fix(amazon): Amazon Business URL + hash-based pagination

[arminfauland]

Problem

The existing Amazon config had two critical bugs for Amazon Business accounts:

1. Wrong URL: your-orders/orders only shows a subset of orders. Amazon Business uses /gp/css/order-history — missing ~60% of orders (6 pages found vs 14 actual pages).

2. Broken pagination: ?startIndex=N is ignored by the React SPA. The page always showed the same first 10 orders regardless of startIndex.

Fix

Navigate directly via hash-based routing which the React SPA uses natively:

https://www.amazon.de/gp/css/order-history#time/2024/pagination/2/

No dropdown interaction needed — year and page are fully encoded in the URL hash.

Additional fixes

• extractedYears selector: option values are "2024" not "year-2024" in #timeFilterDropdown
• extractedYearsNormalized: normalizes to internal year-XXXX format via $substring (not $startsWith — not a JSONata built-in)
• maxPages=0 treated as "no limit" (same as empty)
• shouldBreak disabled when explicit targetYear is set — historical runs must process all pages
• yearDisplayText moved before Navigate to year action (was computed after, causing empty hash)

Tested

Amazon Business DE account — 2024: 14 pages / 136 orders / 198 PDFs successfully downloaded to Paperless-NGX consume directory.

Previously (broken): 6 pages / ~60 orders / same 10 orders repeated on every "page".

[github-actions]

👋 Thanks for opening your first pull request! We're excited to have you contribute to Scrape Dojo.

A maintainer will review this shortly. Please make sure:

• ✅ Your code follows the project's code style
• ✅ You've tested your changes locally
• ✅ All existing tests still pass

Check out our Contributing Guide for more info!

[Disane87]

Review Feedback: Compatibility with Standard Amazon (Consumer) Accounts

Great fix for Amazon Business! 🎉 The hash-based pagination and switch to /gp/css/order-history solve the problem cleanly.

However, we have concerns about whether this fix also works with regular Amazon consumer accounts (non-Business):

Specific Points to Verify

1. Hash-based pagination (#time/2024/pagination/2/): Standard consumer accounts may still use the server-rendered order history, where hash routing is ignored and ?startIndex=N is the correct approach. Can you verify that pagination also works on a regular consumer account?

2. clickByText: "Rechnung": This is DE-locale-specific. On EN accounts or .com domains, this would need to be "Invoice". Does this work on your account because it's set to DE, or have you tested with other locales as well?

3. #timeFilterDropdown vs. #time-filter: The fallback selector for #time-filter is included — good. But have you been able to verify whether standard accounts use the React dropdown or the classic <select>?

Request

Could you test this fix with a regular Amazon consumer account (non-Business) as well? We unfortunately lack experience with the React SPA version of the order history, and we want to make sure there's no regression for existing standard users.

If it turns out that standard and Business accounts get different page versions, a separate amazon-business.jsonc config would be a clean alternative.

Thanks! 🙏

[github-actions]

This PR has been automatically marked as stale because it has not had recent activity. It will be closed in 14 days if no further activity occurs.

[github-actions]

This PR was closed because it has been stale for 14 days with no activity. Feel free to reopen if still relevant.