Damir Radulic
68 lines
2.6 KiB
Python
68 lines
2.6 KiB
Python
#!/usr/bin/env python3
|
|
from dotenv import load_dotenv
|
|
load_dotenv('/opt/rinet-gpu/.env.master')
|
|
# auto-added by patch_scrapers_with_dotenv.sh
|
|
# Fajl: hns_avatar_harvester.py | v1.0 | 05.05.2026
|
|
# Author: Damir Radulić
|
|
# Lokacija: /opt/pgz-sport/scripts/hns_avatar_harvester.py
|
|
# Svrha: Dohvati avatar URL za svakog igrača sa HNS profila
|
|
import os, time, re, json, sys
|
|
import psycopg2
|
|
import requests
|
|
from bs4 import BeautifulSoup
|
|
|
|
DSN = os.environ.get("RINET_DSN", f"host=10.10.0.2 port=6432 dbname=rinet_v3 user=rinet password={os.environ['DB_PASSWORD']}")
|
|
HEADERS = {"User-Agent": "Mozilla/5.0 (Ri.NET PGŽ Sport Bot)"}
|
|
|
|
conn = psycopg2.connect(DSN); conn.autocommit = True
|
|
|
|
def fetch_avatar(hns_id, slug=""):
|
|
url = f"https://semafor.hns.family/igraci/{hns_id}/"
|
|
if slug: url += f"{slug}/"
|
|
try:
|
|
r = requests.get(url, headers=HEADERS, timeout=15)
|
|
if r.status_code != 200: return None
|
|
soup = BeautifulSoup(r.text, "html.parser")
|
|
# Player photo selectors
|
|
for sel in [".playerPhoto img", ".player-photo img", ".playerHeader img", "img.player_photo"]:
|
|
img = soup.select_one(sel)
|
|
if img and img.get("src"):
|
|
src = img["src"]
|
|
if src.startswith("/"): src = "https://hns.family" + src
|
|
return src
|
|
# Generic: first img inside header
|
|
hdr = soup.select_one(".playerHeader, .player-header, .basic_info")
|
|
if hdr:
|
|
img = hdr.find("img")
|
|
if img and img.get("src"):
|
|
src = img["src"]
|
|
if src.startswith("/"): src = "https://hns.family" + src
|
|
return src
|
|
return None
|
|
except Exception as e:
|
|
return None
|
|
|
|
with conn.cursor() as cur:
|
|
cur.execute("""
|
|
SELECT id, hns_igrac_id, ime, prezime
|
|
FROM pgz_sport.clanovi
|
|
WHERE hns_igrac_id IS NOT NULL AND foto_url IS NULL
|
|
LIMIT 200
|
|
""")
|
|
rows = cur.fetchall()
|
|
|
|
print(f"Total: {len(rows)} igrača za avatar fetch")
|
|
hits = 0
|
|
for i, (cid, hns_id, ime, prezime) in enumerate(rows):
|
|
slug = f"{ime}-{prezime}".lower().replace("ć","c").replace("č","c").replace("š","s").replace("ž","z").replace("đ","d").replace(" ","-")
|
|
slug = re.sub(r"[^a-z0-9-]", "", slug)
|
|
avatar = fetch_avatar(hns_id, slug)
|
|
if avatar:
|
|
with conn.cursor() as cur:
|
|
cur.execute("UPDATE pgz_sport.clanovi SET foto_url=%s WHERE id=%s", (avatar, cid))
|
|
hits += 1
|
|
if i % 10 == 0: print(f" [{i+1}/{len(rows)}] {ime} {prezime} → {avatar[:80]}")
|
|
time.sleep(0.5)
|
|
|
|
print(f"\nDONE: {hits}/{len(rows)} avatar URL-ova spremljen")
|