How to connect to Atlassian python API using Auth 2.0 Bearer Token

Question

Hi - I am trying to connect to Confluence with the Atlassian python API using Oauth 2.0 with a bearer token as required by my company because we have our own domain and use multi-factor authentication.

The link below shows the different authentication methods. 
https://atlassian-python-api.readthedocs.io/index.html

I have tested my token and URL using the below on the command line and it works (token and content shown below are fake)
curl -H "Authorization: Bearer XYMWODLW" https://wiki.mycompany.com/rest/api/content/123

I am having trouble translating this to one of the authentication methods in the link. Bearer tokens are typical in Oauth 2.0. When I reference the Oauth 2.0 section, I get the below code.

from atlassian.bitbucket import Cloud

# token is a dictionary and must at least contain "access_token"
# and "token_type".
oauth2_dict = {
"client_id": client_id,
"token": token}

bitbucket_cloud = Cloud(
oauth2=oauth2_dict)

# For a detailed example see bitbucket_oauth2.py in
# examples/bitbucket

My questions: 
1. Does this also work for Confluence?
2. Where do I get client_id?
3. What is the actual format? Example shown below. (I was unable to find bitbucket_oauth2.py to reference)

oauth2_dict = { "client_id": client_id,

"token": {"access_token" : "XYMWODLW",

"token_type" : "Bearer"

}

girija kakani · Answer

Hi everyone,

I’m new to working with the Confluence API and I’m trying to write a Python script to access my organization’s Confluence space and export the content.

Here’s the situation:

Our company uses SSO (Single Sign-On) for Confluence login.

I’m using the atlassian-python-api library.

My script looks like this:

import os

from atlassian import Confluence

confluence = Confluence(

url=os.getenv("CONFLUENCE_URL"),

username=os.getenv("CONFLUENCE_USER"),

password=os.getenv("CONFLUENCE_TOKEN")

)

However, every time I run it, I get a 401 Unauthorized error.

I’m guessing this might be related to our SSO setup - maybe username/password or API tokens aren’t valid with SSO.

* Could someone explain what additional information or credentials I might need from my IT/admin team to successfully connect? For example:
* Do I need an OAuth 2.0 access token instead?
* Should I request a client ID / client secret to generate a token?

Are there specific API scopes or permissions they need to grant me?

Any guidance or steps on how to properly authenticate when SSO is enabled would be really appreciated!

Ishan Kaul_g9wpcl · Answer

UPDATE: I have found the solution finally. Since Atlassian's python client library now points to the v2 API, the url passed to Confluence.__init__ needs to be structured like

(https)"api.atlassian.com/ex/confluence/{client_id}/wiki/" (NOTE: the url provided in the python example for the v2 docs is incorrect as of this writing). The python client will inject "rest/api/content/{page_id}" into the path when it calls get_page_by_id.
So for anyone trying to do this by making a request, the final url in the GET request is 
api.atlassian.com/ex/confluence/{client_id}/wiki/rest/api/content/{page_id}.

Note that I also went on a mad frenzy adding both classic and granular read scopes so I'm not sure which ones actually allowed it to work

Ishan Kaul_g9wpcl · Answer

Hi , @equigley2 @Yasir Kusay I am attempting what I think is the same code, but still getting an HTTPError. My code looks like

s = Session()s.headers["Authorization"] = f"Bearer {access_token}"confluence = Confluence(url=base_url, session=s)page_contents = confluence.get_page_by_id(page_id)

I get my access_token from the oauth 2.0 handshake with confluence, and base url through another request to the atlassian accessible-resources apiThe scopes returned from accessible resources from the token looked like:

#readonly:content.attachment:confluence

#read:confluence-content.all
#read:confluence-content.summary
#read:confluence-user
#search:confluence
#read:confluence-space.summary
so I don't think it's a permissions issue.

Is there anything you guys did differently to get it working?