config.py

# -*- coding: utf-8 -*-
"""
Configuration de l'Assistant Vizyon Ayiti 360 - version Hugging Face Space.
Config for the Vizyon Ayiti 360 Assistant - Hugging Face Space edition.

Sur Hugging Face, definissez les valeurs sensibles dans
Settings > Variables and secrets (elles arrivent comme variables d'env).
On Hugging Face, set sensitive values in Settings > Variables and secrets.
"""
import os
from pathlib import Path

try:
    from dotenv import load_dotenv
    load_dotenv()
except Exception:
    pass

# --------------------------------------------------------------------------
# Site source (seule source autorisee / only allowed source)
# --------------------------------------------------------------------------
SITE_URL = os.getenv("SITE_URL", "https://www.vizyonayiti360.org")
WP_API_BASE = f"{SITE_URL.rstrip('/')}/wp-json/wp/v2"

WP_CONTENT_TYPES = {
    "posts": "Article",
    "pages": "Page",
    "mec-events": "Evenement",
}

# --------------------------------------------------------------------------
# Authentification WordPress (pour lire le contenu NON public)
# WordPress authentication (to read NON-public content)
# --------------------------------------------------------------------------
# Cree un "mot de passe d'application" dans WordPress :
#   Utilisateurs > Profil > Mots de passe d'application.
# Create an "Application Password" in WordPress:
#   Users > Profile > Application Passwords.
WP_USERNAME = os.getenv("WP_USERNAME", "")
WP_APP_PASSWORD = os.getenv("WP_APP_PASSWORD", "")
# Inclure le contenu non public si des identifiants sont fournis.
INCLUDE_NON_PUBLIC = os.getenv("INCLUDE_NON_PUBLIC", "true").lower() == "true"
# Statuts a recuperer quand on est authentifie.
WP_STATUSES = os.getenv("WP_STATUSES", "publish,private,draft,pending,future")

# --------------------------------------------------------------------------
# Backend LLM : sur un Space, on utilise l'API Hugging Face.
# LLM backend: on a Space we use the Hugging Face Inference API.
# --------------------------------------------------------------------------
LLM_BACKEND = os.getenv("LLM_BACKEND", "hf_api").lower()

# Token HF : dans un Space, ajoutez le secret HF_TOKEN (fourni par defaut sur
# certains Spaces) ou HF_API_TOKEN.
HF_API_TOKEN = os.getenv("HF_API_TOKEN", os.getenv("HF_TOKEN", ""))
HF_MODEL = os.getenv("HF_MODEL", "meta-llama/Llama-3.1-8B-Instruct")

# Repli local optionnel (si vous testez hors Space avec Ollama).
OLLAMA_HOST = os.getenv("OLLAMA_HOST", "http://localhost:11434")
OLLAMA_MODEL = os.getenv("OLLAMA_MODEL", "qwen2.5:7b-instruct")

# --------------------------------------------------------------------------
# Embeddings (locaux, multilingues FR/EN) / local multilingual embeddings
# --------------------------------------------------------------------------
EMBEDDING_MODEL = os.getenv(
    "EMBEDDING_MODEL",
    "sentence-transformers/paraphrase-multilingual-MiniLM-L12-v2",
)

# --------------------------------------------------------------------------
# Parametres RAG / RAG parameters
# --------------------------------------------------------------------------
CHUNK_SIZE = int(os.getenv("CHUNK_SIZE", "800"))
CHUNK_OVERLAP = int(os.getenv("CHUNK_OVERLAP", "120"))
TOP_K = int(os.getenv("TOP_K", "5"))
MIN_SCORE = float(os.getenv("MIN_SCORE", "0.25"))
MAX_TOKENS = int(os.getenv("MAX_TOKENS", "1024"))
TEMPERATURE = float(os.getenv("TEMPERATURE", "0.2"))

# Reconstruire l'index a chaque demarrage du Space (contenu frais).
REBUILD_ON_START = os.getenv("REBUILD_ON_START", "true").lower() == "true"

# --------------------------------------------------------------------------
# Stockage / storage (ephemere sur un Space)
# --------------------------------------------------------------------------
BASE_DIR = Path(__file__).resolve().parent
DATA_DIR = BASE_DIR / "data"
INDEX_DIR = DATA_DIR / "index"
LOG_DIR = DATA_DIR / "logs"
for _d in (DATA_DIR, INDEX_DIR, LOG_DIR):
    _d.mkdir(parents=True, exist_ok=True)

INDEX_FILE = INDEX_DIR / "faiss.index"
CHUNKS_FILE = INDEX_DIR / "chunks.pkl"
META_FILE = INDEX_DIR / "meta.json"

# --------------------------------------------------------------------------
# Interface / UI
# --------------------------------------------------------------------------
APP_TITLE = "Assistant Vizyon Ayiti 360"
DEFAULT_LANG = os.getenv("DEFAULT_LANG", "fr")

# Jeton d'acces partage : si defini, l'app n'accepte que les visiteurs dont
# l'URL contient ?access=<jeton>. Protege l'acces direct au Space quand il est
# integre en iframe. Laisser vide pour desactiver.
# Shared access token: if set, the app only serves visitors whose URL contains
# ?access=<token>. Protects direct Space access when embedded via iframe.
APP_ACCESS_TOKEN = os.getenv("APP_ACCESS_TOKEN", "")

LOGO_URL = os.getenv(
    "LOGO_URL",
    "https://www.vizyonayiti360.org/wp-content/uploads/2025/12/"
    "cropped-logo-4_1-300x100.png",
)
FAVICON_URL = os.getenv(
    "FAVICON_URL",
    "https://www.vizyonayiti360.org/wp-content/uploads/2026/06/"
    "cropped-fivicon-192x192.png",
)

# --------------------------------------------------------------------------
# RGPD / GDPR
# --------------------------------------------------------------------------
ENABLE_LOGGING = os.getenv("ENABLE_LOGGING", "false").lower() == "true"
ANONYMIZE_INPUT = os.getenv("ANONYMIZE_INPUT", "true").lower() == "true"
LOG_RETENTION_DAYS = int(os.getenv("LOG_RETENTION_DAYS", "30"))

# Le contenu non public est confidentiel : ne jamais journaliser par defaut.
if INCLUDE_NON_PUBLIC:
    ENABLE_LOGGING = False