python 将hexo部署到阿里云OSS https://www.ficapy.com/2016/03/06/install_hexo_to_aliyun_oss/index.html

Posted 2021-05-09

#!/usr/bin/env python
# -*- coding: utf-8 -*-
# Author: Ficapy

import os
import requests
import base64
import hmac
import datetime
from os.path import join, exists, relpath
from mimetypes import MimeTypes
from hashlib import sha1, md5
from pickle import dump, loads
from cStringIO import StringIO, InputType, OutputType
from lxml import html

KEY             = '---------'
SECRET          = '---------'
BUCKET          = '---------'
url             = '---------'
ABSOLUTE_PATH   = '/Users/Ficapy/CodeSpace/Blog/public'

s = requests.Session()

headers = {
    'Date': datetime.datetime.utcnow().strftime('%a, %d %b %Y %H:%M:%S GMT'),
    'Content-Type': 'text/html',
    'HOST': url,
}


def signature(file_path):
    global headers
    path = join(ABSOLUTE_PATH, file_path)
    headers['Content-Type'] = MimeTypes().guess_type(path)[0] or 'text/html'
    sign = base64.b64encode(hmac.new(SECRET,
                                     'PUT' + "\n"
                                     + '' + "\n"
                                     + headers['Content-Type'] + "\n"
                                     + headers['Date'] + "\n"
                                     + '/' + BUCKET + '/' + file_path, sha1).digest()).strip()
    headers.update({'Authorization': 'OSS ' + KEY + ':' + sign})


def read_in_chunks(file_object, blocksize=1024):
    file_object = file_object if isinstance(file_object, (InputType, OutputType)) else open(
        join(ABSOLUTE_PATH, file_object))
    file_object.seek(0)
    while 1:
        data = file_object.read(blocksize)
        if not data:
            break
        yield data


def upload(file_path, file_object):
    signature(file_path)
    s.put('http://' + url + '/' + file_path, headers=headers, data=read_in_chunks(file_object)).raise_for_status()
    print('upload done: ' + file_path)


# 首先请求oss_fuck_md5文件,没有则全部上传  有则和生成的md5对比,本地文件不存在或变动则上传
# 所有文件上传成功后最后上传.oss_fuck_md5文件
def local_md5():
    file_md5_mp = {}
    for root, dirnames, files in os.walk(ABSOLUTE_PATH):
        for file in files:
            fullpath = join(root, file)
            rel = relpath(fullpath, ABSOLUTE_PATH)
            file_md5_mp[rel] = md5(open(fullpath).read()).hexdigest()
    return file_md5_mp


def remote_md5():
    a = s.get('http://' + url + '/oss_fuck_md5')
    if a.status_code != 200:
        return {}
    else:
        return loads(a.content)


def path_process(file_path):
    file_path = join(ABSOLUTE_PATH, file_path)
    tree = html.parse(file_path).getroot()
    hrefs = tree.xpath('//*[@href]')

    for i, href in enumerate(hrefs):
        rewrite = join(href.attrib['href'].split()[0], 'index.html')
        if exists(join(ABSOLUTE_PATH, rewrite[1:])):
            hrefs[i].attrib['href'] = rewrite
    f = StringIO()
    f.write(html.tostring(tree))
    return f


def main():
    remote = remote_md5()
    locals = local_md5()
    _locals = locals.copy()
    for local in _locals.keys():
        if _locals[local] != remote.get(local, None):
            continue
        else:
            locals.pop(local)

    for path in locals.keys():
        if path.endswith('html'):
            upload(path, path_process(path))
        else:
            upload(path, path)

    f = StringIO()
    dump(_locals, f)
    upload('oss_fuck_md5', f)


if __name__ == '__main__':
    main()