使用Python统计git仓库中热点函数

使用Python3统计git仓库中C++代码改动频繁的函数

# -- coding: utf-8 --
 
import _thread
 
import git
import re
import sys
import threading
import time
 
 
# 全局变量定义
thread_num = 1  # 使用的CPU线程数
commit_num = 3000  # 需要遍历的commit数量
repo = git.Repo.init("E:/ClionProject/serving")  # 仓库路径，根据自己的实际填写
folder_path = "."  # 需要扫描的子文件夹路径
diff_regex = r"@@.*@@\s((\w+)\s+)+[\*,&]*\s*(\w*:*\w+)\s*\("  # 修改函数名匹配正则表达式，根据自己的需要修改
max_modified_filenum = 30  # 允许的单次提交修改的最大文件数，排除分支合并的commit
 
# 正则匹配
commit_regex = r"commit\s(\w{40})"  # commit匹配正则表达式，当前简单匹配40个hash码
 
# 全局线程互斥变量
fun_dict = {}
lock = threading.Lock()
global final_thread_count
final_thread_count = 0
 
 
# 获取正则匹配结果
def get_regex_match(src_str, regex_str):
    partten = re.compile(regex_str)
    match = partten.findall(src_str)
    return match
 
 
# 获取小于特定修改文件数的commit，支持多线程
def get_commit_lessthan(match, thread_id):
    print("id is:", thread_id)
    fo = open("fo" + str(thread_id) + ".txt", "w")
    ok_list = []
    count = 0
 
    # 筛选修改文件数符合要求的commit
    for commit in match:
        filelist_str = repo.git.show(commit, stat=True)
        name_patt = re.compile(commit_regex)
        name_match = name_patt.findall(filelist_str)
        if len(name_match) <= max_modified_filenum:
            ok_list.append(commit)
        else:
            print("toMax----",len(name_match))
 
    # 获取每个commit的具体修改内容
    for commit in ok_list:
        if commit is "":
            continue
        diff_str = get_commit_content(commit, "*")
        # 从内容中提取修改的函数
        change_fun_list = get_change_funcs(diff_str)
        try:
            fo.write(diff_str)
        except BaseException:
            print("happy...except...")
        count += 1
 
        lock.acquire()
        try:
            # 添加到全局数组
            add_funlist2dict(change_fun_list)
            print("Thd" + str(thread_id) + "\t" + commit + "_" + str(count) + "/" + str(len(ok_list)))
            if count == len(ok_list):
                print(repo.git.show(commit, stat=True))
        except BaseException:
            print("happy...except....")
        finally:
            lock.release()
 
    fo.close()
    global final_thread_count
    final_thread_count += 1
 
 
# 按指定数量拆分数组
def chunks(l, n):
    for i in range(0, len(l), n):
        yield l[i:i + n]
 
 
# 获取commit列表
def get_commit_list():
    log = repo.git.log(folder_path)
    all_commit = get_regex_match(log, commit_regex)
    commit_groups = list(chunks(all_commit, int(commit_num / thread_num)))
    print(commit_groups[-6:-1])
 
    # 开始多个线程扫描修改commit，修改函数名
    try:
        for index in range(0, min(thread_num, len(commit_groups))):
            _thread.start_new_thread(get_commit_lessthan, (commit_groups[index], index,))
    except:
        print("start thread fail!")
    return
 
 
# 获取COMMIT中文件的修改内容
def get_commit_content(commit_id, file_name):
    return repo.git.show(commit_id, file_name)
 
 
# 提取差异中的函数
def get_change_funcs(diff_log):
    return get_regex_match(diff_log, diff_regex)
 
 
# 将函数加入全局dict
def add_funlist2dict(list):
    for fun in list:
        if fun in fun_dict:
            fun_dict[fun] = fun_dict[fun] + 1
        else:
            fun_dict[fun] = 1
 
 
# dict按照value值排序
def sort_funcdict(dict):
    items = dict.items()
    backitems = [[v[1], v[0]] for v in items]
    backitems.sort(reverse=True)
    return [[v[1], v[0]] for v in backitems]
 
 
if __name__ == '__main__':
    # 开始处理
    get_commit_list()
    # 等待所有线程处理完毕
    while (final_thread_count < thread_num):
        time.sleep(1)
 
    # 输出结果
    print("out_put result:")
    resfile = open("result.txt", "w")
    sorted_list = sort_funcdict(fun_dict)
    for func_name, times in sorted_list:
        res_str = " times:" + str(times) + "\t" + func_name[1].strip() + " " + func_name[2].strip() + "\n"
        resfile.write(res_str)
    resfile.close()

参考：
https://www.liushuideng.com/efficiency/110
https://www.dazhuanlan.com/zm4015/topics/1298219