Refactor pagination to preserve request parameters

[Copilot]

The pagination logic was setting params = None after the first page, relying on GitHub's Link header URL to contain all query parameters. This is fragile and makes the code harder to maintain.

Changes

• Parse next URL to extract page parameter - Use urllib.parse to extract only the page value from the Link header URL
• Preserve original params dict - Update only the page parameter instead of replacing the entire params object
• Validate page parameter - Ensure page is a positive integer; log warning and stop pagination on invalid values (missing, non-numeric, zero, negative)

Before

# Pagination
next_url = resp.links.get("next", {}).get("url")
if not next_url or len(batch) == 0:
    break
base_url, params = next_url, None  # Fragile: relies on next URL having all params

After

# Pagination
next_url = resp.links.get("next", {}).get("url")
if not next_url or len(batch) == 0:
    break
parsed_url = urlparse(next_url)
query_params = parse_qs(parsed_url.query)
if "page" in query_params and query_params["page"]:
    try:
        page_num = int(query_params["page"][0])
        if page_num > 0:
            params["page"] = page_num
        else:
            logger.warning(f"Invalid page number {page_num}, stopping pagination")
            break
    except (ValueError, IndexError) as e:
        logger.warning(f"Invalid page parameter: {e}, stopping pagination")
        break

This maintains the original request parameters (state, per_page, sort, direction) across all pages and prevents infinite loops from malformed pagination URLs.

---

💬 We'd love your input! Share your thoughts on Copilot coding agent in our 2 minute survey.