fix(docs): Implement exponential backoff for GitHub API to resolve HTTP 429/403 crashes

[williamcancodee]

Description

The PR attempts to resolve the documentation generation crash caused by GitHub API rate limiting (HTTP Error 429/403), as indicated by issue #924.

In the previous implementation, when an HTTP error occurred, fetch_url() would return an empty dictionary. Consequently, when get_paged_request() tried to access f.status, an AttributeError would occur. However, even if this did not occur, the 5-attempt retry loop would terminate instantaneously without a delay.

Changes Made:

• Fixed Error Handling: Modified fetch_url() to return the real HTTPError object (which naturally has a .code/.status attribute in recent versions of Python) or a dummy object for other types of exceptions to avoid AttributeError crashes.
• Exponential Backoff: Introduced a time.sleep() for 429 (Rate Limit) and 403 (Forbidden/Abuse) status codes.
• The script will now pause for progressively longer durations (1s, 2s, 4s, 8s, 16s) between attempts, giving the API quota window a chance to reset.
• Added print statements to inform the user that the script is deliberately pausing rather than freezing.

Fixes #924

Type of change

• Bug fix (non-breaking change which fixes an issue)

Checklist:

• My code follows the style guidelines of this project
• I have performed a self-review of my own code