mirror of
https://github.com/calibrain/shelfmark.git
synced 2026-09-24 20:00:27 +01:00
splitsec2
7c8e89c567
The Google Books search builds `maxResults` as `min(limit, 40)` because the API caps a page at 40 volumes, but advances `startIndex` by the full `limit`. The pages then stop tiling. With `limit=50`, page 1 covers items 0 to 39 and page 2 starts at 50, so items 40 to 49 are never returned and every later page drops another 10. This computes the page size once and uses it for both `maxResults` and `startIndex`. The shipped frontend asks for 40 and is unaffected. `/api/metadata/search` clamps `limit` to 100, so an API caller passing 41 to 100 was hitting it. One thing I left alone. The provider uses the base `search_paginated` heuristic, `has_more = len(books) >= options.limit`, which still reports `has_more: false` for a limit above 40 since Google can never return that many. That was already the behaviour before this change, and fixing it means either touching the shared heuristic or adding a provider override, so I kept this patch to the stride. Happy to follow up if you want it. ## Verification - `tests/metadata/test_googlebooks_parse.py`: pages 1 and 2 at `limit=50` must tile exactly, plus a guard that `limit=25` still strides by 25. The first fails on current main and passes here. - Full suite (3165), ruff, ruff format, basedpyright, vulture green.