# -*- coding: utf-8 -*-
# @Time    : 2022/4/7 21:24
# @Author  : Kaicheng

import os
import urllib
import github



GH = github.Github("your_access_token")
ROOT_DIR = "D:/OSS_Projects"
repo_list = ["jruby/jruby", "netty/netty", "spring-projects/spring-framework", "bumptech/glide", "jfree/jfreechart", "apache/lucene-solr"]


def file_download(local_path, url):
    if not os.path.exists(os.path.dirname(local_path)):
        os.makedirs(os.path.dirname(local_path))
    if not os.path.exists(local_path):
        urllib.request.urlretrieve(url, local_path)


for repo_str in repo_list:
    repo = GH.get_repo(repo_str)
    pr_list = repo.get_pulls(state='open', sort='created')

    for index in range(pr_list.totalCount):
        pr = pr_list[index]

        # select the appropriate PRs
        java_count = 0
        for f in pr.get_files():
            if f.filename.endswith(".java"):
                java_count += 1
        if java_count < 2 or java_count > 100:
            continue

        print(pr.title + "  " + str(pr.number))
        first_commit = pr.get_commits()[0]
        parent_commit = first_commit.parents[0]
        pre_repo_url = f"https://github.com/{repo_str}/raw/{parent_commit.sha}"
        print("parent commit repo resource url: " + pre_repo_url)
        print("parent commit repo front-end url: " + pre_repo_url.replace("/raw/", "/tree/"))

        for f in pr.get_files():
            if not f.filename.endswith(".java"):
                continue

            print(f.filename)
            if f.status == "removed":
                local_old_path = f"{ROOT_DIR}/{repo_str}/{str(pr.number)}/old/{f.filename}"
                file_download(local_old_path, f.raw_url)
            elif f.status == "added":
                local_new_path = f"{ROOT_DIR}/{repo_str}/{str(pr.number)}/new/{f.filename}"
                file_download(local_new_path, f.raw_url)
            else:
                # download new version
                local_new_path = f"{ROOT_DIR}/{repo_str}/{str(pr.number)}/new/{f.filename}"
                file_download(local_new_path, f.raw_url)
                # download old version
                try:
                    old_raw_url = f"{pre_repo_url}/{f.filename.replace('/', '%2F')}"
                    local_old_path = f"{ROOT_DIR}/{repo_str}/{str(pr.number)}/old/{f.filename}"
                    file_download(local_old_path, old_raw_url)
                except:
                    print()
                    print(f"404: {pre_repo_url}/{f.filename.replace('/', '%2F')}")
                    continue
        print("--------------------------------------------------")
    print("===================================================")