Schema math (parts, chapters, sections, activity), seeded with all 100 chapters
(Ch 0-94, appendix A-E), 647 sections, PDF pages, exercise counts and approximate
prereqs. Triggers stamp started_at/completed_at and log status/exercise/note/section
changes; math.apply_progress(jsonb) applies the JSON format bots send. Private by
default via MATH_PRIVATE in #lib/math (owner-only prefix + lock in every tracker list).
Overall + per-part progress, current chapter, prereq hints, quick status buttons,
exercises stepper, notes, optional section ticks, activity log, dependency chart.
Deploy preflight checks the math schema (4 tables, 100 chapters).
- cs.skins catalog (weapons, knives, gloves, stickers, cases, agents, charms, …) refreshed weekly
- every market config: wears in float range, StatTrak, Souvenir, Doppler/Gamma phases
- cs.daily_prices (one row per config/stream/day, idempotent) from Skinport, Steam, CSFloat, Buff163
- cs.price_windows for 24h/7d/30d/90d sale aggregates, cs.fx_rates (ECB)
- scripts/cs-prices.ts daily|catalog|prices; legacy ingest fixed for the new unique index
- skin browser + /projects/cs-prices/[slug] with variant/wear/phase switcher and real charts
- charts show points for short histories and pad sparse time axes
Vault files stay the source of truth (bind-mounted read-write at /notes):
atomic writes chowned to the vault owner, hash-checked saves (409 on a
newer file), deletes to .trash. notes.* is a full-text/link index kept
current by a recursive watcher plus a 60s rescan; live updates over the
owner-only /ws/notes channel (paths and hashes only). Wikilinks, embeds,
backlinks, tags, callouts, image paste. Owner-only everywhere (404).
- /stats/economy defaults multi-country charts to CAN KOR CHN JPN USA NLD CUW,
with add-a-country chips and an EU 27 comparison line
- CPI index levels rebased to a chosen base year, index/YoY toggle,
categories (food, energy, housing, core) where the source has them
- new no-key collectors: oecd (CPI categories, 10y yields), statcan (CPI);
bis WS_LONG_CPI, FRED US CPI categories, Eurostat prc_hicp_minr, WB FP.CPI.TOTL
- Canada added as a region; CBS Curacao fallbacks for growth/unemployment
- timers stats-oecd and stats-statcan
Read-only layers from static/pubg-intel/<map>.v1.json (scripts/build-pubg-intel.py):
positions from pubg.gs (game level files), drops from guides, secret rooms and Deston
keys as approximate rings on the named place. Layer panel (K), per-player choices in
localStorage, drawn below room markers; clicks still place markers.
Long-run charts (Maddison, HYDE, UN via OWID; CBS and Pew religion), a filterable
sourced timeline, Wikidata office holders and turnout, Nager.Date holidays, UNESCO
sites, Wikidata people by sitelinks, curated food and cultural exports. Wikidata
labels now fall back to mul.
- CBS_SERIES: CPI, household tariffs (2007-now), energy balance, renewables, heat pumps,
solar, cars, population (monthly, since 1942, components), house prices, housing stock,
municipal levies for Gouda and its neighbours
- shared StatChart/Kpi/chart helpers with per-chart source attribution
- drop Eurostat nrg_pc_20x_h (404; CBS covers that history)
- search.ts: score on the core name (brand/size stripped): exact or query+base suffix
(kipdij+filet), head-final compounds, prefix, inside; penalties per extra word and per
flavour/processing word (search-words.ts, editable); bonuses for store brands, the
query's dominant category, store count, favourites; trigram tiebreak
- groceries.synonyms table (seeded: kipdij/kippendij, pasta → spaghetti, …) expands queries
- list rows show all pack sizes of a group; group picker uses the same ranking
- scripts/groceries-search-eval.ts prints top results + latency; preflight now 8 tables
dataholders schema (orgs + identity) with migration and personal_app grants,
idempotent CSV import (upsert on name+login, never wipes, keeps site edits and
ticks), owner-only /trackers/data-holders with search, filters, pagination,
inline edit/add and a moving checklist per identifier
- Range picker (?range=7d|30d|1y|all) drives every section; the hours chart
covers the whole range (daily for 7d/30d, weekly for 1y, monthly for all).
- Genres from Last.fm artist top tags (music.artist_tags, normalised and
filtered), split by artist listening time, with a per-bucket stacked chart.
- Listening clock, new artists, comparison with the previous period and top
movers, streaks, milestones, top albums with art, rank changes.
- music.play_facts materialized view (refreshed by the app when plays arrive)
plus a per-range cache keep the page fast.
- scripts/music-fetch-tags.ts backfills tags, most-played artists first.
