Add cross-repo suite badges to README

Add suite links to README

Merge celery-reliability into dev

Merge rename-armour-to-armor into dev

Rename armour to armor + gitignore celerybeat

New models: DataAchievement, DataFaction, ProfileAltAchievement,
ProfileAltReputation with migration 0011.

fullDataScan now fetches achievement and reputation-faction indexes
from the Blizzard static API to populate the Data* tables.

scan_single_alt now fetches /achievements and /reputations per alt;
both sync helpers create Data* records lazily on first encounter.

purge_stale_profiles extended to clean up expired achievement and
reputation records.

New endpoints: /api/profile/altachievements/, /api/profile/altreputations/,
/api/data/achievements/, /api/data/factions/ — all support ?user=,
?alt=, ?realm= query params with session auth enforcement.

Excludes migrations from ruff lint.

- Add DataAchievement, DataFaction, ProfileAltAchievement, ProfileAltReputation models
- Add alt_ilvl field to ProfileAlt (populated from character summary on scan)
- Refactor fullAltScan: move mounts/pets/achievements to scan_user_collection
  which scans one highest-level/ilvl alt per faction instead of all alts
- Split fullDataScan into 5 independent subtasks with individual endpoints
- Add logout endpoint that flushes Django session
- Fix standing_value to IntegerField (Blizzard returns negative for Hated)

…ions

Merge phase-2a-achievements-reputations into dev

…view

- DataFaction.faction_category: new field (migration 0014) populated via
  hardcoded FACTION_EXPANSION map (283 factions, Classic → Midnight)
- _sync_reputations: use standing.name not standing.type (was always empty);
  derive faction_category from FACTION_EXPANSION map
- ProfileAltAchievementView: add ?summary=1 (category aggregates) and
  ?category=X (per-category lazy load) to avoid returning all achievements
- Achievement/rep serializers: expose alt_name and faction_category
- _sync_achievement_data: add progress logging every 100 achievements

Merge phase-2a-ui-data-fixes into dev

Keybind scraping/parsing/display was too niche to maintain. Deletes
keybind_builder.py, keybind_mapping.py, and their tests; trims
ProfileUserView.list to the generic page=header response plus the
parse-and-cache scaffolding (renamed cache key from keybinds:{user}
to userfile:{user}) for future addon-only data. tier_sort_key was
also used for profession tier sorting, not just keybinds, so it
moves to a new apicore/libs/expansion_order.py rather than being
deleted.

A lingering Django admin session (from triggering DataScan via
/api/admin/) made DRF's SessionAuthentication enforce CSRF on every
API call, since profile/custom endpoints use the custom IsSessionUser
permission rather than Django auth. Removed SessionAuthentication from
the default authenticators and scoped it explicitly to the
IsAdminUser-gated DataScan* viewsets that actually need it.

Adds DataMythicDungeon (static catalog), ProfileAltMythicPlus
(per-alt season rating summary), and ProfileAltMythicPlusDungeon
(per-alt per-dungeon best run) with migration 0015.

scanMythicDungeonData populates the dungeon catalog via
/data/wow/mythic-keystone/dungeon/index (namespace dynamic-eu, not
static-eu) and is added to fullDataScan. _sync_mythic_plus is wired
into scan_single_alt: the main mythic-keystone-profile endpoint only
lists season refs with no "current" flag, so max(season.id) picks
the current one, then a second call to the season-specific endpoint
fetches mythic_rating and best_runs. Verified against a live
character that best_runs can list two entries per dungeon
(best-timed vs best-overall, same map_rating but different
keystone_level) — deduped to the higher level per dungeon before
writing.

New endpoints: altmythicplus, altmythicplusdungeons, mythicdungeons,
datascan/mythicdungeons.

Runs every Sunday 03:00 UTC alongside the existing daily
purge_stale_profiles schedule. The admin-triggered datascan
endpoints are kept for ad-hoc manual scans.

perform_update closed the uploaded file's stream via a context manager
before handing it to the storage backend, which then failed trying to
re-read it. Validate and normalise the content directly, then save it
as a fresh ContentFile instead of reusing the original upload stream.

ProfileUserView.list opened the stored .lua file with mode "r" (text),
so readlines() already returned str, and the subsequent .decode("utf-8")
call crashed. This path was previously unreachable in practice (no page
read the file before page=addon was added) so the bug went unnoticed.

Replaces the page=addon read-through cache (which re-parsed the raw
.lua file on every cache miss, with no expiry) with parse-on-upload
into a new ProfileAltAddonData model, matching how every other
Profile* model is populated. Adds the altaddondata endpoint and drops
the now-dead generic file-read branch from ProfileUserView.list.

Merge remove-codecov into dev