Just as the title describes

In the Usage section, both Client constructors pass a dummy account_type value of "by defect common". This is potentially confusing to people who don't realize you meant "default" instead of "defect", such as in #5.

Since "common" fits the vast majority of use cases and is already the default value, I think the account_type parameter should just be removed from the examples.

If you feel like it should remain, please add a note about what the valid parameter values are or at least correct the typo.

How can i use this with proxy settings?

Hi,

I'm getting this error after updating to 1.1.1. Not sure if this is a local issue or a bug? Noted some changes where made in nearby code in this PR: #29

Error in Microsoft login hook error Traceback (most recent call last):
  File "/workspaces/.../.../integrations/azure/aad/auth_hook.py", line 43, in handle_login
    user = client.users.get_me()
  File "/usr/src/.venv/lib/python3.8/site-packages/microsoftgraph/decorators.py", line 11, in helper
    return func(*args, **kwargs)
  File "/usr/src/.venv/lib/python3.8/site-packages/microsoftgraph/users.py", line 32, in get_me
    return self._client._get(self._client.base_url + "me", params=params)
  File "/usr/src/.venv/lib/python3.8/site-packages/microsoftgraph/client.py", line 227, in _get
    return self._paginate_response(response)
  File "/usr/src/.venv/lib/python3.8/site-packages/microsoftgraph/client.py", line 215, in _paginate_response
    data = list(response.data["value"])
KeyError: 'value'

If you run a request specifying, for example, the $top parameter, it will fail if it needs to fetch a second page, with the following error:

microsoftgraph.exceptions.BadRequest: {'error': {'code': 'BadRequest', 'message': "Query option '$top' was specified more than
once, but it must be specified at most once.", 'innerError': {'date': '2022-02-18T23:21:56', 'request-id': 'ae804f84-62aa-4bb4-a521-8c5ad8be2ed7', 'client-request-id': 'ae804f84-62aa-4bb4-a521-8c5ad8be2ed7'}}}

The reason for this error is that, when we want to fetch the next page, we use "@odata.nextLink" as request url which already contains all query params from the original request but we still pass **kwargs to the requests.request constructor, so the values in the "params" dictionary get added again to the request, causing all request query params to be duplicated.

Hey there, with this flow, what step do you recommend to pass "code" to exchange code for access token. I understand the user would login in the browser. Sorry if the question is low level, still learning more about Graph API

This line of code started raising sometimes for me, as of 2 months ago

Screenshot from Sentry with some more context:

Some responses (like delete calendar event) don't include a Content-Type header.

Will send pull-request

In the file mail.py:

line 32 url = "me/mailFolders/{id}/messages".format(folder_id) if folder_id else "me/messages"

should be

url = "me/mailFolders/{id}/messages".format(id=folder_id) if folder_id else "me/messages"

or better again use f-strings!

Here's the simplistic code for a head start that works for me. Of course you can edit parts to make it conventional, but adding this in is a huge value add and makes it easier for people to pick up the API.

@token_required
    def update_event(
        self,
        start_datetime: datetime,
        start_timezone: str,
        end_datetime: datetime,
        end_timezone: str,
        id: str,
        **kwargs,
    ) -> Response:
        """Update an event in the user's default calendar or specified calendar.

        https://docs.microsoft.com/en-us/graph/api/user-post-events?view=graph-rest-1.0&tabs=http

        Additional time zones: https://docs.microsoft.com/en-us/graph/api/resources/datetimetimezone?view=graph-rest-1.0

        Args:
            subject (str): The text of the event's subject line.
            content (str): The body of the message associated with the event.
            start_datetime (datetime): A single point of time in a combined date and time representation ({date}T{time};
            start_timezone (str): Represents a time zone, for example, "Pacific Standard Time".
            end_datetime (datetime): A single point of time in a combined date and time representation ({date}T{time}; for
            end_timezone (str): Represents a time zone, for example, "Pacific Standard Time".
            location (str): The location of the event.
            calendar_id (str, optional): Calendar ID. Defaults to None.
            content_type (str, optional): It can be in HTML or text format. Defaults to HTML.

        Returns:
            Response: Microsoft Graph Response.
        """
        if isinstance(start_datetime, datetime):
            start_datetime = format_time(start_datetime)
        if isinstance(end_datetime, datetime):
            end_datetime = format_time(end_datetime)

        body = {
            "start": {
                "dateTime": start_datetime,
                "timeZone": start_timezone,
            },
            "end": {
                "dateTime": end_datetime,
                "timeZone": end_timezone,
            },
        }
        body.update(kwargs)
        url = "me/events/{}".format(id)
        return self._client._patch(self._client.base_url + url, json=body)

    @token_required
    def delete_event(
        self,
        id: str
    ) -> Response:
        """Delete an event in the user's default calendar or specified calendar.

        https://docs.microsoft.com/en-us/graph/api/user-post-events?view=graph-rest-1.0&tabs=http

        Additional time zones: https://docs.microsoft.com/en-us/graph/api/resources/datetimetimezone?view=graph-rest-1.0

        Args:
            id (str): Event ID.

        Returns:
            Response: Microsoft Graph Response.
        """
        url = "me/events/{}".format(id)
        return self._client._delete(self._client.base_url + url)

Need to add attributes for the subscription webhook attributes like
lifecycleNotificationUrl(important to receive notification whenever user changes password of account subscription would be cancelled)
includeResourceData(Really important to avoid multiple api calls to retrieve resource)

For some reason AUTHORITY_URL and OFFICE365_AUTHORITY_URL are switched, so that the behavior of the office365 parameter to Client's constructor is the opposite of the expected behavior.

Just wondering, since there were some fixes done, is 1.1.1 going to be released soon?

We could copy the library to our local project but I prefer to actually pull it from pypy and avoid having local changes to a library, but rather PR the fixes here as I did

I got

python3.9/site-packages/microsoftgraph/client.py", line 250, in _request
_headers["Authorization"] = "Bearer " + self.token["access_token"]
TypeError: 'Response' object is not subscriptable

when I call client.files.drive_root_items() after the authentication phase.

I saw that self.token is an object with attributes data, original, status_code, throttling and data contains the key access_token

So I solved by changing line 250 in client.py to _headers["Authentication"] = "Bearer " + self.token.data["access_token"]

But now in line 215 of client.py value is not a key of data

BG

Would be nice if there was a version that utilised aiohttp instead of requests :) Might make a PR for it but I'd like to hear your thoughts

I've used exchange_token_result = client.exchange_code(REDIRECT_URL, code) to successfully exchange a code for a token, but there is no refresh_token in the response, which I anticipated from https://docs.microsoft.com/en-us/graph/auth-v2-user?context=graph%2Fapi%2Fbeta&view=graph-rest-beta#token-response

In the docs of this repo it is possible to call token = client.refresh_token(redirect_uri, refresh_token), but I don't know how I would obtain the refresh token to begin with. I tried using the access_token returned from exchange_code but that returned an Invalid Request exception.

The _paginate_response method in the client performs pagination via recursion:

• You call and endpoint, which calls _get, which calls _paginate_response, which on a while loop calls _get, which calls _paginate_response, etc

There are a few problems with this approach:

• If the collection being requested is very big, and the page size is the default (10 items) this can lead to hundreds of stacked calls, and ultimately an exception will be raised
• If we hit a faulty response (i.e. one where response.data is not a dict instance, which is one of the conditions for just returning response, that return will come inside the while loop, and we will do response.data["value"] += data, but if we reach through that path, response.data is not a dict, so response.data["value"] might fail if data is not a subscriptable object.

This could be solved by calling self._request directly on the while loop inside the _paginate_response method, rather than _get, which eliminates recursion. The check for response.data being a dict instance should be performed there as well to avoid the potential exception.

Calling 'client.mail.list_messages()' and 'client.mail.get_message({message_id})' end up returning only the response status code and not what they're intended to return; i.e., a list of messages and the contents of a message.

Also tried fetching the json of the response object 'msgs', but it resulted in an attribute error which stated that the response object has no attribute 'json'.

Please add a public "get" method in case I need some API access point that's not provided by the client class. All this function needs to do is paste the argument to base_url and send the request.

So you would use it like this:
item = client.get("/me/drive/root:/path/to/file", params=None)

And same for post and delete. Thanks for this module!

Can you include a working example of sending mail with multiple attachments?

Initially I thought that the library doesn't include mail functions as they were not documented in the readme. Just figured out which I was browsing the source code.

Thanks in advance!

Does this library support uploading files to OneDrive? Are there any examples?

Here are the relevant API docs: https://docs.microsoft.com/en-us/graph/api/driveitem-createuploadsession?view=graph-rest-1.0

Currently getting the above exception with the attached message: 'message': '/me request is only valid with delegated authentication flow.'

Any suggestions on how to resolve?

I've seen on the docs during Azure app registration, redirect_uri is optional here
https://docs.microsoft.com/en-us/graph/auth-register-app-v2?context=graph%2Fapi%2F1.0&view=graph-rest-1.0

So refresh token and get token does not require it as necessary, it is optional

When using the client with office365=True, the decorator says you require a token even if it is set.

this solved it for me:

def token_required(func):
    @wraps(func)
    def helper(*args, **kwargs):
        client = args[0]
        if not (client.token or client.office365_token):
                raise TokenRequired('You must set the Token.')
        return func(*args, **kwargs)

    return helper

When specifying a folder in list_messages, there's an error with the string substitution in the URL. Trivial fix, submitted as #43