1) HNS direct link u research_links: za sportaš s profile_url/source_url
(npr. https://semafor.hns.family/igraci/X/...) generira [⭐DIRECT] link na vrhu liste,
umjesto generic Google search. _research_links sada prima row dict.
2) Avatar cache buster: applyMeToHeader dodaje ?t=Date.now() na sve avatar img tagove.
Avatar upload handler dodatno persistira novi avatar_url u localStorage.pgz_user
tako da preživi page refresh + cross-page navigacije.
3) Logo home link: <div class='logo'> → <a href='/' class='logo'> u app.html i sport2.html.
Klik na PGŽ SPORT logo vodi na public portal.
4) Klub → Sportaši drill-down: u klub Info tabu dodan button
'👥 Vidi sportaše ovog kluba (N)' koji prebacuje na k-clan tab.
Plus '🌐 Službena stranica' link kad klub ima web.
5) Smarter klub enrichment:
- URL validacija (skip placeholder strings poput 'godisnjak_zspgz_2025')
- Domain candidate guesser (slug → 16 candidate URLs s common HR TLD-ovima i sport prefix-ima)
- Parallel HEAD probe (8 threads, 10s budget) — first 200 + name token match wins
- Subpage scrape (/kontakt, /uprava, /o-nama, /o-klubu, /predsjednik) za richer evidence
- HNK Orijent (id 3766) test: pogađa https://www.orijent.hr/, predlaže web+email+telefon+opis
E2E verified:
- 9/9 sidebar URL-ova → 200
- /users/me/gdpr-export → 200 (28KB JSON)
- /users/me/request-deletion → 200 (DB row pgz_sport.gdpr_erasure_requests)
- /enrich/klub/3766 → 4 proposed fields (web, email, telefon, opis)
- HNS sportaš research_links: ⭐ HNS profil DIRECT link na vrhu
Backend: routers/enrich_router.py
Frontend: static/app.html, static/sport2.html
Backups: _backups/sprint_1777940670/
Tag: R7-demo-ready
#1 JWT middleware extended:
- Was: /api/admin/* only
- Now: any POST/PUT/PATCH/DELETE under /api/* requires Bearer JWT
- Whitelist (still anonymous): /api/auth/login, /refresh, /forgot-password,
/password/reset, /reset-password, /setup-password, /google;
/api/gdpr/consent; any path ending /avatar
- 14 mutating endpoints verified to return 401 without token
#2 Avatar upload demo mode (routers/clan_panel_router.py):
- Anonymous → returns {demo_mode:true, slika_url:null,
message:'Demo mode — slika nije spremljena. Prijavite se za pravu pohranu.'},
no FS write, no DB write
- Authenticated (valid JWT, allowed role) → real save as before
- Auth check now uses auth.auth_v2.decode_token (proper secret + revocation)
instead of the broken local _resolve_role
#3 Mock mailer (auth/mailer.py):
- send_email writes RFC 822 .eml to /tmp/pgz_mailbox + appends to INDEX.jsonl
- send_password_reset, send_invite helpers with HR text + HTML alt
- Real SMTP active when PGZ_SMTP_HOST is set (env-driven, off by default)
- forgot-password and admin invite both call mailer; audit logs mail status
#5 Rate limiting on /api/auth/login:
- Per-user: 5 wrong attempts → 5-minute DB-backed lockout
(was 5 → 15 min). Configurable via PGZ_LOGIN_LOCK_THRESHOLD/MINUTES.
- Per-IP: 10 fails / 5-min sliding window in-memory → HTTP 429
Configurable via PGZ_LOGIN_IP_THRESHOLD/WINDOW_SEC. Successful
login clears the IP counter.
- Failed attempts respond '(N/5) — račun je zaključan na 5 minuta'
- New audit actions: login.ratelimit.ip; login.fail meta now
includes fails count, locked, lock_minutes
#4 Live test report: 46/46 across 6 demo users — login, JWT gate on 14
mutating endpoints, public path whitelist, demo-mode avatar +
real save, forgot-password e-mail to mailbox, no-leak unknown email,
5-fail lockout, 423 during lockout, audit coverage.
- New auth.gdpr.me_router prefix /api/users/me with:
- GET/POST /gdpr-export → Art.20 JSON download with Content-Disposition
- POST /gdpr-erase → Art.17 erasure request
- GET /gdpr-consent → consent history for caller
- jsonable_encoder fixes datetime serialisation in JSONResponse
- admin_users.html: 'Izvezi moje podatke' now POSTs to alias and uses
filename from Content-Disposition header
- 401 enforced on no-auth, 200 on valid Bearer (verified live)
- enrich_apply now imports audit_seal_router.audit_log and writes a sys_audit
row after every successful UPDATE: action='enrich.apply', target_type=kind,
target_id=eid, payload={applied: {...}, sources: [...]}, user from headers.
- Other modules (cc2 users, cc4 invoices/putni_nalozi, cc5 clanarine/lijecnicki/
obrasci) can call the same helper:
from audit_seal_router import audit_log
audit_log(action='users.update', target_type='users', target_id=u['id'],
payload={'changed':[...]}, user_email=actor)
- Verified: real apply on klub 4528 produced sys_audit id 102.
Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
Critical bug fix: /v2/enrich/sportas/{id} returned proposed:{} for athletes
because the v3 pipeline was still relying on Wikipedia-only evidence and never
actually fetched semafor.hns.family/igraci/.
- enrich_router._propose_for_sportas now:
• Resolves a HNS Semafor URL from profile_url, source_url, hns_igrac_id,
vanjski_id JSONB ('hns_comet'+'hns_slug'), or source='hns_semafor'+source_id.
• Fetches and parses the player page (BS4, regex fallback) and proposes
profile_url, source_url, slika_url, hns_igrac_id, datum_rodenja,
mjesto_rodenja, broj_dresa, biografija (DeepSeek synthesis from HNS+Wiki).
- _load_row(sportas) widened to read every relevant column + vanjski_id.
- _TABLE_MAP['sportas'] writeback whitelist expanded to 12 fields.
- workers/enrichment_worker.py: 24/7 daemon, picks under-enriched
clanovi/klubovi/savezi every 5 min via SQL, calls /apply for each.
- systemd unit pgz-sport-enricher.service installed + enabled.
- Tested end-to-end: id=2222 (Abdija) and id=449 (Zec) now have
profile_url, slika_url, hns_igrac_id, biografija persisted.
Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
- enrichment/playwright_scraper.py: fetch_rendered(), scrape_sport_pgz_klub(),
scrape_federation(). Headless Chromium, 12s timeout, returns rendered text.
Import-safe when playwright is missing.
- enrich_router._sport_pgz_search() now falls back to the JS path when the
cheap urllib fetch returns empty or unparseable HTML.
Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
- POST /v2/enrich/{kind}/{eid} now scrapes Wikipedia HR + sport-pgz.hr +
primary site, runs relevance filter so contact info from off-topic pages
isn't lifted, optionally calls DeepSeek for opis_djelatnosti, returns
{current, proposed, sources, last_enriched_at} for diff UI.
- POST /v2/enrich/{kind}/{eid}/apply UPDATES klubovi/savezi/clanovi for
whitelisted empty fields, sets metadata.enriched_at +
metadata.enrichment_source + metadata.enrichment_history, writes a row
to pgz_sport.enrichment_log (new table).
- GET /v2/enrich/log read-back endpoint.
- Tested on klub 3 (KK Kvarner 2010): opis_djelatnosti persisted; metadata
carries enriched_at + sources.
- New tables/columns: pgz_sport.enrichment_log; metadata jsonb on klubovi/savezi.
Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
- geocode_objekti_v2.py + DB updates (Kastav, Rujevica, Platak, Petehovac, Crikvenica, Krk hand-curated)
- Maps URL → /maps/search/?api=1 format for proper pin
- Dashboard: year selector for nositelji, click → klub/PDF panel; top savezi clickable
- Universal sort (asc/desc) on Savezi/Klubovi/Sportaši/Objekti/Manifestacije/Financije
- Card↔Table toggle on Financije
- Manifestacije: source_url direct open, Google fallback
- Forenzika: severity/tip filter, search, run-scan, Liverić PEP custom findings + DB alerts
- Enrich endpoint /api/v2/enrich/{kind}/{id} + button on savez/klub/sportaš panels
- New 'Mreža' section: D3 force graph from /api/v1/presenter/graph-real
Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
damir
Damir Radulić