Overview

Project version control commit histories are the authoritative open history of an open source community. That is not to say that contributions outside of commits are less important or interesting, but version control commit histories are clearly documented points in time associated with a project’s artifacts.

Traditional analysis of commit histories tend to focus on simple summary statistics like numbers of commits or lines of code often to create a leaderboard ranking of authors. This report takes history analysis a step further by looking at author trends from a macro level, rather than at an indvidual level. The goal of this analysis is to discover what the commit history can tell us about a project’s overall activity level, trends, and growth potential.

Commits and Authors

The traditional commit and author count paradigm tells us little about the nature of an organization’s involvement in a project. This notebook demonstrates the process of extracting email addresses from commits in the git log for the Tensorflow project and shows the distribution of domains of those email addresses to illustrate the challenges of identification using this method. In response to the challenges of accurate identification, this notebook poses the suggestion that the use of time intervals instead of frequency totals is sufficient to indicate an organization engagement in a project through commits. Note that project engagement is a much larger topic and commits alone are insufficient to determine an organization’s overall project engagement.

This notebook lives in Github and is the first in a series: https://github.com/countering-bean-counting/commit-log-chronology/tensorflow-commit-log

Setup Instructions

To use this notebook, you first need to clone the Tensorflow repository (ideally in the “data” folder within the same location as this notebook).

# Clone the repo into the data directory for this project
paste0("git clone ", params$git_url, " ", getwd(),'/', params$git_path, "/", params$git_repo)
## [1] "git clone git@github.com:tensorflow/tensorflow.git /home/auggy/dev/R/countering-bean-counting/git-commit-log-engagement/tensorflow-commit-log/data/tensorflow"
# Get the current SHA and update the params
gitlog_sha <- system(paste0("cd ", params$git_path, "/", params$git_repo, 
                                   "; git rev-parse HEAD"),  intern = TRUE)

This notebook sets the SHA used for the analysis as a parameter to ensure reproducibility. If you want to run this against the latest changes, update the SHA in the parameters to the latest one in your local repository.

Get Git Commit Log

git_log_cmd <- paste0('cd ', params$git_path,"/", params$git_repo, 
                     '; git log ', gitlog_sha, 
                     #' --no-merges ',
                     ' --date=short --pretty=tformat:"%ad|%an|%ae|%cd|%cn|%ce|%h" > ', "../", params$gitlog_out)
system(git_log_cmd)

git_log_cmd
## [1] "cd data/tensorflow; git log ce41d2f95e1e5883f1808030c94fd9aaa57d9f10 --date=short --pretty=tformat:\"%ad|%an|%ae|%cd|%cn|%ce|%h\" > ../gitlog_tensorflow_tensorflow.txt"
gitlog_raw <- read.csv(paste0(params$git_path, "/", params$gitlog_out),
                       header = FALSE, sep = "|", quote="",
                       col.names=c("author_date", "author_name", "author_email", 
                                 "committer_date", "committer_name", "committer_email", 
                                 "sha"))

Create Time Intervals

# fix names and emails to be all lowercase
gitlog_commits_dates <- gitlog_raw %>% 
  mutate(
         author_date=as.Date(author_date, tz="UTC"),
         committer_date=as.Date(committer_date, tz="UTC")
    ) %>% 
  mutate(
    commit_date=ymd(committer_date),
    commit_year=floor_date(commit_date, "year"),
    commit_halfyear=floor_date(commit_date, "halfyear"),
    commit_quarter=floor_date(commit_date, "quarter"),
    commit_month=floor_date(commit_date, "month"),
    commit_bimonth=floor_date(commit_date, "bimonth"),
    commit_week=floor_date(commit_date, "week")
  )

Extract Email Domains

# these will be labeled as "Personal"
email_providers <- c("126",
                     "163",
                     "github",
                     "gmail",
                     "googlemail",
                     "hotmail",
                     "live",
                     "me",
                     "naver",
                     "outlook",
                     "qq",
                     "yahoo")

googler_domains <- c("google", 
                     "tensorflow", 
                     "petewarden", 
                     "vomjom", 
                     "babuschk", 
                     "naml")
gitlog_commits <- gitlog_commits_dates %>% 
  mutate(
         author_name=str_to_lower(author_name),
         author_email=str_to_lower(author_email),
         committer_name=str_to_lower(committer_name),
         committer_email=str_to_lower(committer_email)
    ) %>% 
  separate(author_email, c("author_username", "author_host"), sep="@", remove=FALSE) %>%
  separate(committer_email, c("committer_username", "committer_host"), sep="@", remove=FALSE) %>%
  mutate(
    author_domain=suffix_extract(author_host)$domain,
    author_suffix=suffix_extract(author_host)$suffix,
    author_is_edu=str_detect(author_suffix, "edu")|str_detect(author_suffix, "ac."),
    committer_domain=suffix_extract(committer_host)$domain,
    committer_suffix=suffix_extract(committer_host)$suffix,
    committer_is_edu=str_detect(committer_suffix, "edu")|str_detect(committer_suffix, "ac."),
    author_hosted_email=author_domain %in% email_providers,
    committer_hosted_email=committer_domain != "github" & committer_domain %in% email_providers,
    author_is_google = author_domain %in% googler_domains,
    committer_is_google = committer_domain %in% googler_domains | committer_domain == "github",
    author_domain_type = 
           ifelse(author_is_google, "Google",
                  ifelse(author_hosted_email, "Personal",
                         ifelse(author_is_edu, "Edu", "Other")
                                )
                  ),
    author_domain_type=ifelse(is.na(author_domain), "Local", author_domain_type),
    committer_domain_type =  
           ifelse(committer_is_google, "Google",
                  ifelse(committer_hosted_email, "Personal",
                         ifelse(committer_is_edu, "Edu", "Other")
                         )
                  ),
    committer_domain_type=ifelse(is.na(committer_domain), "Local", committer_domain_type)
    )
## Warning: Expected 2 pieces. Missing pieces filled with `NA` in 2 rows
## [8420, 8494].
## Warning: Expected 2 pieces. Missing pieces filled with `NA` in 1 rows
## [8494].
gitlog_commits <- gitlog_commits %>%
  filter(commit_month < '2018-10-01')
gitlog_commits <- gitlog_commits %>%
  group_by(author_domain) %>%
  mutate(author_first_commit_age=as.numeric(ymd("2017-10-01") - min(commit_date))) %>%
  ungroup() %>%
  group_by(committer_domain) %>%
  mutate(committer_first_commit_age=as.numeric(ymd("2018-10-01") - min(commit_date)))
saveRDS(gitlog_commits, paste0("data/", params$repo, "_gitlog_commits.Rds"))
gh_committers_by_email <- gitlog_commits %>%
  rename(name=committer_name, email=committer_email) %>%
  arrange(desc(commit_date)) %>%
  group_by(email, name) %>%
  summarise(last_commit=max(commit_date)) %>%
  arrange(desc(last_commit)) 

gh_committers_join1 <- gh_committers_by_email %>%
  inner_join(gh_committers_by_email %>% select(name, email) %>% rename(name2=name), "email") %>%
  unique()

gh_committers_join <- gh_committers_join1 %>%
  inner_join(gh_committers_join1 %>% select(name, email) %>% rename(email2=email), "name") %>%
  unique()

rm(gh_committers_join1)
# group commits by email address and name
gh_authors_by_email <- gitlog_commits %>%
  rename(name=author_name, email=author_email) %>%
  arrange(desc(commit_date)) %>%
  group_by(email, name) %>%
  summarise(num_commits = n(), 
            last_commit=max(commit_date)) %>%
  arrange(desc(last_commit))

# join on name to show emails tied to the same names
gh_authors_join1 <- gh_authors_by_email %>%
  inner_join(gh_authors_by_email %>% select(name, email) %>% rename(name2=name), "email") %>%
  unique()

# join on email to show names tied to the same emails
gh_authors_join <- gh_authors_join1 %>%
  inner_join(gh_authors_join1 %>% select(name, email) %>% rename(email2=email), "name") %>%
  unique()

rm(gh_authors_join1)
gh_emails <- bind_rows(gh_authors_join %>% select(email, email2), gh_committers_join %>% select(email, email2))
gh_emails <- gh_emails %>% unique()

gh_emails %>% filter(str_detect(email, "keveman"))
## # A tibble: 4 x 2
## # Groups:   email [2]
##   email              email2            
##   <chr>              <chr>             
## 1 keveman@google.com keveman@google.com
## 2 keveman@google.com keveman@gmail.com 
## 3 keveman@gmail.com  keveman@google.com
## 4 keveman@gmail.com  keveman@gmail.com
# this might need to be directed in the future based on commit dates
gh_emails_graph_big <- graph_from_data_frame(gh_emails,
                                             directed=FALSE,
                                             vertices=unique(gh_emails$email2))

E(gh_emails_graph_big)$weight <- 1
gh_emails_graph <- simplify(gh_emails_graph_big, 
                            edge.attr.comb=list(
                              weight = "sum", 
                              transaction_amount = "sum", 
                              function(x)length(x))
                            )

# identify clusters
gh_emails_networks <- clusters(as.undirected(gh_emails_graph))
V(gh_emails_graph)$network <- gh_emails_networks$membership

# extract vertices
gh_emails_nodes_vert <- get.data.frame(gh_emails_graph, what="vertices")

# create nodes with fields used by Visnetwork for plotting
gh_emails_nodes <- data.frame(id = gh_emails_nodes_vert$name,
                              title = gh_emails_nodes_vert$name, 
                              group = gh_emails_nodes_vert$network)
gh_emails_nodes <- gh_emails_nodes[order(gh_emails_nodes$id, decreasing = F),]

# extract edges
gh_emails_edges <- get.data.frame(gh_emails_graph, what="edges")[1:2]

# remove data structures we no longer need
rm(gh_committers_emails_graph, gh_emails_graph, gh_emails_networks, gh_emails_nodes_pre)

gh_emails_nodes_is_google <- gh_emails_nodes %>%
  select(id, group) %>%
  rename(email=id) %>%
  group_by(group) %>%
  mutate(is_googler=any(str_detect(email, "google")))

# join by committer email address with git log data to get the clusters
gitlog_networks <- gitlog_commits %>% 
  ungroup() %>%
  inner_join(gh_emails_nodes_is_google %>% 
               select(email, group, is_googler) %>% 
               rename(committer_group=group, committer_is_googler=is_googler), 
             by=c("committer_email"="email"))

# join by author
gitlog_networks <- gitlog_networks %>% 
  ungroup() %>%
  inner_join(gh_emails_nodes_is_google %>% 
               select(email, group, is_googler) %>% 
               rename(author_group=group, author_is_googler=is_googler), 
             by=c("author_email"="email"))

paste("identified", max(gitlog_networks$committer_group),"unique committers from", n_distinct(gh_emails$email),"emails")
## [1] "identified 1778 unique committers from 1956 emails"
paste("identified", max(gitlog_networks$author_group),"unique authors from", n_distinct(gh_emails$email),"emails")
## [1] "identified 1760 unique authors from 1956 emails"
# This will show false because we haven't identified bsed on groups yet
gitlog_networks %>% 
  ungroup() %>% 
  filter(author_username=="keveman") %>% 
  select(author_name, author_host, author_is_google, author_is_googler, author_group) %>% 
  unique()
## # A tibble: 2 x 5
##   author_name   author_host author_is_google author_is_googl… author_group
##   <chr>         <chr>       <lgl>            <lgl>                   <dbl>
## 1 manjunath ku… google.com  TRUE             TRUE                      978
## 2 manjunath ku… gmail.com   FALSE            TRUE                      978
# just plot Google people, otherwise this is too big
gh_emails_nodes_google <- gh_emails_nodes %>% 
  group_by(group) %>%
  mutate(is_google=any(str_detect(id, "google"))) %>%
  filter(is_google) %>%
  select(-is_google)

saveRDS(gh_emails_nodes_google, "data/tf_google_email_nodes.Rds")

# filter out non-google emails
gh_emails_edges_google <- gh_emails_edges %>% 
  group_by(from) %>%
  mutate(is_google=any(str_detect(to, "google") | str_detect(from, "google"))) %>%
  filter(is_google) %>%
  select(-is_google)
  
saveRDS(gh_emails_edges_google, "data/tf_google_email_edges.Rds")

gh_emails_graph <- visNetwork(gh_emails_nodes_google, gh_emails_edges_google) %>%
  visOptions(highlightNearest = TRUE, nodesIdSelection = TRUE)

gh_emails_graph
# Export for embedding into the blog entry
# manually open HTML and save as in web browser

graph_file <- paste0("gh_", params$git_repo, "_emails_graph.html")
visSave(gh_emails_graph, graph_file, 
        selfcontained = TRUE, background = "white")

# visSave doesn't use relative paths
system(paste0("mv ", graph_file, " data/gh_emails_graph" ))

Committers

# is any in network google

gitlog_commits_is_google <- gitlog_networks %>%
  mutate(
    committer_domain_type=ifelse(committer_is_googler, "Google", committer_domain_type),
    author_domain_type=ifelse(author_is_googler, "Google", author_domain_type)
    )
committers_month_is_google <- gitlog_commits_is_google %>%
  ungroup() %>%
  group_by(committer_group) %>%
  arrange(desc(commit_month)) %>% # arrange by most recent commit
  mutate(
    committer = first(committer_name),
    committer_email = first(committer_email),
    committer_type=ifelse(committer_email=="noreply@github.com", "Github", 
                               ifelse(any(committer_domain_type=="Google"), "Google", "Not Google")),
    gmail=any(committer_domain=="gmail")
    
  )

committers_month_is_google_summary <- committers_month_is_google %>%
  ungroup() %>%
  group_by(commit_month) %>%
  mutate(total_committers=n_distinct(committer_group)) %>% # total number unique committers each month
  group_by(commit_month, committer_type) %>%
  summarise(
    num_committers=n_distinct(committer_group),
    pct_committers=round(num_committers/first(total_committers),4)
      ) %>%
  mutate(commit_year=year(commit_month))

saveRDS(committers_month_is_google_summary, "data/tf_committer_diversity.Rds")

ggplot(committers_month_is_google_summary, 
       aes(x=commit_month, y=pct_committers)) +
  geom_bar(aes(fill=committer_type), position="dodge", stat="identity") +
  theme_few() +
  labs(x="Month", y="Committers", title="Tensorflow Commmiters") +
  guides(fill=guide_legend(title="Google?")) +
  scale_y_continuous(labels = percent, breaks=pretty_breaks()) +
  facet_wrap(~ commit_year, scales="free_x")

committers_month_is_gmail_summary <- committers_month_is_google %>%
  ungroup() %>%
  group_by(commit_month, committer_type) %>%
  mutate(total_committers=n_distinct(committer_group)) %>% # total number unique committers each month
  group_by(commit_month, committer_type, gmail) %>%
  summarise(
    num_committers=n_distinct(committer_group),
    pct_committers=round(num_committers/first(total_committers),4)      
    ) %>%
  mutate(commit_year=year(commit_month))

ggplot(committers_month_is_gmail_summary %>% filter(gmail), 
       aes(x=commit_month, y=pct_committers)) +
  geom_bar(aes(fill=committer_type), position="dodge", stat="identity") +
  theme_few() +
  labs(x="Month", y="Committers w/in Type", title="Tensorflow Commmitters using Gmail address") +
  guides(fill=guide_legend(title="Committer Type")) +
  scale_y_continuous(labels = percent, breaks=pretty_breaks()) +
  facet_wrap(~ commit_year, scales="free_x")

commits_month_is_google <- gitlog_commits_is_google %>%
  ungroup() %>%
  group_by(commit_month) %>%
  mutate(total_commits=n()) %>% # total number of commits made each month
  ungroup() %>%
  group_by(committer_group) %>%
  arrange(desc(commit_month)) %>% # arrange by most recent commit
  mutate(
    gmail=any(committer_domain=="gmail"),
    committer = first(committer_name),
    committer_email = first(committer_email),
    committer_type=ifelse(committer_email=="noreply@github.com", "Github", 
                               ifelse(any(committer_domain_type=="Google"), "Google", 
                                      "Not Google")) # are any of the committers' email addresses Google?
  ) %>% 
  group_by(commit_month, committer_group) %>%
  mutate(num_commits=n()) %>% # number of commits per user
  ungroup() %>%
  select(commit_month, committer_group, committer_name, committer_email, committer_type, num_commits, total_commits, gmail) %>%
  unique()

commits_month_is_google_summary <- commits_month_is_google %>%
  group_by(commit_month, committer_type) %>%
  summarise(num_commits=sum(num_commits),
            pct_commits=round(num_commits/first(total_commits), 4)) %>%
  mutate(commit_year=year(commit_month))

saveRDS(commits_month_is_google_summary, "data/tf_commit_diversity.Rds")

ggplot(commits_month_is_google_summary, 
       aes(x=commit_month, y=pct_commits)) +
  geom_bar(aes(fill=committer_type), position="dodge", stat="identity") +
  theme_few() +
  labs(x="Month", y="Commits", title="Tensorflow Commits") +
  guides(fill=guide_legend(title="Google?")) +
  scale_y_continuous(labels = percent, breaks=pretty_breaks()) +
  facet_wrap(~ commit_year, scales="free_x")

commits_month_is_gmail_summary <- commits_month_is_google %>%
  group_by(commit_month, committer_type) %>%
  mutate(total_commits=sum(num_commits)) %>%
  group_by(commit_month, committer_type, gmail) %>%
  summarise(num_commits=sum(num_commits),
            total_commits = first(total_commits),
            pct_commits=round(num_commits/first(total_commits), 4)) %>%
  mutate(commit_year=year(commit_month))

ggplot(commits_month_is_gmail_summary %>% filter(gmail), 
       aes(x=commit_month, y=pct_commits)) +
  geom_bar(aes(fill=committer_type), position="dodge", stat="identity") +
  theme_few() +
  labs(x="Month", y="Commits w/in Committer Type", title="Tensorflow Commits from Committers w/ Gmail address") +
  guides(fill=guide_legend(title="Committer Type")) +
  scale_y_continuous(labels = percent, breaks=pretty_breaks()) +
  facet_wrap(~ commit_year, scales="free_x")