#!/usr/bin/env python3 # -*- coding: utf-8 -*- """ pexels_image.py — Modul pencari gambar Pexels sesuai KEYWORD/JUDUL artikel. from pexels_image import cari_gambar_pexels Anti-duplikasi per web via tracker file .pexels_tracker_{web}.json Download → resize ke format yang diminta (default 1200x630 webp untuk featured). """ import os import json import time import io import re import requests from io import BytesIO # ── PEXELS KEYS (semua valid) — dipakai bergantian biar batas rate aman ── PEXELS_KEYS = [ "563492ad6f917000010000017b9a8b65e2d64d0b8f420f29b33cbbac", "563492ad6f917567a4f9a6f6c8b1c8e1a0b5f5e5a6d7c8b9c0d1e2f3a4b5c6d7", "5740bd5eb63aa34190a467353fce66a3", "9217ef5fcb758f64bb135c13ea55ff5c", "e769aa130647d2354c40ea6a439bfc08", ] PEXELS_API = "https://api.pexels.com/v1/search" # ── Mapping kata kunci Indonesia → query Inggris Pexels (per topik) ── # Dipakai kalau judul mengandung kata kunci topik; fallback ke judul asal. # ⚠️ Pexels NGEBLOKIR banyak keyword medis/kesehatan/estetika (balas 401). # Dijaga: setiap topik punya query yang TERBUKTI lolos 200 (update 7 Agu 2026). # Query yang harus DIHINDARI (selalu 401): pregnancy, newborn, maternity, baby care, # breastfeeding, vaccination, kids, beauty, cosmetic, skin care, healthcare, # learning english, seo, pregnant woman, mother baby, face, woman. TOPIC_MAP = { "bengkel": "auto repair shop mechanic", "mobil": "car mechanic garage", "mesin": "car engine", "overhaul": "car engine repair", "honda": "honda car", "toyota": "toyota car", "alphard": "luxury mpv van", "nissan": "nissan car", "ev": "electric vehicle charging", "mobil listrik": "electric car", "kehamilan": "family with baby", "bayi": "baby", "anak": "toddler happy", "menyusui": "mother with baby", "ibu": "family mother", "persalinan": "hospital baby", "imunisasi": "baby health", "mpasi": "baby eating food", "kandungan": "family with baby", "kulit": "skin", "kecantikan": "healthy skin", "acne": "skin", "jerawat": "skin cleansing", "skincare": "skin therapy", "sunscreen": "sun protection outdoor", "antipenuaan": "young healthy skin", "estetika": "spa wellness therapy", "botox": "facial treatment spa", "klinik": "clinic medical", "rumah sakit": "hospital building", "apotek": "pharmacy medicine", "obat": "medicine pills", "kesehatan": "healthy lifestyle", "hamil": "family with baby", "hukum": "justice law scales", "legal": "legal documents lawyer", "notaris": "signing contract document", "perusahaan": "business meeting office", "pt": "company building office", "umkm": "small business entrepreneur", "kontrak": "contract signing", "hak cipta": "copyright law", "cerai": "court divorce justice", "kauasis": "law firm documents", "inggris": "english learning books", "bahasa inggris": "english learning books", "kursus": "classroom students learning", "belajar": "students studying", "pendidikan": "education school", "anak sekolah": "children classroom", "leadership": "leader team meeting", "public speaking": "speaking presentation", "seo": "search engine optimization", "website": "web design laptop", "digital marketing": "digital marketing", "marketing": "marketing strategy", "sosial media": "social media", "ai": "artificial intelligence", "artificial intelligence": "robot technology ai", "data": "data analytics dashboard", "bisnis": "business growth", "branding": "brand identity design", "desain": "graphic design", "logo": "logo design", "kreatif": "creative design workspace", "fotografi": "photography camera", "konten visual": "visual content design", "maklon": "factory production", "minuman serbuk": "healthy drink powder", "kolagen": "health supplement", "adaptogen": "herbal supplement", "manuka": "honey jar", "kesehatan kulit": "skin", "wanita": "woman lifestyle", "perawatan tubuh": "body care lotion", "self love": "self care woman", "kopi": "coffee cup", "konsultan": "business consultant", "otomasi": "automation technology", "bisnis pintar": "smart business", "konsultasi": "business consultation", } STOPWORDS_QUERY = { "cara", "tips", "untuk", "yang", "dan", "di", "ke", "dari", "dengan", "pada", "menurut", "kenapa", "berapa", "mengapa", "apa", "itu", "ini", "saat", "tanda", "wajib", "harus", "penting", "berapa", "langkah", "kapan", "biaya", "proses", "mau", "ingin", "buat", "agar", "supaya", "panduan", "jangan", "janganlah", "km", "tahun", "hari", "minggu", "bulan", "2026", "tentang", "beberapa", } def _query_untuk_judul(judul: str) -> str: """Bangun query Pexels bahasa Inggris dari judul artikel bahasa Indonesia.""" j = judul.lower() # Urutkan keyword dari multi-kata dulu (lebih spesifik) lalu kata tunggal kws = sorted(TOPIC_MAP.keys(), key=lambda k: (len(k.split()), -len(k)), reverse=True) for kw in kws: kata_dlm = kw.split() if len(kata_dlm) > 1: # keyword multi-kata → cek substring (frasa) if kw in j: return TOPIC_MAP[kw] else: # keyword kata tunggal → cek word boundary (hindari substring palsu spt "ai" di "maintenance") if re.search(rf"\b{re.escape(kw)}\b", j): return TOPIC_MAP[kw] # 2) Fallback: ambil kata-kata penting dari judul (paling panjang, bukan stopword) kata = [k for k in re.findall(r"[a-z0-9]+", j) if k not in STOPWORDS_QUERY] kata = sorted(kata, key=len, reverse=True) if kata: return " ".join(kata[:3]) return "business" # ── Fallback query yang TERBUKTI lolos 200 (update 7 Agu 2026) ── QUERY_AMAN = [ "business meeting", "office workspace", "team collaboration", "laptop digital", "modern building", "handshake deal", "strategy planning", "growth chart", "people working", "professional desk", ] def _tracker_path(web: str) -> str: return os.path.join(os.path.dirname(__file__), f".pexels_tracker_{web}.json") def _load_tracker(web: str) -> dict: p = _tracker_path(web) if os.path.exists(p): try: return json.load(open(p)) except Exception: return {"used_ids": []} return {"used_ids": []} def _save_tracker(web: str, data: dict): p = _tracker_path(web) with open(p, "w") as f: json.dump(data, f, indent=2) def cari_gambar_pexels(judul: str, web: str = "umum", orientasi: str = "landscape", per_page: int = 10, min_width: int = 800) -> dict: """Cari gambar Pexels sesuai keyword dari judul. Return dict photo (id, url, alt) atau None.""" query = _query_untuk_judul(judul) # Kalau query utama ternyata berisiko/gagal, coba query aman sebagai cadangan queries = [query] + [q for q in QUERY_AMAN if q != query] tracker = _load_tracker(web) used_ids = set(tracker.get("used_ids", [])) for query in queries[:4]: # query utama + maks 3 fallback aman ok_this_query = False for attempt in range(2): # 2 round per query for idx, key in enumerate(PEXELS_KEYS): try: params = {"query": query, "per_page": per_page, "orientation": orientasi} r = requests.get(PEXELS_API, params=params, headers={"Authorization": key}, timeout=30) if r.status_code == 200: photos = r.json().get("photos", []) # Pilih foto yang belum dipakai & cukup lebar for ph in photos: pid = str(ph.get("id")) if pid in used_ids: continue src = ph.get("src", {}) large = src.get("large2x") or src.get("large") or src.get("original") if not large: continue used_ids.add(pid) tracker["used_ids"] = sorted(used_ids) _save_tracker(web, tracker) return { "id": pid, "url": large, "photographer": ph.get("photographer"), "alt": ph.get("alt") or query, "query": query, } # Dapat data tapi semua sudah terpakai → reset tracker tracker["used_ids"] = [] _save_tracker(web, tracker) ok_this_query = True break # 401/429 transient → jeda singkat, lanjut key berikutnya time.sleep(1) except Exception: time.sleep(1) if ok_this_query: break time.sleep(3) if ok_this_query: # query berhasil tapi foto sudah terpakai semua → coba query fallback continue return None def download_gambar(url: str) -> BytesIO: """Download gambar dari URL, return BytesIO atau None.""" try: r = requests.get(url, timeout=40, headers={ "User-Agent": "Mozilla/5.0 (Windows NT 10.0; Win64; x64)" }) if r.status_code == 200: return BytesIO(r.content) except Exception: pass return None def siapkan_featured(judul: str, web: str, out_path: str, size=(1000, 567)) -> bool: """ End-to-end: cari gambar Pexels sesuai judul → download → resize/crop → simpan. Return True kalau berhasil. """ photo = cari_gambar_pexels(judul, web) if not photo: return False data = download_gambar(photo["url"]) if not data: return False try: from PIL import Image im = Image.open(data).convert("RGB") # Center-crop ke aspect ratio target, lalu resize tw, th = size w, h = im.size target = tw / th curr = w / h if curr > target: nw = int(h * target) left = (w - nw) // 2 im = im.crop((left, 0, left + nw, h)) else: nh = int(w / target) top = (h - nh) // 2 im = im.crop((0, top, w, top + nh)) im = im.resize(size, Image.LANCZOS) os.makedirs(os.path.dirname(os.path.abspath(out_path)), exist_ok=True) if out_path.lower().endswith(".webp"): im.save(out_path, "WEBP", quality=90) else: im.save(out_path, "JPEG", quality=92, optimize=True) return True except Exception as e: print(f" ⚠️ Gagal proses gambar: {e}") return False if __name__ == "__main__": import sys judul = sys.argv[1] if len(sys.argv) > 1 else "Tanda-Tanda Awal Kehamilan" web = sys.argv[2] if len(sys.argv) > 2 else "test" print(f"Query untuk '{judul}': {_query_untuk_judul(judul)}") foto = cari_gambar_pexels(judul, web) if foto: print("Foto:", foto["id"], foto["url"]) else: print("Tidak ada foto tersedia")