Commit Graph

178 Commits

Author SHA1 Message Date
wasrusgen
efa2046a97 gitignore: cat_refresh.json 2026-05-12 07:10:06 +03:00
wasrusgen
fe472b0827 catalog: filter junk + background refresh + clear endpoint
FILTERING (catalog.py _save_results):
- CATEGORY_KEYWORDS: must contain category word ('холодильник', 'варочн', 'духов', etc.)
- CATEGORY_MIN_PRICE: filters parts/accessories (fridge >20k, hood >5k, etc.)
- PART_BLACKLIST: 'фильтр', 'лампочк', 'термодатчик', 'шланг', 'тэн', 'компрессор', etc.
- Previously had Asko light bulb (155₽), Miele dryer filter (376₽), Siemens cooktop in fridge category — all now filtered out

ASYNC REFRESH (main.py):
- POST /api/catalog/refresh queues background task, returns immediately
  (was sync, taking 3+ min → Cloudflare tunnel was killing connection)
- New GET /api/catalog/refresh_status for progress polling
- Concurrent refresh blocked (one at a time)

CLEAR ENDPOINT:
- POST /api/catalog/clear?cat=fridge clears one category
- POST /api/catalog/clear clears entire catalog (start over)

NEXT: clear current dirty data, re-seed fridge with filters
2026-05-12 07:09:33 +03:00
wasrusgen
9e652c4a34 catalog: models cache in Sheets — AI picks from real list, no SKU hallucination
NEW MODULE app/catalog.py:
- refresh_catalog(cats, sources, per_brand, delay) — runs parsers for seed brand+category pairs
- list_catalog(cat, tier, brand) — reads from Sheets
- list_for_ai(cats, tiers) — compact text for AI prompt context
- SEED_BRANDS_BY_TIER + CATEGORY_QUERIES — 22 brands × 8 cats = 176 combos
- Saves top-2 relevant results per (brand × cat), filters by brand presence in title
- Dedup by title hash within (cat, brand) bucket

SHEETS:
- ensure_sheet(name, headers) — auto-creates Catalog tab on first refresh
- Schema: id, category, brand, tier, model_name, search_query, price_min/max, image_url, source, url, last_seen_at

ENDPOINTS:
- POST /api/catalog/refresh?cat=X&per_brand=N — manual refresh (1 cat ~2-5 min)
- GET /api/catalog/list?cat=&tier=&brand= — read with filters
- GET /api/catalog/preview_ai?cats=fridge — debug what AI receives

AI PROMPT:
- Rule #0: if catalog passed in user prompt — MUST select only from there
- _build_catalog_context: filters by checklist.budget_preset → tier subset
  (luxe→premium, premium→premium+middle, middle→middle, budget→middle+budget)

_handle_podbor:
- Loads catalog subset, appends to user_prompt as 'ДОСТУПНЫЙ КАТАЛОГ МОДЕЛЕЙ'
- AI 'выбирай ТОЛЬКО из этого списка' rule reinforced

NEXT: trigger refresh manually for 1 category (~3 min), then real podbor test
to verify AI uses catalog models instead of hallucinating SKUs
2026-05-12 06:32:39 +03:00
wasrusgen
1a57374020 parsers: better image extraction — real product photos in report cards
CITILINK:
- Now reads data-src / data-original / srcset / src in priority order
- srcset → picks largest size variant (last in comma-list)
- Filters only _next/static/images (placeholder) and 'placeholder' in URL
- Accepts cs.citilink.ru / c.citilink.ru / images.citilink.ru product photos

ЯНДЕКС.МАРКЕТ:
- Collects all img attrs (data-src, data-original, srcset, data-srcset, src)
- Prefers avatars.mds.yandex.net (real product CDN), skips yastatic (icons/logos)
- Auto-appends /300x300 suffix to avatars.mds URLs without size

ENRICH_ONE (aggregator):
- Image picked by source priority: yamarket > wb > ozon > citilink > dns
- Yamarket photos are cleanest (avatars.mds.yandex.net)
- WB has product photos via basket-XX.wbbasket.ru
2026-05-11 23:43:25 +03:00
wasrusgen
0b48dd2371 simplify: remove 'features' step from all 8 categories
USER FEEDBACK:
'Особенности везде убрать, их можно в SWOT анализе приводить в качестве примечания
не акцентируя на них особого внимания. Современные фичи на 95% одинаковые.'

REMOVED features step from:
- fridge (NoFrost, Inverter, Wi-Fi, etc.)
- hob (Booster, FlexZone, FFD, Hob2Hood, etc.)
- oven (Wi-Fi, autoprogram, probe, softclose, etc.)
- dw (Wi-Fi, AutoOpen, AutoDose, AquaStop, Inverter)
- hood (touch, LED, auto, silent, turbo, wifi, perimeter)
- microwave (Wi-Fi, humid sensor, defrost, antibac)
- coffee (Wi-Fi, touch, grinder, autoclean)
- washer (inverter, steam, wifi, autodose, silent, aquastop)

KEPT: hood.color (about visible material/aesthetics, not feature)
KEPT: oven.location (where in kitchen — design-relevant)

NEW STEP COUNTS:
- fridge: 3 (was 4)
- hob: 4 (was 5)
- oven: 3-4 (was 4-5)
- dw: 3 (was 4)
- hood: 3-4 (was 4-5)
- microwave: 3 (was 4)
- coffee: 1-4 (was 2-5)
- washer: 5 (was 6)

AI PROMPT updated:
- Features no longer come from user — AI mentions important ones in highlights/pros
- Emphasis on MEASURABLE advantages in pros (N dB quieter, Y l more, N% cheaper)
- Не делать акцент на стандартных фичах — 95% одинаковые

USER WIZARD теперь короче и проще: тип → размер → ключевые параметры → готово
2026-05-11 23:37:41 +03:00
wasrusgen
cecb8d3444 review screen: fix text wrap + cleanup stale answers from removed steps
CSS:
- .rev-val: flex:1, min-width:0, overflow-wrap:break-word — длинные значения
  больше не ломаются мид-словом ('энергоэффективнос·...')
- .rev-label: max-width:40% — лейбл не съедает всё место
- hyphens:auto для перенос длинных слов на дефис

JS (getCatState):
- При загрузке per_cat фильтруем answers — оставляем только ключи которые
  есть в текущем config.steps
- Это убирает stale-поля типа 'class' у ПММ, оставшиеся в localStorage
  после рефакторинга шагов
- Безопасно: меняет только в памяти, не перезаписывает state (renderReview
  всё равно итерирует config.steps)
2026-05-11 23:25:23 +03:00
wasrusgen
6915bba845 user feedback: oven proportions + dw simplification
OVEN PICTOGRAMS (per user: 'духовка не очень похожа, прямоугольные, фасадом не закрываются'):
- oven_install_builtin: REMOVED dashed niche outline (ovens don't close with façade — sit in open cabinet)
- Made body wider+shorter — 78×74 viewBox area (was 68×112, too tall)
- Real 60×60 cm proportions, control panel at top + handle + glass window with racks
- oven_install_stove: тщательнее прорисован — cooktop (with concentric burners), control strip,
  oven door with handle + window, ножки чётче, линия пола

DW LOGIC SIMPLIFIED (per user: 'энергопотребление уже перебор'):
4 шага вместо 5:
1. Тип встройки (full/partial/freestanding) — was step 1, kept
2. Размер ширина (45/60) — was step 3, moved up to step 2
3. Корзины + программы — merged in one step:
   - 2 корзины · базовый (5-6 программ)
   - 3 корзины · стандарт  (8-10 программ)
   - 3 корзины · расширенный (12+ программ, стекло, авто, кастрюли)
4. Особенности (multi) — теперь содержит Wi-Fi, AutoOpen, AutoDose, Beam, AquaStop, ≤44dB,
   Inverter (включая A+++), GlassZone

Removed: separate 'class' step (energy efficiency moved into features as Inverter option)
2026-05-11 23:06:25 +03:00
wasrusgen
03c95fe13a wb: relevance filter — discard anti-bot trash products (платья/обувь in fridge search)
WB sometimes responds with 1-2 unrelated products instead of 429 status.
Was returning 'Платье вечернее' on 'Haier холодильник' query.

Fix: _is_relevant(product, query) checks that at least 1 significant query word (>=3 chars)
appears in product name or brand. Discards full result if zero matches.

Tradeoff: may sometimes reject valid product if query is overly specific (e.g. exact SKU).
But that's OK — we fall through to next query variant.
2026-05-11 23:02:37 +03:00
wasrusgen
d84a53f91d wb: skip proxy pool (use direct VPS-IP — residential were rate-limited) 2026-05-11 23:01:34 +03:00
wasrusgen
e9b0db6772 wb: API v9 → v18 (WB сменил endpoint и структуру) + brand+category fallback query
DISCOVERED in real test:
- WB API v9 (/exactmatch/ru/common/v9/search) теперь возвращает только метаданные
  (name, query, shardKey, filters, search_result={}) — products пусто
- WB API v18 (/exactmatch/ru/common/v18/search) — рабочий
  Структура: {metadata, products, total} — products НА ВЕРХНЕМ уровне (не data.products)
- Подтверждено: query='Haier холодильник' → 100 products via v18

CHANGES:
1. _SEARCH_URL → v18 endpoint
2. Парсинг products: сначала data.products (legacy fallback), потом products top-level
3. _build_item: цены теперь читаются из sizes[].price.{product, total, basic}
   (v18 формат), с fallback на priceU/salePriceU (v9 legacy)
4. _generate_query_variants: добавлен brand+category fallback
   ('Bosch холодильник' если не нашли по модели)

TEST: Haier холодильник → 100 results (first: 'Холодильник двухкамерный C2F619CFU1')
2026-05-11 22:59:14 +03:00
wasrusgen
555c5568ff miniapp: 25 SVG pictograms for 7 remaining categories (style D · 3D)
PODBOR_PICTS additions (25 total):
- HOB (3): elec — induction concentric circles, gas — burners with grid + knobs, combi — split panel
- OVEN (2): built-in with niche + control panel + glass window, stove (combo unit on legs)
- DW (3): full (hidden facade with handle strip), partial (control bar on top), freestanding (full controls + door + feet)
- HOOD (7): drawer (cabinet + sliding panel), hidden (cabinet only), dome (chimney shape), inclined (angled glass), island (ceiling tubes), downdraft (counter panel rising), hob-in-combo (cooktop with center exhaust slot)
- MICROWAVE (2): built-in (in niche with window+keypad), freestanding (countertop with feet)
- COFFEE (5): built-in (display + buttons + spout + cup), free-grinder (bean hopper + display), capsule (small + capsule slot), manual (with portafilter + steam wand + pressure gauge), tap (faucet integrated into countertop)
- WASHER (3): built-in (hidden facade), under-top (control panel visible, big door), freestanding (full unit + feet + powder tray)

CONFIG wiring: all 25 pict keys referenced in podbor.config.js step 1 of each category

PREVIEW: new preview-all-picts.html shows all 25 in one page for visual review
2026-05-11 21:27:40 +03:00
wasrusgen
da8a98f34f market 2026 update: Kuppersberg budget default, Haier mid, full RF brand realism
PODBOR_SINGLE_BRAND_OPTIONS (single-brand kitchen picker):
- Premium: + Gaggenau ⚠, V-Zug ⚠, Liebherr ⚠ (all parallel-import)
- Middle: + NEFF ⚠, Haier  marked recommended
- Budget: + Kuppersberg  recommended, Maunfeld, Weissgauff, Gorenje, Hotpoint, Indesit, Midea, Candy
- Removed budget-only Бирюса/Pozis/DEXP (not popular in built-in segment)

PODBOR_BRANDS per category — fully refreshed with realistic 2026 lists:
- Coffee: Bosch/Siemens/NEFF in mid, Kuppersberg/Maunfeld in budget (instead of obscure ones)
- All categories now include Kuppersberg/Maunfeld/Weissgauff in budget tier
- Premium adds Gaggenau, V-Zug consistently

AI PROMPT — new section 'РЫНОК РФ 2026':
- Documents exact tier composition with brands and price ranges
- 'Типичный выбор клиента ЗОВ-СПб: Bosch + Haier + Maunfeld'
- Premium combo: Bosch + Miele washer (для кухонь 600к+)
- Trends: parallel import normalized, Haier #2 after Bosch, Kuppersberg builder default
- СВЧ category fading — combined ovens with microwave taking over
- Induction wins, gas only in private houses

EXAMPLES в prompt expanded:
- Haier C4F744CMG, Kuppersberg NRS 1857 X, Maunfeld MBL 88LU, Weissgauff WBI 30 ATX
- Clear 'НЕ выдуманное' guidance
2026-05-11 20:23:45 +03:00
wasrusgen
c97b8dce3c parsers: skip sponsored/ad URLs (cpc/sponsored=1) — they expire in 2-3 hours
User reported clicking matrix prices led to 'Произошла ошибка!' on OZON home page.
Cause: parsers captured /product/?sponsored=1&cpc=Jtiito95... links that died after few hours.

Fix:
- ozon.py: skip href with 'sponsored=1', '/promo/', 'cpc='. Strip query string from final URL.
- yamarket.py: skip 'sponsored=1', 'cpc=', 'advUuid' (Я.Маркет sponsored marker)
- citilink.py: strip query string from final URL (defensive)

Now matrix links go to canonical product pages that don't expire.
2026-05-11 17:20:59 +03:00
wasrusgen
ef500fa446 user feedback batch: model count, specs, manual link, dimensions, export
1. MODEL COUNT SELECTOR (strategy step):
   - new PODBOR_MODEL_COUNTS [3/5/7]
   - state.model_count default '5'
   - UI on strategy page with description (быстро/оптимально/максимум)

2. AI PROMPT EXPANDED:
   - new field: manual_search_query — for Google search instruction PDF
   - new specs object per model: dimensions_mm/volume_l/weight_kg/noise_db/energy_class/color
   - 'specs ОБЯЗАТЕЛЬНЫ для проектирования кухни' explicit rule
   - reads checklist.model_count to determine how many models per category
   - max_tokens 4000 → 8000 (room for richer responses)

3. MODEL CARD RICHER:
   - _renderSpecsBlock — characteristics in 2-col grid, dimensions highlighted
   - _renderUtilityLinks — Google search buttons for инструкция (PDF) + Схема установки
   - Specs critical for ZOV kitchen design (manager needs to verify niche fits)

4. EXPORT BUTTONS:
   - 'Скачать HTML' — generates standalone HTML with inline styles, downloads as file
   - 'Печать → PDF' — opens new window with cleaned layout + auto-prints
   - User can save as PDF via system print dialog

5. PREVIEW updated with realistic specs/manual_query for all 3 fridges
2026-05-11 17:11:30 +03:00
wasrusgen
7f417da7e0 gitignore: wb.json test artifact 2026-05-11 16:49:02 +03:00
wasrusgen
5ceffa4f69 miniapp: phone validation on intro — blocks transition with bad number
- New isValidPhone(raw): checks 11-digit Russian after normalization (8/7/+7/9-prefix)
- Intro 'Начать' button now custom click handler instead of data-go
- Validates name (non-empty) and phone (Russian format)
- Inline .field-error red message under invalid field
- .field-hint shows format help under phone input
- Haptic 'warning' feedback on invalid submit
- Phone is auto-normalized to '+7 900 123-45-67' before transition
2026-05-11 16:48:52 +03:00
wasrusgen
0f2635d5f8 dns+ozon: 4 retries with proxy rotation (residential pool has dirty IPs) 2026-05-11 16:37:28 +03:00
wasrusgen
796e20bc73 gitignore: r.json (test artifacts) 2026-05-11 16:34:51 +03:00
wasrusgen
aa569a8ed1 dns: switch to Playwright (Qrator JS challenge); ozon: fix false-positive antibot detector
- DNS: использовали httpx + proxy_pool но Qrator кидал 401 даже с residential
  → теперь Playwright + residential — браузер сам решает JS challenge
- OZON: теперь проверяем только <title>='Доступ ограничен' (точная), а не подстроку '/robotcheck/'
2026-05-11 16:34:04 +03:00
wasrusgen
b27cf02aa2 yamarket: clean React JSON noise + extract title from URL slug
Я.Маркет рендерит SnippetConstructor виджет с JSON-стейтом ВНУТРИ a-тега.
Поэтому link.get_text() возвращает мусор типа {'widgets':{...}}.

Фикс:
- copy.copy(card) и удаление <script>/<noscript>/<noframes>/<template>
- Title теперь берётся из URL slug первым приоритетом (всегда чистый)
- _slug_to_title: транслитерация и капитализация
  'bosch-kgn39ul30u-dvukhkamernyy-kholodilnik-no-frost-seryy-metallik' →
  'Bosch KGN39UL30U Двухкамерный Холодильник NoFrost Серый Металлик'
2026-05-11 16:30:34 +03:00
wasrusgen
839e775151 yamarket: rewrite for /card/{slug}/{id} URL pattern (Я.Маркет 2026)
- Old /product--{id} URLs deprecated
- Walks up from a[href*='/card/'] to nearest article/zone-div
- Extracts title from link text or h2/h3/itemprop=name
- Price: min from card text (with sanity bounds 100..10M)
- Image filters yastatic / _next placeholders
- Rating: '4.7★' or '4.7 N оценок' pattern
- Reviews: 'N отзывов' / 'N оценок'
- Stores count: 'от N магазинов / предложений'
2026-05-11 16:26:28 +03:00
wasrusgen
b1d8f3e38a gitignore: exclude .tmp_* files (accidentally committed proxy creds in previous commit)
User needs to rotate Proxys.io password after migration tests are done.
2026-05-11 16:06:05 +03:00
wasrusgen
e7f6e64e38 playwright_engine: route through proxy_pool — random residential IP per request
- New use_proxy param (default True)
- Per-request random proxy from pool
- _parse_proxy_url_for_playwright converts http://user:pass@host:port to playwright.proxy dict
2026-05-11 16:05:36 +03:00
wasrusgen
811bed31a4 backend: proxy_pool supports PROXY_LIST_FILE + format auto-conversion
- New env: PROXY_LIST_FILE — path to file with one proxy per line
- _normalize_proxy_entry accepts: http://user:pass@host:port, host:port:user:pass (Proxys.io format), host:port
- _load_from_file reads file, dedup with static list
- /api/proxy_status returns file_path, file_loaded count, sample (first 3 masked)
2026-05-11 15:52:02 +03:00
wasrusgen
d7be644aed miniapp: price comparison matrix as PRIMARY view per category
WHAT CHANGED:
- New _renderPriceMatrix(models) — table with rows=models, columns=stores
- Inserted as PRIMARY view above model cards (was secondary accordion)
- Columns dynamically include only stores that returned data
- Sticky model column (left) — scrolls horizontally on mobile
- Best price per row highlighted: green bg + ✓ badge + green text
- Empty cells: '—' if no URL, 'смотреть →' if URL but no price yet
- 'Мин' column on far right — explicit cheapest price summary

CSS:
- .report-matrix-wrap with rounded card
- Sticky col-model with box-shadow on right edge
- Cell-price.best with rgba green background
- .best-mark circle badge

PREVIEW:
- Updated mock with 3 fridges + 3 hobs across multiple stores (real pricing spread)
- Demonstrates min-price highlighting working

UX:
- User can now visually compare 'where is it cheapest' at a glance
- Tap any cell with price → opens store page
- Tap empty cell with URL → opens search in store

NEXT: same matrix can become PDF/Excel export for client briefcase
2026-05-11 14:56:41 +03:00
wasrusgen
ca342c0641 ai+report: deeper analysis — required pros/cons, category insights, source visibility
AI PROMPT (ai.py):
- Requires minimum 3 pros + 2 cons per model with NUMBERS (36 dB, 463 L, A++, не 'тихий/большой')
- New field 'reasoning' — 1-sentence why-this-model justification
- New per-category 'analysis' — 2-3 sentences about trade-offs
- Strict rules: no fake article numbers, account for parallel-import price markup
- Russian market 2026 awareness: Haier/Korting up, Bosch/Siemens ⚠

TELEGRAM FORMAT (main.py):
- Renders category analysis as italic prelude
- Lists pros/cons as bullet lists (up to 4 pros, 3 cons)
- Shows '🛒 Нашли в: OZON · Citilink · WB' line listing successful sources
- Rating + reviews + stores count line: '📊 ★ 4.7 · 1242 отзыв. · 12 магаз.'
- Direct link to best store: '🔗 Открыть в магазине'

WB PARSER:
- Generates 3 query variants per request: full → brand+model → model only
- Increases hit rate when AI search_query is too verbose
- First non-empty variant wins

MINIAPP REPORT (podbor.js + podbor.css):
- Category analysis block above models (italic, walnut left-border)
- Pros block: green tinted bg, bullet list, header 'Плюсы'
- Cons block: terracotta tinted bg, bullet list, header 'Минусы'
- Reasoning chip: 💡 italic in warm background
- Source badges with per-store price '<store> · 89 990 ₽'
- Color-coded source links: OZON blue, Citilink yellow, WB pink, Я.Маркет red, DNS orange
- 'X магазинов нашли товар' header + plural fix
- '— не найден' fallback if 0 sources

PREVIEW (preview-report.html):
- Mock updated with Haier as flagship (more relevant for 2026 RF)
- Shows analysis, reasoning, source spread (4 stores with different prices)
2026-05-11 14:34:08 +03:00
wasrusgen
4b04f2de54 miniapp: summary page hides Подключение/Вентиляция if hob/hood not picked 2026-05-11 14:26:12 +03:00
wasrusgen
80580db446 miniapp: 4 UX fixes from user feedback
1. PHONE NORMALIZATION
   - On blur (or before submit): '9001234567' -> '+7 900 123-45-67'
   - Handles 8XXX, 7XXX, +7XXX, 10-digit mobile prefixes
   - Leaves untouched if not Russian-looking number

2. BRAND LIST FOR RF 2026
   - PODBOR_SINGLE_BRAND_OPTIONS updated with realistic 2026 brands
   - Promoted: Haier, Korting, Midea, Hisense, Бирюса, Атлант, Pozis, DEXP
   - Bosch/Siemens marked with ⚠ (parallel-import)
   - Miele/Liebherr/Smeg also marked ⚠
   - PODBOR_BRANDS per-category fully refreshed

3. BUDGET ADAPTIVE HINTS
   - Hints now scale by selected categories share of full kitchen
   - Just fridge picked → 'Средний' shows ~88-175 тыс instead of 350-700к
   - Full 8 categories → original 350-700к
   - PODBOR_BUDGET_SHARES + PODBOR_BUDGET_RANGES constants

4. INFRA STEP CONDITIONAL
   - Stove power question only shown if hob category picked
   - Vent question only shown if hood category picked
   - If neither → step auto-skips to summary (with brief notice)
   - Summary 'Назад' button respects skip — goes to strategy if needed
2026-05-11 14:25:25 +03:00
wasrusgen
44281b1e07 citilink: dedup by product ID + filter Next.js placeholder images 2026-05-11 13:59:07 +03:00
wasrusgen
c5f662f53d citilink: rewrite parser to walk up from a[href*=/product/] (CSS-in-JS resistant) 2026-05-11 13:57:18 +03:00
wasrusgen
1a948ebf02 ozon: fix false-positive challenge detector (was catching 'challenge' in normal JS) 2026-05-11 13:54:13 +03:00
wasrusgen
e8b487891f backend: working parsers — OZON + Citilink (DOM via Playwright) + WB
DIAGNOSTIC RESULTS:
- OZON: 19 product links via Playwright on naked VPS-IP ✓
- Citilink: 112 data-meta-name Snippets ✓
- Wildberries: JSON API works with delays ✓
- Я.Маркет, DNS: blocked by ASN (need residential proxy)

OZON PARSER:
- Pure Playwright DOM (composer-api dropped — was blocked)
- Selects a[href*='/product/'], walks up to card div, extracts title/price/img
- Filters fake 'titles' like Распродажа, Скидка

CITILINK PARSER (new):
- Selects [data-meta-name*='Snippet'] or ProductCard markers
- Multiple title selectors fallback chain
- Filters out non-product hits

PARSERS/__init__.py:
- DEFAULT_SOURCES = (ozon, citilink, wb) — all work without proxy
- Я.Маркет, DNS kept but not default — usable when residential proxy added

NEW ENDPOINT:
- GET /api/parse_citilink?q=...&limit=N
2026-05-11 13:53:07 +03:00
wasrusgen
5fdae262ef backend: parse_* endpoints sync (FastAPI threadpool) — fix Playwright asyncio conflict 2026-05-11 13:30:51 +03:00
wasrusgen
d5f290bd0a backend: Playwright + Chromium for JS-rendered sites (Я.Маркет, OZON fallback)
DOCKERFILE:
- + Chromium system deps (libnss3, libxkbcommon0, libgbm1, libgtk-3-0, etc.)
- + RUN python -m playwright install chromium (~150MB)
- + ENV PLAYWRIGHT_BROWSERS_PATH

REQUIREMENTS:
- + playwright >= 1.45

PARSERS:
- new playwright_engine.py — singleton browser, isolated context per request,
  blocks images/fonts/CSS to save memory, waits for selector + JS hydration
- yamarket.py — rewritten to use Playwright (Я.Маркет is React SPA)
- ozon.py — Playwright fallback when composer-api returns challenge (403)
- wb.py — exponential backoff on 429, still uses direct HTTP (JSON API, no JS needed)

STRATEGY (Hybrid Path C):
- Я.Маркет: Playwright (rendering JS)
- OZON: composer-api first, Playwright fallback
- WB: direct HTTP with backoff (JSON API, fast)
- DNS: kept but lower priority (Qrator hard to crack)
- No more proxy needed for primary path

DEPLOY: removed PROXY_STATIC_LIST from .env, expect ~5min for first build (Chromium download)
2026-05-11 13:25:05 +03:00
wasrusgen
3ee5275ea0 backend: PROXY_STATIC_LIST support (manual proxies without API token)
- proxy_pool now loads from both PROXY_STATIC_LIST (env, comma-separated) and PROXY6_TOKEN (API)
- Static list has priority, merged with API list (dedup by URL)
- /api/proxy_status returns masked proxy URLs for diagnostic (passwords hidden)
- Supports formats: 'http://user:pass@host:port' or 'host:port' (assumed http://)
2026-05-11 13:03:29 +03:00
wasrusgen
c2be5e846f miniapp: inline report after submit + standalone preview-report.html
REPORT RENDERER (podbor.js):
- New renderReport(ai, leadId) function — beautiful inline report after submit success
- Shows by_category with up to 5 models per category
- Model card: photo (88x88), brand · name, price range, rating + reviews + stores
- Highlights (with tech translations), pros (green), cons (orange)
- External links to WB / Я.Маркет / OZON / DNS (when enriched data present)
- Comparison table per category (accordion details)
- Total price block (dark theme contrast)
- Warnings block (when AI returns concerns)

CSS (podbor.css):
- .report-* classes: head, summary, cat, model, links, compare, total, warnings
- Editorial Calm palette — walnut accents, paper bg, Newsreader for titles
- Responsive: model card grid 88px image + 1fr body
- Placeholder gradient when no image (camera emoji)

STANDALONE PREVIEW (preview-report.html):
- Mock AI response with 3 fridges + 2 hobs
- Same render logic, runs without backend
- Visit: https://wasrusgen.github.io/zov-tech/preview-report.html

NEXT: integrate proxy6 token → real photos/prices instead of placeholders
2026-05-11 12:26:58 +03:00
wasrusgen
82425dbd88 backend: Proxy6 pool + parsers WB / OZON / Я.Маркет / DNS
PROXY POOL (app/proxy_pool.py):
- Loads active proxies from Proxy6.net API every 10 min
- Random rotation per request via proxied_client(timeout, headers)
- Graceful fallback to direct HTTP if PROXY6_TOKEN not set
- Config: PROXY6_TOKEN env var

PARSERS (app/parsers/):
- dns.py — refactored to use proxy_pool with retry+rotation on Qrator block
- wb.py — Wildberries JSON API (search.wb.ru), retries on 429
- ozon.py — OZON composer-api JSON (widgetStates extraction)
- yamarket.py — Я.Маркет HTML + embedded JSON parser
- __init__.py — enrich_one() fans out to all sources, aggregates min/max prices, max rating, sum reviews
- enrich_models() — batch enrich for AI by_category output

NEW DIAGNOSTIC ENDPOINTS (main.py):
- GET /api/parse_wb?q=...&limit=N
- GET /api/parse_ozon?q=...&limit=N
- GET /api/parse_yamarket?q=...&limit=N
- GET /api/parse_all?q=... — fan-out + aggregate
- GET /api/proxy_status — pool diagnostics (count, token configured, age)

PODBOR (main.py):
- _enrich_ai_with_dns -> _enrich_ai_marketplaces (uses all sources)

DEPLOY: needs PROXY6_TOKEN in /opt/zov-tech/deploy/.env on VPS, then docker compose build + up -d backend
2026-05-11 12:18:04 +03:00
wasrusgen
64edb76035 backend: new state-shape AI prompt + DNS parser MVP
AI PROMPT (ai.py):
- Документирует новую форму checklist (per_cat.answers, brand_strategy, single_brand, brands, budget_preset, pick_strategies)
- Просит вернуть 3-5 моделей по КАЖДОЙ категории (не одну)
- Новый формат ответа: by_category[cat].models[] с brand/model/price_min/price_max/search_query/pros/cons/tier
- Подробные правила для бренд-стратегий (single → вся техника одной марки; different → preferred/acceptable/avoid)
- Бюджет-пресеты с авто-распределением по категориям (fridge ~25%, hob ~12% и т.д.)

DNS PARSER (parsers/dns.py):
- search_dns(query, limit) — HTTP + BeautifulSoup
- Реалистичный User-Agent, фолбэк на JSON-LD если HTML-селекторы не сработали
- enrich_models(models) — обогащает список моделей от AI, добавляя dns: {title, price, image, url, rating, reviews}
- Вежливая задержка 0.4с между запросами

MAIN.PY:
- /api/parse_dns?q=... — тестовый эндпоинт для проверки парсера
- _handle_podbor теперь после AI вызывает _enrich_ai_with_dns для каждой модели
- _format_podbor_for_telegram переписан под новый формат by_category — выводит 3-5 моделей в каждой категории с pros/cons
- Fallback на старый формат items[] для совместимости

REQUIREMENTS:
- + beautifulsoup4 >= 4.12
- + lxml >= 5.2

DEPLOY: после пуша на VPS нужно пересобрать backend контейнер (docker compose up --build -d backend)
2026-05-11 11:42:37 +03:00
wasrusgen
717c6ea138 miniapp: hierarchical wizard for all 8 categories + condition support
CATEGORIES MIGRATED to steps[] schema:
- hob: Источник нагрева → Подтип (multi, optionsBy) → Размер → Конфорки → Особенности
- oven: Установка → Функции (multi) → Размер → Где ставим (cond:built_in) → Особенности
- dw: Тип встройки → Класс (multi) → Ширина → Корзины → Особенности
- hood: Форм-фактор → Подключение → Ширина → Цвет (cond:visible-types) → Особенности
- microwave: Установка → Функции (multi) → Размер (optionsBy) → Особенности
- coffee: Тип → Молоко (cond:grinder/manual) → Вода (cond:built-in/tap) → Размер (cond:built-in) → Особенности
- washer: Установка → Функция → Глубина → Загрузка → Объём → Особенности

NEW PODBOR.JS FEATURES:
- isStepActive(step, answers) — predicate for condition field
- findNextActiveIdx / findPrevActiveIdx — skip inactive steps in navigation
- Auto-advance through inactive on single-select pick
- Review screen filters inactive steps
- isCategoryFilled checks only active single-steps
- buildPerCatSummary skips inactive
- Clearing dependent answers when condition's parent changes (in addition to optionsBy)

NEXT: pictograms for step 1 of each category (currently text-pin layout)
2026-05-11 11:28:50 +03:00
wasrusgen
dd400b71ac miniapp: new pricing flow — brand strategy + budget presets + multi pick strategy
NEW STRUCTURE:
- Step 4 'Бренд' — ai/single/different + brand picker or per-cat chips (now 4-state with 'avoid')
- Step 5 'Бюджет' — Люкс/Премиум/Средний/Бюджет/Точные цифры presets
- Step 6 'Стратегия' — multi: Лучшее по отзывам / Цена-качество / Топ-бренды / Доступное / Tech / Стиль
- Step 7 'Инфра' — перенесено после стратегии
- Step 8 'Итог' — обновлённый summary с новыми полями

FIXES:
- Keyboard-disappearing bug in price inputs — removed render() on input, total recomputed locally
- localStorage merge with defaults for backward compat with new fields
- Bumped STORAGE_KEY to v4

REMAINING:
- Backend still reads checklist.priorities (old shape) — needs update to read pick_strategies + brand_strategy + budget_preset
2026-05-11 10:43:54 +03:00
wasrusgen
496ddf793c miniapp: persistent category strip with active highlight + tap-to-jump
- Visible on all steps after categories are selected
- Highlights current category when inside its wizard
- Filled categories show checkmark
- Tap chip jumps directly to that category's wizard
- Horizontal scroll if many categories don't fit
2026-05-11 00:46:43 +03:00
wasrusgen
d289f7601e miniapp: compact pin layout for wizard steps without pictograms
- Steps with pict (1-3 fridge: install/chamber/size) keep grid cards
- Steps without pict (4 features) render as flex-wrap pill pins
- Auto-detect via options.some(o => o.pict)
2026-05-11 00:02:25 +03:00
wasrusgen
17b112f061 miniapp: hierarchical wizard for fridge category (style D pictograms)
- New PODBOR_PARAMS schema with steps[] supporting single/multi + optionsBy branches
- 11 fridge SVG pictograms in podbor.picts.js (style D — 3D perspective with shadow)
- renderCategoryWizard with step-by-step flow, chips for prior answers, review screen
- Legacy renderCategoryDetail still used for other 7 categories until migrated
- Auto-advance on single-select, Дальше button for multi-select
- Backend-compatible: per_cat[catKey].answers replaces .params/.features
2026-05-10 23:57:03 +03:00
wasrusgen
cd5d92ea17 miniapp: redraw fridge pictograms in style D (3D perspective with shadow) 2026-05-10 23:33:07 +03:00
wasrusgen
8991e7890d miniapp: add fridge style preview — 4 styles × 4 variants for comparison 2026-05-10 23:26:30 +03:00
wasrusgen
fe51f44bd9 fix(preview): correct fridge taxonomy — columns moved to built-in (3 variants: cold, freeze, pair); morozilka stays in freestanding 2026-05-10 23:16:43 +03:00
wasrusgen
f2e3333846 feat(miniapp/preview): SVG pictogram preview for fridge types (Editorial Calm, line-art, walnut) 2026-05-10 22:56:36 +03:00
wasrusgen
b7fa20dc69 fix(backend/sheets): write ISO-string for datetime (gspread can't serialize datetime) 2026-05-10 22:28:13 +03:00
wasrusgen
a849491f56 fix(miniapp): /api/me path for FastAPI backend (was ?path=me from Apps Script) 2026-05-10 22:19:34 +03:00
wasrusgen
f85d3a9d1e feat(miniapp): switch BACKEND_URL to Cloudflare Tunnel → VPS backend (GigaChat live) 2026-05-10 22:18:53 +03:00