def add_company_column_for_issues(org, repo):
users = c.get_issue_authors_with_company(org, repo)
issues = c.get_issues(org, repo)
issues_with_company = merge_issues_with_company_column(issues, users)
issues_with_company.to_csv(org + "_" + repo + "_" + c.issue_file_suffix +
"with_employer",
sep='\t')
def calculate_issue_processing_time(org, repo):
issues = c.get_issues(org, repo)
for based_on_devstats_data in True, False:
issues_w_processing_time = p.calculate_issue_time_difference(
org, repo, issues, "created_at", "closed_at",
based_on_devstats_data)
issues_w_processing_time.rename(
columns={'time_difference': 'processing_time'}, inplace=True)
issues_w_processing_time.to_csv(
org + "_" + repo + "_" + c.issue_file_suffix +
"_with_processing_time_based_on_devstats_" +
str(based_on_devstats_data),
sep='\t')
def print_logistic_regression_for_pr_acceptance_rate(
org, repo, based_on_devstats_data=False):
users = p.get_users(org, repo, based_on_devstats_data)
issues = c.get_issues(org, repo)
pulls = c.get_pulls(org, repo)
pulls = p.merge_pulls_with_issue_priority_and_kind(pulls, issues)
pulls = _merge_pulls_with_company_column(pulls, users,
based_on_devstats_data)
pulls = p.add_dummy_column_for_pr_merge_state(pulls)
pulls = _add_controlling_variables(pulls)
pulls = _prepare_independent_company_variable(pulls,
based_on_devstats_data)
result = sm.logit(formula=_ols_formula("pr_is_merged",
based_on_devstats_data),
data=pulls).fit()
_print_company_representation_in_data(pulls)
_print_and_save_result(result)
def calculate_issue_reponse_time(org, repo):
issues_w_comments = c.get_issues_with_comments(org, repo)
if issues_w_comments.empty:
issues = c.get_issues(org, repo)
issue_comments = c.get_issue_comments(org, repo)
issue_comments = p.extract_first_comment_per_issue(issue_comments)
issues_w_comments = p.merge_issues_with_issue_comments(
issues, issue_comments)
for based_on_devstats_data in True, False:
issues_w_response_time_df = p.calculate_issue_time_difference(
org, repo, issues_w_comments, "created_at", "commented_at",
based_on_devstats_data)
issues_w_response_time_df.rename(
columns={'time_difference': 'response_time'}, inplace=True)
issues_w_response_time_df.to_csv(
org + "_" + repo + "_" + c.issue_file_suffix +
"_with_response_time_based_on_devstats_" +
str(based_on_devstats_data),
sep='\t')
def calculate_avg_issue_response_time_by_company(org, repo):
issues = c.get_issues(org, repo)
issue_comments = c.get_issue_comments(org, repo)
issues_w_comments = p.merge_issues_with_issue_comments(
issues, issue_comments)
companies = c.get_companies(org, repo)
for employer in companies.keys():
companies[employer]["response_time"] = timedelta(0)
companies[employer]["issue_count"] = 0
time_format = "%Y-%m-%d %H:%M:%S"
for _, issue in issues_w_comments.iterrows():
employer = p.get_employer(
issue.user_login_x, org,
repo) # TODO make more generic / change merge behavior
open_issue = type(issue.created_at_y) is float and math.isnan(
issue.created_at_y)
if employer is None or open_issue:
continue
companies[employer]["response_time"] = companies[employer][
"response_time"] + p.determine_processing_time(
issue.created_at_y, issue.created_at_x,
time_format) # TODO make more generic / change merge behavior
companies[employer]["issue_count"] += 1
for employer in companies.keys():
companies[employer]["avg_response_time"] = companies[employer][
"response_time"].total_seconds() / companies[employer][
"issue_count"] if companies[employer]["issue_count"] else 0
print(
str(employer) + " - avg_response_time: " +
str(companies[employer]["avg_response_time"]))
print(
str(employer) + " - issue_count: " +
str(companies[employer]["issue_count"]))
return companies
def calculate_avg_issue_processing_time_by_company(org, repo):
# note that PRs are included here as they are a special type of an issue
time_format = "%Y-%m-%d %H:%M:%S"
companies = c.get_companies(org, repo)
for employer in companies.keys():
companies[employer]["processing_time"] = timedelta(0)
companies[employer]["issue_count"] = 0
issues = c.get_issues(org, repo)
print("Start iterating over issues...")
for _, issue in issues.iterrows():
employer = p.get_employer(issue.user_login, org, repo)
still_open = type(issue.closed_at) is float and math.isnan(
issue.closed_at)
if employer is None or still_open:
continue
print(issue.title)
companies[employer]["processing_time"] = companies[employer][
"processing_time"] + p.determine_processing_time(
issue.created_at, issue.closed_at, time_format)
companies[employer]["issue_count"] += 1
for employer in companies.keys():
companies[employer]["avg_processing_time"] = companies[employer][
"processing_time"].total_seconds() / companies[employer][
"issue_count"] if companies[employer]["issue_count"] else 0
print(
str(employer) + " - avg_processing_time: " +
str(companies[employer]["avg_processing_time"]))
print(
str(employer) + " - issue_count: " +
str(companies[employer]["issue_count"]))
return companies
def determine_company_share_of_issues_based_on_devstats_data(org, repo):
issues = c.get_issues(org, repo)
users = get_formatted_devstats_user()
issues_with_company = merge_issues_with_company_column(issues, users)
return Counter(issues_with_company.company.values)