+  mutate(case = "c1") |>
+  select(authored_date, commit_date, author_name, author_email, message)
 
+c2_relevant <- core_df |>  
+  mutate(week_index = relative_week(commit_date, c2_event_date)) |>
+  filter(week_index >= -104 & week_index <= 13) |>
+  filter(author_email != "jenkins-bot@gerrit.wikimedia.org") |>
+  filter(is_http) |>
+  mutate(case = "c1") |>
+  select(authored_date, commit_date, author_name, author_email, message)
+
+c3_relevant <- core_df |>  
+  mutate(week_index = relative_week(commit_date, c3_event_date)) |>
+  filter(week_index >= -83 & week_index <= 13) |>
+  filter(author_email != "jenkins-bot@gerrit.wikimedia.org") |>
+  filter(is_http) |>
+  mutate(case = "c1") |>
+  select(authored_date, commit_date, author_name, author_email, message)
+
+possibly_relevant <- rbind(c1_relevant, c2_relevant, c3_relevant)
+write.csv(possibly_relevant, "012926_possibly_relevant_sample.csv", row.names = FALSE)
 
 total_share <- core_df |>
   filter(commit_date >= as.Date('2012-11-11') & commit_date <= as.Date("2015-10-02"))|>

analysis_data/scratch.R

@@ -3,6 +3,8 @@ library(dplyr)
 library(stringr)
 main_csv <-"~/analysis_data/121625_unified.csv"
 main_df <- read.csv(main_csv, header = TRUE) 
+dsl_csv <-"~/dsl/121625_DSL_frame.csv"
+dsl_df <- read.csv(dsl_csv, header = TRUE) 
 
 #01-10-26 look for affil rosters
 affils_ <- main_df |>
@@ -87,7 +89,7 @@ tasks_flagged <- main_df %>%
   filter(comment_type == "task_description") |>
   left_join(first_task, by = c("source", "AuthorPHID")) %>%
   mutate(is_first_time_author = week_index == first_task_week)
-
+# Results for RQ2
 summary_df <- tasks_flagged %>%
   mutate(
     period = case_when(
@@ -96,7 +98,7 @@ summary_df <- tasks_flagged %>%
       TRUE ~ NA
     )
   ) |>
-  group_by(period, source) %>%
+  group_by(period, source, isAuthorWMF) %>%
   summarize(
     total_tasks = n(),
     first_time_tasks = sum(is_first_time_author),