How to access Confluence use atlassian-python-api and a Personal Access Token

[nirabe]

I can use the following for basic auth:

from atlassian import Confluence

confluence = Confluence(url='https://internalconfluenceurl.com', username = 'myusername', password='mypassword')

This also works if I use an API token value instead of password.

How do I alter this to use a Personal Access Token (as created within Confluence) instead? I tried substituting that in for the password, but it gives an error.

Is there a further authentication step that I am missing? I have managed to do something similar in Jira, but the same approach does not work for Confluence. 

[nirabe · Accepted solution]

I found a different approach:

import requests

response = requests.post(base_url, data=page_data_json, headers={'Authorization': f'Bearer {pat}',"Accept": "application/json",'Content-Type': 'application/json'})

if response.status_code == 200:
print("Child page created successfully")
else:
print("Failed to create the child page. Status code:", response.status_code)

It is not quite as convenient as using the inbuilt Confluence function, but it seems to work. 

[Akshayaa Pandiyan]

what is page_data_json?

[weitzj]

In case you are using the Python SDK to authenticate against Confluence using a Personal Access Token (PAT), the example above did not work for me, since somehow the urllib3 HTTPConnectionPool "forgets" about the "Bearer" Token, but uses a "Basic" Auth token, which is wrong.

The hack to circumvent is, is to introduce your own HTTP Adapter:

import os
import requests
from atlassian import Confluence
from requests.adapters import HTTPAdapter

class CustomHTTPAdapter(HTTPAdapter):
    def __init__(self, headers=None, *args, **kwargs):
        self.default_headers = Confluence.default_headers
        self.default_headers.update(headers)
        super().__init__(*args, **kwargs)

    def send(self, request, **kwargs):
        # Merge default headers with request-specific headers
        request.headers.update(self.default_headers)
        return super().send(request, **kwargs)


session = requests.Session()
adapter = CustomHTTPAdapter(headers={"Authorization": f"Bearer {os.environ["CONFLUENCE_TOKEN"]}"})
session.mount(os.environ["CONFLUENCE_HOST"], adapter)
confluence = Confluence(session=session, url=os.environ["CONFLUENCE_HOST"])

[GMK]

Hello All, 

First of all it is worth noting  atlassian-python-api   is an open source project, it is not owned/maintained by Atlassian. 

You can use personal assess token with this library, the only difference is in syntax between cloud and data-center:

from atlassian import Jira
# To be used with Jira cloud
jira_cloud = Jira(url="<url>", username="username", password="password")

# Use with data-center
jira_dc = Jira(url="url", token="<token>>")

[Sagar Mahajan]

Hi @nirabe Welcome to the Atlassian community.

Have you check the below document for how to use the PAT in Jira/Confluence instead of basic auth?

https://confluence.atlassian.com/enterprise/using-personal-access-tokens-1026032365.html

Below is the example of using PAT same way you need to modify to use in Python script

curl -H "Authorization: Bearer <yourToken>" https://{baseUrlOfYourInstance}/rest/api/content

[nirabe]

Hi @Sagar and thanks for the prompt reply.

Yes, I had seen the link you shared. I cannot use cURL. There have been some suggestions of using requests instead.

I may go down this route, but it seems much more complicated than the approach for Jira, which is odd.