… pages

These two pages live in this repo (not migrated from clickhouse-docs)
and contained anchor typos that the broken-link sweep flagged.

invertedindexes.mdx (backward-compat alias for textindexes):
- mapkeys/mapvalues/mapcontains -> mapKeys/mapValues/mapContainsKey
  to match the actual heading IDs in tuple-map-functions.
- Tokens cache settings: refresh the table to match the current setting
  names (text_index_dictionary_block_cache_* -> text_index_tokens_cache_*
  and use_text_index_dictionary_cache -> use_text_index_tokens_cache).
  The settings were renamed upstream; this alias page hadn't followed.

pg_clickhouse/reference.mdx:
- Lowercase function anchors -> camelCase (trimBoth, concatWithSeparator,
  lowerUTF8/upperUTF8/substringUTF8/lengthUTF8/reverseUTF8, positionUTF8,
  bitCount, toUInt8..128).
- ltrim/rtrim links point at trimLeft/trimRight (the actual ClickHouse
  function names).
- arrayStringConcat now lives in splitting-merging-functions, not
  array-functions.
- reverse(bytea) link points to array-functions#reverse (the bare reverse
  function is on the array-functions page).
- regexpExtract is documented in string-functions, not string-search-functions.
- Composite #dictget-dictgetordefault-dictgetornull split into three
  separate links to dictGet, dictGetOrDefault, dictGetOrNull.
- Window functions now have their own per-function pages, so links go
  to /window-functions/<name> instead of /window-functions#<name>.

Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>

This page was a hand-authored alias of textindexes.mdx, kept to make the
old Docusaurus URL /engines/table-engines/mergetree-family/invertedindexes
render. The content had drifted from textindexes (stale setting names,
typo anchors) and was a perpetual maintenance burden.

External links to the old URL are already handled:
- _site/redirects.json redirects /engines/.../invertedindexes ->
  /core/reference/.../textindexes
- _migration/slug-aliases.csv rewrites any upstream link to the
  old slug at migration time

The previous commit's anchor fixes to this file are obviated by the
deletion. Also removed from core/reference/navigation.json.

Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>

The corresponding upstream commit on ClickHouse/clickhouse-docs#6278 was
reverted, so the local anchor fixes here would have been clobbered the
next time migrate.py --force regenerates the file from upstream.

This PR now contains only the invertedindexes.mdx deletion.

Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>

This page (converted VLDB paper) uses ~100 inline <a id="page-N-M">
figure/citation anchor tags produced by migrate.py's transform_inline_anchor
from upstream <Anchor> components. The links work in-browser, but
mint broken-links --check-anchors only recognizes markdown heading slugs,
so all 106 anchor references on this page report as broken.

Adding to .mintignore drops the checker noise from 219 to 118 entries.
Trade-off: 5 inbound links from other pages (mostly query-parallelism.mdx)
now report as broken since the checker can't see academic-overview's real
heading slugs anymore. Those links work in browser; the false positives
are mild.

Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>