Two UX nits:
- Caveats badge had whiteSpace:nowrap on a 250-char data_caveat → overflowed
visibly outside the section in the screenshot. Drop nowrap, keep small
font + light grey background. Reads as a compact note.
- Headline KPIs (369 млн ₽ · 19.4 мес · темп 6.0 · ликв 97/100) are
computed by backend at price_factor=1.0, while live KPI cards below
recompute with the current slider position. When user moves price to
+40% the discrepancy is confusing. Append small grey '(базовая цена)'
marker after headline with a tooltip explaining baseline vs live.
The comparables block on /analytics/recommend was showing the same ЖК
multiple times (ЖК СТАРТ × 3, ЭХО ЛЕСА × 2 etc) because domrf_kn_objects
has ~3 historical snapshots per obj_id. The previous query joined all
rows and LIMIT 5 cherry-picked duplicates by flat_count.
Fix: wrap base in CTE `latest_obj` with `DISTINCT ON (obj_id) ORDER BY
obj_id, snapshot_date DESC` to pick only the latest snapshot per object
before sorting by flat_count.
Previously `docker image prune -af --filter until=72h` only removed
images older than 72h. With multiple deploys/day we accumulate many
fresh (<72h) tagged images — each gendesign-backend / worker / frontend
SHA-tag is 500MB-1GB, filling /var/lib/docker.
New logic: for each ghcr.io/lekss361/gendesign-* repo, list tags sorted
by Created DESC, skip :latest, keep top 2, remove the rest. Still runs
the standard prune after for dangling layers + build cache.
Root cause of "auto-resume never fired": the kn_scrape_runs resume section
hit `if not rows: return` (and similar `return` in except) before reaching
the nspd_geo_jobs resume section. Whenever there were no zombie kn-runs
(the normal case), the handler bailed out and geo jobs stayed forever
'running' with stale heartbeats — users had to manual cancel/resume after
every deploy.
Fix: don't return early. Initialize `ids = []`, only run UPDATE if rows
exist, drop the inner `return` from exception branch. The for-loop over
ids becomes a no-op when empty, and execution falls through to the geo
section. Same pattern as the breadcrumb above — fail soft, continue.
cleanup_zombies beat task (added in caa467f) stays as belt-and-suspenders
in case worker_ready signal ever misbehaves again.
worker_ready signal handler was NOT firing in our setup (verified via
DB breadcrumb after 3 deploys — zero rows of stage='worker_ready' in
nspd_geo_log). Root cause of unreliability unknown — possibly Celery
internals, possibly compose recreate timing. Either way, after every
redeploy users had to manually cancel/resume jobs to keep them moving.
Replace signal-based resume with periodic beat task:
- cleanup_zombies runs every minute (* * * * *)
- Finds nspd_geo_jobs in status running/paused with heartbeat >2 min stale
- Sets status='queued' + apply_async with queue=geo
- Idempotent — if no zombies, no-op
worker_ready handler kept (with FK-fix breadcrumb on NULL job_id) for
diagnostic purposes — if signal ever does fire, we'll have evidence.
The 'stale' badge in /admin/scrape/* tables was a UI-only label
computed as Date.now() - heartbeat_at > 60s (90s for geo). Heartbeats
naturally lag past 60s during WAF backoff (30s) or NSPD HTTP timeout
(15s) — every page render flashed running jobs as 'stale' even though
they were processing.
300s threshold matches the real auto-resume threshold in the worker
(jobs are zombies only after >5 min of silence). Less false-positive
churn in the UI.
Two related fixes:
1. worker_process_init handler disposes the SQLAlchemy engine in each
prefork child. Without this, child processes inherit open psycopg
sockets from the parent. First use in a child raises
ProgrammingError: can't change 'autocommit' now: connection in
transaction status INTRANS. This was killing 1 of every 5 parallel
geo jobs on cold start (job 13 in latest bulk run).
2. Add logger.info at start/end of worker_ready resume handler so we
can see in worker logs whether it actually fired and how many jobs
it resumed.
Auto-resume bug: kn and nspd_geo resume-on-worker_ready required
heartbeat >5min / >10min stale. After redeploy worker boots in
~1-2 min, so jobs killed seconds before deploy had fresh heartbeat
and were NEVER auto-resumed — required manual cancel + resume.
Fix: drop time threshold entirely. On worker_ready ANY 'running'
or 'paused' job is by definition a zombie (no worker exists yet),
safe to resume all of them.
Concurrency bump: 5 -> 8 prefork slots. Headroom for 5 geo jobs +
1 kn sweep + 2 objective tasks running simultaneously. Each slot
~150MB RSS -> ~1.2GB total, well within 6GB VPS RAM budget.
- ObjectBuilding type + useObjectBuildings hook
- Objects detail page: render Корпуса ЖК table (cad_num, floors, area,
purpose, name, address) when buildings_count > 0
- Developers page: drop hardcoded compareIds, use top-3 from
useTopDevelopers sorted by jk_count DESC (fallback to legacy static
array when API has no data yet)
Move JobSettings UI from /admin/jobs/settings into JobSettingsPanel component,
render it alongside new BulkGeoPanel on /admin/scrape/all. BulkGeoPanel has a
parallelism input (1-10, default 5) and POSTs /admin/scrape/geo/bulk. Remove
the standalone /admin/jobs/ route and the nav tab pointing to it.
- POST /api/v1/admin/scrape/geo/bulk — splits pending Sverdlovsk cad-nums
into N chunks (parallelism 1..10, default 5), creates N jobs and enqueues
each in queue=geo. source_kind='rosreestr_pending_chunk' for tracking.
- analytics_queries.complex_buildings(db, obj_id) — returns list of buildings
from cad_buildings (cad_num, floors, area, purpose, name, address, geom).
- object_detail: LEFT JOIN v_complex_buildings, adds buildings_count.
- top_developers: adds complexes_count via correlated subquery.
- GET /api/v1/analytics/object/{obj_id}/buildings → list[ComplexBuilding].
Add GET+PUT table for all job_types (scrape_kn, nspd_geo, objective_etl,
objective_sync). Per-row drafts, collapsible extra_config JSON editor,
toast notifications. Tab added to admin layout. Hand-rolled types until
codegen picks up new endpoints.
Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
- JobSetting ORM model (JSONB extra_config) + Pydantic schemas
- GET/PUT /api/v1/admin/jobs/settings + /{job_type} (X-Admin-Token auth)
- celery_app.py builds beat_schedule from job_settings DB (env fallback
retained for safety on first deploy / DB unreachable)
- nspd_geo task reads rate_ms from job_settings when per-job row has
no override
- enqueue/resume geo jobs route to queue_name from job_settings
- Worker container: --queues=celery,scrape_kn,geo (one container,
three named queues — kn sweep no longer blocks nspd_geo)
Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
Add `changes` job (dorny/paths-filter@v3) to detect which parts of the
codebase changed. Build jobs now run only when their source tree is
touched; deploy runs as long as no build failed (skipped builds are OK).
Split IMAGE_TAG (latest) from SENTRY_RELEASE_VAL (git sha) so Sentry
release tracking still points to the exact commit while compose pulls
the :latest tag.
Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
Split single build-and-deploy job into 4 jobs: build-backend,
build-worker, build-frontend run in parallel (all 3 independent),
deploy job has needs:[...] on all three. Removes ~2 min of sequential
build time on warm cache runs.
Also: scope GHA cache keys per image (backend-lean / worker-chromium /
frontend) so builds don't invalidate each other's layer cache. Replace
sleep 8 health-check with a 30-iteration polling loop (fails fast on
bad deploy, doesn't over-wait on fast ones). Add --mount=type=cache to
frontend npm ci for node_modules cache across builds.
_save_quarter использовал str(dict).replace("'", '"') как hack для GeoJSON
→ JSON. Это ломалось на Python booleans (True/False/None), которые в JSON
должны быть true/false/null. После dumps=False фикса rosreestr2coord
возвращает Python dict с булями в properties (is_actual=True) → INSERT
падал с psycopg.errors.InvalidTextRepresentation. _save_building уже
использовал json.dumps корректно.
Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
Three incremental bugs surfaced while wiring rosreestr2coord v5 into our
worker. All in backend/app/services/scrapers/nspd_lite.py.
1. `Area.__init__()` removed `delay` kwarg in v5 (was throttling). Drop
it from our call. Rate limit now happens in worker via
time.sleep(rate_ms/1000) after each fetch.
2. `Area(..., use_cache=True)` default writes media cache to ./tmp/
relative to CWD. In our docker image CWD=/app (owned by root), worker
runs as non-root → PermissionError on every target. Fix:
use_cache=False, media_path="/tmp/rosreestr2coord".
3. `Area.to_geojson_poly(dumps=True)` default returns JSON-serialized
STRING, not dict (changed in v5). Worker's _persist_target expects
dict with .get("properties") → AttributeError. Fix: dumps=False.
Vault entries:
- fixes/Bug_Rosreestr2coord_Delay_Arg_v5_Fixed.md
- fixes/Bug_Nspd_Geo_Tmp_Permission_Fixed.md
- fixes/Bug_Nspd_Geo_Str_Object_No_Get_Fixed.md
rosreestr2coord.Area() by default creates ./tmp/ media cache relative
to CWD. In our worker container CWD=/app (owned by root), runs as
non-root user `app` → PermissionError: '/app/tmp' on every target.
Fix: pass explicit kwargs to Area():
- use_cache=False (we don't need cache — each cad_num is unique)
- media_path='/tmp/rosreestr2coord' (world-writable in container)
Vault: fixes/Bug_Nspd_Geo_Tmp_Permission_Fixed.md
Объединяю несколько связанных изменений вокруг NSPD geo bulk-fetcher:
Adapt to rosreestr2coord v5 API:
- nspd_lite.fetch_via_rosreestr2coord: drop `delay` kwarg from Area()
(removed upstream in v5); keep it in our function signature for
backward-compat, comment why.
- nspd_geo worker: add explicit time.sleep(rate_ms/1000) after lib-branch
fetch — in v4 the library throttled internally via delay, in v5
rate limiting is the caller's job. Без этого получали Area.__init__()
unexpected kwarg `delay` на каждом target.
Drop use_rosreestr2coord switch:
- Removed urllib-vs-lib choice everywhere. We always use community
rosreestr2coord library — авторы регулярно обновляют WAF-tricks,
наш urllib-fetcher (fetch_geoportal) уже неактуален.
- admin_scrape.py: Pydantic schema, INSERT, SELECT, API response
cleaned of `use_rosreestr2coord`.
- nspd_geo.enqueue_geo_job: param dropped, INSERT shrunk.
- worker process loop: dropped `if use_lib:` branch + import of
fetch_geoportal.
- frontend/geo/page.tsx: removed checkbox + GeoJob.use_rosreestr2coord
field + POST body field.
DB column drop:
- data/sql/78_drop_use_rosreestr2coord.sql (NEW):
DROP COLUMN nspd_geo_jobs.use_rosreestr2coord + CREATE OR REPLACE
VIEW v_scrape_runs_unified (которая depended on the column).
- data/sql/77_nspd_geo_jobs.sql: cleaned historical DDL for fresh setups.
- Migration applied to prod (in-conversation via postgres MCP).
Frontend polish:
- Thematic ID changed from free-form number input to labeled select
(1=parcel / 2=quarter / 4=admin / 5=building / 7=zone / 15=complex).
- Auto-sync thematic_id from Job kind on change (override possible).
- ScrapeLogsPanel: extended union type with "nspd_geo" + fixed
/admin/scrape/geo to pass scraperType="nspd_geo" (was "nspd",
filtering empty legacy nspd_scrape_log table; real logs live in
nspd_geo_log via v_scrape_log_unified).
Verified: ruff ✓, tsc --noEmit ✓, migration ran (BEGIN..COMMIT clean).
Deploy order safe: prod column уже удалена → новый backend код, который
не INSERT'ит use_rosreestr2coord, совпадёт со схемой после deploy.
Two fixes for the NSPD geo bulk-fetcher:
1. rosreestr2coord was listed in pyproject.toml but missing from uv.lock,
so Docker's `uv sync --frozen` didn't install it. NSPD jobs failed every
target with `No module named 'rosreestr2coord'`. Ran `uv lock` to add it
(v5.3.3) plus requests + charset-normalizer transitive deps.
2. /admin/scrape/geo and /admin/scrape/objective lacked the log panel that
/admin/scrape (DomRF) had. Extracted ScrapeLogsPanel component over the
unified /api/v1/admin/scrape/all/logs endpoint with scraper_type filter
and wired into both pages.
Tests: ruff ✓, tsc --noEmit ✓
Worker крашился на старте с `ModuleNotFoundError: No module named 'psycopg2'`
при импорте app/workers/tasks/objective_etl, потому что app/services/objective_etl
писал на psycopg2 API, а в pyproject.toml есть только psycopg[binary]>=3.2.0.
Изменения:
- psycopg2 → psycopg (v3)
- psycopg2.extras.execute_values → _bulk_upsert helper, эмулирующий тот же
behavior через cur.executemany (в psycopg v3 это pipeline-mode, overhead
на сетевые round-trips минимален)
- psycopg.connect совместим с тем же connection string
Тесты: AST parse ✓, ruff ✓, import ✓
Obsidian Self-hosted LiveSync (CouchDB) выведен в отдельный docker-compose
stack и GHA workflow. Любой стек теперь редеплоится независимо.
* docker-compose.obsidian.yml: только couchdb в shared external network
* docker-compose.prod.yml: убрана couchdb-секция; caddy подключён к
shared network для маршрута obsidian.gendsgn.ru → couchdb:5984
* .github/workflows/deploy-obsidian.yml: новый workflow с path-filters
(триггерится только при изменениях obsidian-related файлов)
* .github/workflows/deploy.yml: + path-filters (триггерится только при
main-related файлах) + создание shared network в bootstrap
* docs/obsidian-livesync.md: обновлены инструкции под split-архитектуру
После push:
1. ssh gendesign 'docker network create gendesign_shared' (один раз)
2. Прописать COUCHDB_USER/PASSWORD в backend/.env.runtime
3. Добавить DNS A-record obsidian.gendsgn.ru → IP VPS
4. git push → GHA задеплоит обе части независимо
Преимущества:
- Изменения main не разрывают LiveSync клиентов
- Изменения obsidian не рестартуют main
- Падение obsidian-стека не влияет на основное приложение
ОТКРЫТИЕ: stdlib urllib проходит WAF nspd.gov.ru (TLS-fingerprint stdlib
отличается от mainstream HTTP-clients). Это позволяет полностью убрать
Playwright + Chromium для NSPD-задач.
* nspd_lite.py: urllib + ssl._create_unverified_context(). 4 публичных
fetcher'а (geoportal/quarter/parcel/building) + fetch_via_rosreestr2coord
fallback на community-lib
* schema 77: nspd_geo_jobs (журнал) + nspd_geo_targets (cad-номера со
статусом pending/done/failed). Resume-state в БД.
* tasks/nspd_geo.py: Celery task process_nspd_geo_job(job_id). Heartbeat
каждые 5 items, WAF backoff 30s × 2^N (max 8 → paused). UPSERT в
cad_quarters_geom / cad_buildings → идемпотентно.
* worker_ready hook: stale jobs (>10min без heartbeat) автоматически
re-enqueue после redeploy. Никаких потерь прогресса.
* 4 admin endpoint + UI /admin/scrape/geo с формой запуска (manual_list /
rosreestr_pending для авто-наполнения из ДДУ-кварталов), progress-bars
и cancel/resume кнопками per job.
* settings: use_nspd_lite=True, nspd_lite_rate_ms=600.
* dep: rosreestr2coord>=4.0.0 (community lib для fallback).
TODO следующих шагов:
- smoke-test на проде через SSH (validation RU-IP не банит urllib)
- feature toggle USE_NSPD_LITE в scrape_nspd.py для переключения с старого
Playwright-пути
- docker lean image (-300 MB после убирания Chromium)
- расширение SCRAPE_NSPD_DEFAULT_REGIONS на 66,74,72,59 (Челябинск/Тюмень/Пермь)
Objective API (api.objctv.ru) интегрирован как новый source of truth для
per-flat данных по новостройкам УрФО. Заменяет промежуточный Anton-SQLite
(legacy bootstrap-ETL остался как fallback в свёрнутом блоке UI).
Schema (data/sql/68_v2 — applied на проде, 6 таблиц):
- objective_lots (UPSERT по lot_id; 303 677 rows)
- objective_corpus_room_month (long-формат месяц×корпус×room_bucket; 19 738 rows)
- objective_lots_history (append-only weekly snapshots для elasticity)
- objective_complex_mapping (Objective ComplexName ↔ domrf_kn_objects.obj_id)
- objective_raw_reports (jsonb страховка на смену схемы API)
- objective_scrape_runs (журнал прогонов)
+ data/sql/72_objective_sync_config (single-row динамический конфиг)
Backend:
- services/scrapers/objective.py: ObjectiveClient — Bearer-токен (Redis +
in-memory fallback), retry на 401/429/5xx, Retry-After header support
- services/objective_etl.py: ETL SQLite Антона → PG (legacy)
- services/objective_sync_config.py: read/update single-row config
- workers/tasks/scrape_objective.py:
* sync_objective_group: 2 рабочих отчёта (corp_sum, lots_pf), inline-парсинг
* sync_all_groups: wrapper, перебирает группы из БД-конфига с
inter-group паузой; PATCH-merge explicit args > DB config
- workers/tasks/objective_etl.py: Celery task для legacy bootstrap
- workers/celery_app.py: beat читает cron из БД при старте (fallback на env)
- api/v1/admin_scrape.py: 5 новых endpoints для /objective/*
Frontend (frontend/src/app/admin/scrape/objective/page.tsx):
- PRIMARY blue блок «🌐 Наш sync» с input для override групп
- Collapsible «⚙️ Настройки» с формой (cron + 8 параметров) → PUT в БД
- Coverage-панель с PG counts + строка про SQLite Антона как legacy
- Collapsible «🛠 Bootstrap ETL» — legacy-инструмент
Beat schedule: вторник 06:00 МСК, ~10-15 мин на 4 группы (Свердл.обл +
Челябинск + Тюмень + Пермь = ~700K квартир УрФО). Расписание и параметры
меняются через админку без редеплоя (cron требует restart beat).
Эмпирические находки об API (probe 2026-05-10):
- 13 из 21 проверенных group_name доступны на тарифе (включая «Свердловская
область», «Челябинск», «Тюмень», «Пермь», но НЕ «Свердловская обл» —
формат имени строгий)
- ComplexName требует БЕЗ префикса «ЖК» и БЕЗ кавычек
- Поле «Банк» для всех 303k = NULL (тариф не отдаёт), но «Тип обременения»
работает (36% строк = ипотека) → ipoteka_share возможен, банковская
атрибуция — нет
Docker:
- bind-mount /opt/gendesign/site-finder:/data/anton-sqlite:ro в worker
GitHub backlog: добавлены #22-25 (recommend_mix v3 — 4 сабтаска по
улучшению алгоритма на основе фидбэка про POI / границы районов /
конкурентов / окно данных / success-driven mix).
Knowledge graph (memory/memory-gendesign.jsonl): обновлены entities
Objective_Integration_May07_2026, Schema_Objective_v2_May07,
Objective_API_Findings_May07, Module_Objective_Client + новый
Session_End_May07_2026.
TODO для прода: прописать OBJECTIVE_API_KEY=<key> в backend/.env +
docker compose restart worker beat.
В nspd_kn.py:get_pending_cads и admin_scrape.py:nspd_coverage regex-фильтры
quarter_cad_number !~ '^00:00:' и !~ ':0000000$' падали с psycopg.errors.
DataError — двоеточие в text() трактуется как ':param'. Заменил на NOT LIKE
с bindparams (быстрее regex и без issue с парсингом).
Эффект на проде:
- /api/v1/admin/scrape/nspd/coverage возвращал 500 → теперь 200
- Кнопка «Запустить sweep» больше не дизейбленa (pending=964 теперь видно)
- Сам scraper тоже бы валился на этом же баге при beat-запуске
