liluo / py-oauth2 Goto Github PK

Hi there, it's me again.
Since the last time I asked about download large file streamly via OAuth2 , I eventually found there's possibly a bug in your code. I noticed an unnormal RAM usage which would even use up all RAM on old computers. This happens during the last few seconds of downloading a file (50+ MB).

And I found this is somethins related to this line:

If changed to:

        options = {'parse': 'text'}
        options.update(opts)
        if options['parse'] in ('text', 'query', 'json', ):
            self.body = response.text

would solve this problem, and also lower the RAM usage during small size file downloading.

https://github.com/liluo/py-oauth2/blob/master/pyoauth2/client.py#L39

里面处理了 response，但是没有保留这个 response 对象，导致外层无法获取 response.body 对象，有时候服务器报了异常，比如 code error / invalid user id，就没法判断哪里出问题了，
我建议将这个 response 保存到 self.response。

Hi. I was using this little tool for OAuth 2 authentication, and I recently found something confusing.

Think about downloading a large file via OAuth 2. 'Cause the file is very large, I don't want the client to read it all into memory at once. Instead, I thought I could use the iter_content() method of requests.models.Response to write the file to disk block by block, since your tool uses requests. Yet I've read your code, and did not see the stream parameter being passed to the client.request() method, and I think this might be a bug.

As far as I know, requests does require this parameter to be passed to Request to work as the way I described above.

Could you please explain some of your code? If this is rather a feature or something designed than a bug, I'd be glad to hear from you about the right way to do my work.

和
https://github.com/simplegeo/python-oauth2
的包名冲突了，改个名吧。

client = Client(CLIENT_ID, CLIENT_SECRET,

                site='https://www.googleapis.com/oauth2/v1',

                authorize_url='https://accounts.google.com/o/oauth2/auth',

                token_url='https://accounts.google.com/o/oauth2/token')



print '-' * 80

authorize_url = client.auth_code.authorize_url(redirect_uri=REDIRECT_URL, 

                                               scope=SCOPE)





app = Flask(__name__)







@app.route("/")

def hello():

    print "hello world"

    return redirect(authorize_url)



@app.route("/oauth2callback")

def oauth2callback():

How do i get access to user info in oauth2callback in this example?

I got the error as:
TypeError: exceptions must derive from BaseException
for the line of:
pyoauth2/libs/access_token.py", line 39, in refresh
in python 3.4. I have not test on 3.3 yet.

When access_token returns in format 'query', then the access_token.expires_in is str, not a int.

This situation caused access_token.expires_at is not set.

see https://github.com/liluo/py-oauth2/blob/master/oauth2/libs/access_token.py#L13

首先感谢作者提供了这么简洁的工具。已经在使用了，总体感觉很好用，但是用的时候还是发现了两处bug。

第一个地方是在libs/access_token.py 82行左右，原文是：
opts['body'][self.opts['param_name']] = token
我想作者的原意应该是：
opts['body'][self.opts['param_name']] = self.token

第二个地方是在libs/request.py ，因为url参数和配置项都是通过 **opt 传递，但是如果接口参数和配置项重名的话就报错了，我觉得应该有一个params的配置项，这个配置项传递url的参数。奇怪的是我在libs/access_token.py 的__set_token 这个函数中看到了params配置项，但是在libs/request.py中没有看到对其进行处理。