IMIO / imio.smartweb.core / 32344333884

89%

push

github

Add a control panel to include/exclude and cap (max 50, with ordering) each authentic source in the HTML and XML sitemaps. (#90)

* Add a control panel to include/exclude and cap (max 50, with ordering) each authentic source in the HTML and XML sitemaps.

* Sitemap config: drop the per-source filter, cap up to 1000 items

The ordering is fixed per authentic source (upcoming events for the
agenda, most recent news, most recently modified contacts), so the filter
column had a single value and only invited drift. The cap ceiling goes
from 50 to 1000, with defaults of 50 for the agenda and the news and 200
for the directory: a catalogue capped low keeps the same handful of
contacts in the sitemap forever and leaves the rest undiscovered.

* Fetch sitemap source items page by page with a time budget

A single @search for the whole cap did not answer within the endpoint's
20 s timeout, and get_json() returns None on timeout, so the source
disappeared from the sitemap entirely. Pages of 100 items are now fetched
until the cap or the budget is spent, keeping whatever came back: items
are sorted newest-first, and the tail stays crawlable through seo_html.

* Make the seo_html crawlable fallback actually crawlable

Its link list sat in a <noscript> under a window.location.href redirect,
so Googlebot, which executes JavaScript, only ever saw the redirect. The
list is now rendered plainly and the page keeps its X-Robots-Tag:
noindex, follow.

Its entry point is declared in sitemap.xml only: in the HTML sitemap that
citizens browse, clicking it landed them on the fallback instead of the
React listing. seo_html no longer self-links either, and the six
"... : SEO Links" msgids become one per source: All contacts, All events,
All news.

* Align the seo_html rel=prev/next links with the page batch size

They fell back to a hardcoded b_size of 10 while the page lists 100
items, so the chain announced to crawlers disagreed with the page's own
pagination links and walked... (continued)

125 of 137 new or added lines in 8 files covered. (91.24%)

5995 of 6759 relevant lines covered (88.7%)

0.89 hits per line