atlassian.confluence: export_page throws 404 Client Error

[psys-om]

We've been using this library (version 4.0.7) in an automated daily job for quite a while with no issue. Since 10 April the page action confluence.export_page(page_id) fails with the error

404 Client Error: Not Found for url: https://<org>.atlassian.net/wiki/services/api/v1/task/module-<some_uuid>/progress

for all our pages I tested against.

The connection is set up like this:

confluence = Confluence(
    url=<url>,
    username=<username>,
    password=<token>,
    api_version='cloud',
    cloud=True
)

I also tried with the new implementation mentioned in the docs, but I get

from atlassian.confluence import ConfluenceCloud
ImportError: cannot import name 'ConfluenceCloud' from 'atlassian.confluence'

[paulbrandariz-netizen]

Not yet familiar with submitting code changes, but the issue involves modifying:

.local/lib/python3.9/site-packages/atlassian/confluence/init.py

[paul@vmbanzai confluence]$ diff save__init__.pyw __init__.py
3353c3353
<             poll_url = f"/services/api/v1/task/{task_id}/progress"
---
>             poll_url = f"/api/v2/pdfexporttask/progress/{task_id}"
3366c3366
<                     task_result_url = progress_response.get("result")
---
>                     download_url = progress_response.get("result")
3370,3373d3369
<             log.debug("Task successfully done, querying the task result for the download url")
<             # task result url starts with /wiki, remove it.
<             task_content = self.get(task_result_url[5:], not_json_response=True)
<             download_url = task_content.decode(encoding="utf-8", errors="strict")

It appears the new Confluence updated in March forces use of the v2 API.

good luck

[psys-om]

Thanks for the patch!

[maconfr]

I also run into this issue with v4.0.7. I solved it by sub-classing the Confluence class in my project and just overwrite the Confluence.get_pdf_download_url_for_confluence_cloud() Method.

class CustomConfluence(Confluence):
    def get_pdf_download_url_for_confluence_cloud(self, url):
        try:
            running_task = True
            headers = self.form_token_headers
            log.info("Initiate PDF export from Confluence Cloud")
            response = self.get(url, headers=headers, not_json_response=True)
            task_id = (
                response.split(b'name="ajs-taskId" content="', 1)[1]
                .split(b'">', 1)[0]
                .decode(encoding="utf-8", errors="ignore")
            )
            poll_url = f"/api/v2/pdfexporttask/progress/{task_id}"
            download_url = None
            while running_task:
                log.info("Check if export task has completed.")
                progress_response = self.get(poll_url)
                percentage_complete = int(progress_response.get("progress", 0))
                task_state = progress_response.get("state")
                if task_state == "FAILED":
                    log.error("PDF conversion not successful.")
                    return None
                elif percentage_complete == 100:
                    running_task = False
                    log.info(f"Task completed - {task_state}")
                    log.debug("Extract task results to download PDF.")
                    download_url = progress_response.get("result")
                else:
                    log.info(f"{percentage_complete}% - {task_state}")
                    time.sleep(3)
            if not isinstance(download_url, str):
                log.error("PDF conversion not successful.")
                return None
            log.debug("Task successfully done, querying the task result for the download url")
            log.debug("Successfully got the download url")
            return download_url
        except IndexError as e:
            log.error(e)
            return None

i just copied the original method and changed the url and a few other things to make it work again with the new API endpoint.
Here is a diff to outline the changes:

diff --git a/atlassian/confluence/__init__.py b/atlassian/confluence/__init__.py
index 08b8bbe..7826543 100644
--- a/atlassian/confluence/__init__.py
+++ b/atlassian/confluence/__init__.py
@@ -3348,9 +3348,13 @@ class Confluence(AtlassianRestAPI):
             headers = self.form_token_headers
             log.info("Initiate PDF export from Confluence Cloud")
             response = self.get(url, headers=headers, not_json_response=True)
-            response_string = response.decode(encoding="utf-8", errors="ignore")
-            task_id = response_string.split('name="ajs-taskId" content="')[1].split('">')[0]
-            poll_url = f"/services/api/v1/task/{task_id}/progress"
+            task_id = (
+                response.split(b'name="ajs-taskId" content="', 1)[1]
+                .split(b'">', 1)[0]
+                .decode(encoding="utf-8", errors="ignore")
+            )
+            poll_url = f"/api/v2/pdfexporttask/progress/{task_id}"
+            download_url = None
             while running_task:
                 log.info("Check if export task has completed.")
                 progress_response = self.get(poll_url)
@@ -3363,14 +3367,14 @@ class Confluence(AtlassianRestAPI):
                     running_task = False
                     log.info(f"Task completed - {task_state}")
                     log.debug("Extract task results to download PDF.")
-                    task_result_url = progress_response.get("result")
+                    download_url = progress_response.get("result")
                 else:
                     log.info(f"{percentage_complete}% - {task_state}")
                     time.sleep(3)
+            if not isinstance(download_url, str):
+                log.error("PDF conversion not successful.")
+                return None
             log.debug("Task successfully done, querying the task result for the download url")
-            # task result url starts with /wiki, remove it.
-            task_content = self.get(task_result_url[5:], not_json_response=True)
-            download_url = task_content.decode(encoding="utf-8", errors="strict")
             log.debug("Successfully got the download url")
             return download_url
         except IndexError as e:

Hope this helps others with this problem.