knitr::opts_chunk$set( collapse = TRUE, comment = "#>", fig.path = "README-" )
Package is work in progress! If you encounter errors / problems, please file an issue or make a PR.
This package parses a git repository history to collect comprehensive information about the activity in the repo. The parsed data is made available to the user in a tabular format. The package can also generate reports based on the parse data. You can install the development version from GitHub.
remotes::install_github("lorenzwalthert/gitsum")
There are two main functions for parsing the history, both return tabular data:
parse_log_simple()
is a relatively fast parser and returns a tibble with
one commit per row. There is no file-specific information.parse_log_detailed()
outputs a nested tibble and for each commit, the names
of the amended files, number of lines changed ect. available. This function
is slower.report_git()
creates a html, pdf, or word report with the parsed log data
according to a template. Templates can be created by the user or a template
from the gitsum
package can be used.
Let's see the package in action.
library("gitsum") library("tidyverse") library("forcats")
We can obtain a parsed log like this:
remove_gitsum()
init_gitsum() tbl <- parse_log_detailed() %>% select(short_hash, short_message, total_files_changed, nested) tbl
Since we used parse_log_detailed()
, there is detailed file-specific
information available for every commit:
tbl$nested[[3]]
Since the data has such a high resolution, various graphs, tables etc. can be produced from it to provide insights into the git history.
Since the output of git_log_detailed()
is a nested tibble, you can work on it
as you work on any other tibble.
Let us first have a look at who comitted to this repository:
log <- parse_log_detailed() log %>% group_by(author_name) %>% summarize(n = n())
We can also investigate how the number of lines of each file in the R directory evolved. For that, we probaly want to view files with changed names as one file. Also, we probably don't want to see boring plots for files that got changed only a few times. Let's focus on files that were changed in at least five commits.
lines <- log %>% unnest_log() %>% set_changed_file_to_latest_name() %>% add_line_history() r_files <- grep("^R/", lines$changed_file, value = TRUE) to_plot <- lines %>% filter(changed_file %in% r_files) %>% add_n_times_changed_file() %>% filter(n_times_changed_file >= 10) ggplot(to_plot, aes(x = date, y = current_lines)) + geom_step() + scale_y_continuous(name = "Number of Lines", limits = c(0, NA)) + facet_wrap(~changed_file, scales = "free_y")
Next, we want to see which files were contained in most commits:
log %>% unnest_log() %>% mutate(changed_file = fct_lump(fct_infreq(changed_file), n = 10)) %>% filter(changed_file != "Other") %>% ggplot(aes(x = changed_file)) + geom_bar() + coord_flip() + theme_minimal()
We can also easily get a visual overview of the number of insertions & deletions in commits over time:
commit.dat <- data.frame( edits = rep(c("Insertions", "Deletions"), each = nrow(log)), commit = rep(1:nrow(log), 2), count = c(log$total_insertions, -log$total_deletions)) ggplot(commit.dat, aes(x = commit, y = count, fill = edits)) + geom_bar(stat = "identity", position = "identity") + theme_minimal()
Or the number of commits broken down by day of the week:
log %>% mutate(weekday = factor(weekday, c("Mon", "Tue", "Wed", "Thu", "Fri", "Sat", "Sun"))) %>% ggplot(aes(x = weekday)) + geom_bar() + theme_minimal()
Add the following code to your website.
For more information on customizing the embed code, read Embedding Snippets.