mgaughan/24_deb_pkg_gov
1
0
Fork 0
Code Issues Pull Requests Packages Projects Releases Wiki Activity

big data collection run

Browse Source
This commit is contained in:
Matthew Gaughan 2023-12-11 15:42:49 -06:00
parent 29a6ef7074
commit 27d336594f
2 changed files with 6 additions and 3 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

5
expanded_data_collection.py
View File

@ -46,7 +46,10 @@ def main():
perceval_data = pt.main(project_dict["upstream_vcs_link"], early_cutoff) perceval_data = pt.main(project_dict["upstream_vcs_link"], early_cutoff)
if perceval_data == {}: if perceval_data == {}:
continue continue
project_dict["age_of_project"], project_dict["contributors"], project_dict["collaborators"] = perceval_data['age_of_project'], perceval_data['contributors'], perceval_data['collaborators'] project_rosters = {}
project_dict["age_of_project"], project_dict["contributors"], project_dict["collaborators"], project_rosters['contributors_list'], project_rosters['collaborators_list'] = perceval_data['age_of_project'], perceval_data['contributors'], perceval_data['collaborators'], perceval_data['contributors_list'], perceval_data['collaborators_list']
with open("/data/users/mgaughan/kkex_roster_data_121123/" + "rosters_" + project_dict["project_name"] + '.json', 'w') as roster_path
json.dump(project_rosters, roster_path)
if "github" in project_dict["upstream_vcs_link"]: if "github" in project_dict["upstream_vcs_link"]:
project_dict['milestone_count'] = gha.main(project_dict["upstream_vcs_link"], early_cutoff) project_dict['milestone_count'] = gha.main(project_dict["upstream_vcs_link"], early_cutoff)
with open('/data/users/mgaughan/kkex_comment_data_121123/' + 'gh_comments_' + project_dict["project_name"] + '.json', 'w') as data_path: with open('/data/users/mgaughan/kkex_comment_data_121123/' + 'gh_comments_' + project_dict["project_name"] + '.json', 'w') as data_path:

4
perceval_tasks.py
View File

@ -11,7 +11,7 @@ def main(vcs_path, begin_date):
perceval_info['list_of_commits'] = get_perceval_log(vcs_path, begin_date) perceval_info['list_of_commits'] = get_perceval_log(vcs_path, begin_date)
if len(perceval_info['list_of_commits']) > 0: if len(perceval_info['list_of_commits']) > 0:
perceval_info['age_of_project'] = get_repo_age(perceval_info['list_of_commits']) perceval_info['age_of_project'] = get_repo_age(perceval_info['list_of_commits'])
perceval_info['contributors'], perceval_info['collaborators'] = get_all_actors(perceval_info['list_of_commits']) perceval_info['contributors'], perceval_info['collaborators'], perceval_info['contributors_list'], perceval_info['collaborators_list'] = get_all_actors(perceval_info['list_of_commits'])
del perceval_info['list_of_commits'] del perceval_info['list_of_commits']
print(perceval_info) print(perceval_info)
return perceval_info return perceval_info
@ -63,7 +63,7 @@ def get_all_actors(all_commits):
for committer in committers: for committer in committers:
if committer in authors: if committer in authors:
authors.remove(committer) authors.remove(committer)
return len(authors), len(committers) return len(authors), len(committers), authors, committers
if __name__ == "__main__": if __name__ == "__main__":
main(" https://github.com/pali/0xFFFF".strip(),dt.datetime(2008,2, 8)) main(" https://github.com/pali/0xFFFF".strip(),dt.datetime(2008,2, 8))