24_deb_pkg_gov/pr_data_get.py

import csv
from perceval.backends.core.git import Git
import os
import datetime as dt
import time
import shutil 
import pandas as pd
import dateutil
from tqdm import tqdm
import math

key = os.environ.get('KKEXKEY')

early_cutoff = dt.datetime(2008,2, 8)
temp_dir = "/data/users/mgaughan/tmp"

'''
- rate of change, rate of PRs/day
'''
def file_get_pr(upstream_vcs_link, me_read):
    # if we're looking at readmes me_read is true and if not, if we're looking at contributing files, it's false
    #this is the window of days on either side of the event that we're looking at
    window = 182
    #print(upstream_vcs_link.split('/')[4])
    project_dict = {}
    project_dict['upstream_vcs_link'] = upstream_vcs_link
    full_temp_path = temp_dir + upstream_vcs_link.split('/')[4] + ".git"
    repo = Git(uri=upstream_vcs_link, gitpath=full_temp_path)
    try:
        commits = repo.fetch()
    except:
        print("perceval issue")
        return
    has_readme = False
    has_contributing = False
    merge_pre_rm, merge_post_rm, merge_pre_cont, merge_post_cont = 0, 0, 0, 0
    #list of tuples which has date and whether it was a merge
    commit_list = []
    for commit in commits:
        if "Merge" in commit['data'].keys():
            commit_list.append([commit['data']['CommitDate'], True])
            if has_contributing:
                merge_post_cont += 1
            else:
                merge_pre_cont += 1
        else: 
            commit_list.append([commit['data']['CommitDate'], False])
        files = commit['data']['files']
        #print(commit['data']['CommitDate'])
        #print(type(dateutil.parser.parse(commit['data']['CommitDate'])))
        for file in files:
            if "CONTRIBUTING" in file['file'] and has_contributing == False:
                has_contributing = True
                first_date_contributing = dateutil.parser.parse(commit['data']['CommitDate'])
            if "README" in file['file'] and has_readme == False:
                has_readme = True
                first_date_readme = dateutil.parser.parse(commit['data']['CommitDate'])
    shutil.rmtree(full_temp_path, ignore_errors=True)
    if me_read:
        project_dict['first_readme'] = first_date_readme
        before_read = pr_count(first_date_readme+ dt.timedelta(days=-window, hours=0), first_date_readme, commit_list)
        project_dict['b6w_prs_read'] = before_read[0]
        project_dict['b6w_mrg_read'] = before_read[1]
        after_read = pr_count(first_date_readme, first_date_readme + dt.timedelta(days=window, hours=0), commit_list)
        project_dict['a6w_prs_read'] = after_read[0]
        project_dict['a6w_mrg_read'] = after_read[1]
    else:
        project_dict['first_contributing'] = first_date_contributing
        before_cont = pr_count(first_date_contributing + dt.timedelta(days=-window, hours=0), first_date_contributing, commit_list)
        project_dict['b6w_prs_cont'] = before_cont[0]
        project_dict['b6w_mrg_cont'] = before_cont[1]
        after_cont = pr_count(first_date_contributing, first_date_contributing + dt.timedelta(days=window, hours=0), commit_list)
        project_dict['a6w_prs_cont'] = after_cont[0]
        project_dict['a6w_mrg_cont'] = after_cont[1]
    print(project_dict)
    return project_dict


#TODO: pr_count should return an array of values for weekly/6mo
def pr_count(start, end, commits):
    count = 0
    merge_count = 0
    by_week = [0] * 27
    by_week_merge =[0] * 27
    current_week = 0
    for commit in tqdm(commits):
        if dateutil.parser.parse(commit[0]) > start:
            if math.floor((dateutil.parser.parse(commit[0]) - start).days / 7) <= 26:
                by_week[math.floor((dateutil.parser.parse(commit[0]) - start).days / 7)] += 1
                if commit[1]:
                    by_week_merge[math.floor((dateutil.parser.parse(commit[0]) - start).days / 7)] += 1
        if dateutil.parser.parse(commit[0]) > end:
            print(len(by_week))
            return [by_week, by_week_merge]

#TODO: need to do this for all files in the dataset of readme or contributing
def for_files():
    csv_path = "final_data/kk_final_readme_roster.csv"
    count = 0 
    with open(csv_path, 'r') as file:
        csv_reader = csv.DictReader(file)
        with open('kk_test_031024_pr_did.csv', "w") as writing_file:
            keys = ['upstream_vcs_link', "first_readme", "b6w_prs_read", "b6w_mrg_read", "a6w_prs_read", "a6w_mrg_read"]
            dict_writer = csv.DictWriter(writing_file, keys)
            dict_writer.writeheader()
            #training wheels on right now
            for row in csv_reader:
                count += 1
                print(row['upstream_vcs_link'])
                dict_row = file_get_pr(row['upstream_vcs_link'].strip(), True)
                dict_writer.writerow(dict_row)
    

if __name__ == "__main__":
    for_files()
    #file_get_pr("https://github.com/tqdm/tqdm", False)
    #file_get_pr("https://github.com/GameServerManagers/LinuxGSM")