sAbakumoff

SELECT
  id,
  FIRST(name) AS repo_name,
  FIRST(copies) AS num_of_copies
FROM
  [githubdataqueries:NpmStat.package_json_duplicates]
GROUP BY
  id
ORDER BY
  num_of_copies DESC

sAbakumoff

{
  "name": "angular2",
  "version": "<%= packageJson.version %>",
  "description": "Angular 2 - a web framework for modern web apps",
  "homepage": "<%= packageJson.homepage %>",
  "bugs": "<%= packageJson.bugs %>",
  "contributors": <%= JSON.stringify(packageJson.contributors) %>,
  "license": "<%= packageJson.license %>",
  "repository": <%= JSON.stringify(packageJson.repository) %>,
  "devDependencies": <%= JSON.stringify(packageJson.defaultDevDependencies) %>,

sAbakumoff

var toString = {}.toString;

module.exports = Array.isArray || function (arr) {
  return toString.call(arr) == '[object Array]';
};

sAbakumoff

library(dplyr)
library(stringr)
react_angular_commits<-read.csv("react_angular_commits.csv")  %>% 
  mutate(message=str_replace_all(message, "\n", " "), 
         message_len=str_length(message), 
         name=str_sub(repo_name, start=regexpr('/', repo_name) + 1))
boxplot(message_len ~ name, react_angular_commits, outline=FALSE, 
        horizontal=TRUE, las=2, 
        col=c("#d20013", "#53d2fa"), 
        main="Length of the commits messages", par(mar=c(5,6,4,6)))         

sAbakumoff

library(tidytext)
data("stop_words")
tidy_commits <- react_angular_commits %>% 
  select(name, message) %>% 
  unnest_tokens(word, message) %>% 
  anti_join(stop_words)

sAbakumoff

tidy_commits %>% 
      filter(name=="angular.js") %>% 
      count(word, sort = TRUE) %>%
      head(15) %>%
      mutate(word = reorder(word, n)) %>%
      ggplot(aes(word, n)) +
      geom_bar(stat = "identity", fill=I("#d20013")) +
      labs(y="Occurrences", 
      title="Most Common Words in Angular Commit Messages") +
      coord_flip()

sAbakumoff

angular_commits_type_scope <- react_angular_commits %>% 
  filter(name=="angular.js", 
         str_detect(message, "^(feat|fix|docs|style|refactor|test|chore)\\s*\\(")) %>%
  mutate(type=trimws(str_sub(message, start=0, end=regexpr( "\\(", message ) - 1)), 
         scope=trimws(str_sub(message, 
                       start=regexpr( "\\(", message ) + 1, 
                       end=regexpr( "\\)", message ) - 1))) %>%
  select(type, scope)

top_scope<-angular_commits_type_scope %>% 

sAbakumoff

tidy_commits %>% 
  filter(name=="react") %>% 
  count(word, sort = TRUE) %>%
  head(15) %>%
  mutate(word = reorder(word, n)) %>%
  ggplot(aes(word, n)) +
  geom_bar(stat = "identity", fill=I("#53d2fa")) +
  labs(y="Occurrences", title="Most Common Words in React Commit Messages") +
  coord_flip()

sAbakumoff

bing <- sentiments %>% 
filter(lexicon == "bing") %>% 
select(-score)

angular_word_counts<-tidy_commits %>% 
  filter(name=="angular.js") %>% 
  inner_join(bing) %>% 
  count(word, sentiment, sort = TRUE) %>% 
  ungroup()

sAbakumoff

#step 1 : remove rows with merge, pull, request words as they useless for analysis
custom_stop_words<-data.frame("word"=c("merge", "pull", "request"))
tidy_commits<-tidy_commits %>% anti_join(custom_stop_words)

#step 2: calculate tf-idf and sort the rows by its value in descending order  
commit_words<-tidy_commits %>% 
  count(name, word, sort=TRUE) %>% 
  ungroup() %>% 
  bind_tf_idf(word, name, n) %>%
  arrange(desc(tf_idf)) %>%