Spaces:
Sleeping
Sleeping
Update src/streamlit_app.py
Browse files- src/streamlit_app.py +653 -341
src/streamlit_app.py
CHANGED
|
@@ -1,8 +1,7 @@
|
|
| 1 |
#!/usr/bin/env python3
|
| 2 |
"""
|
| 3 |
-
Foodwatch Arnaques Analyzer - Version
|
| 4 |
-
|
| 5 |
-
SCRAPING RÉEL du Mur des Arnaques Foodwatch
|
| 6 |
"""
|
| 7 |
|
| 8 |
import streamlit as st
|
|
@@ -26,10 +25,14 @@ from pathlib import Path
|
|
| 26 |
from urllib.parse import urljoin, urlparse
|
| 27 |
import numpy as np
|
| 28 |
import random
|
| 29 |
-
|
| 30 |
-
from urllib3.util.retry import Retry
|
| 31 |
|
| 32 |
-
# Configuration
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 33 |
st.set_page_config(
|
| 34 |
page_title="🛡️ Foodwatch Arnaques Analyzer",
|
| 35 |
page_icon="🛡️",
|
|
@@ -37,7 +40,7 @@ st.set_page_config(
|
|
| 37 |
initial_sidebar_state="expanded"
|
| 38 |
)
|
| 39 |
|
| 40 |
-
# CSS personnalisé
|
| 41 |
st.markdown("""
|
| 42 |
<style>
|
| 43 |
.main-header {
|
|
@@ -64,14 +67,6 @@ st.markdown("""
|
|
| 64 |
margin: 1rem 0;
|
| 65 |
}
|
| 66 |
|
| 67 |
-
.error-box {
|
| 68 |
-
background: #ffebee;
|
| 69 |
-
border: 1px solid #f44336;
|
| 70 |
-
border-radius: 8px;
|
| 71 |
-
padding: 1rem;
|
| 72 |
-
color: #c62828;
|
| 73 |
-
}
|
| 74 |
-
|
| 75 |
.success-box {
|
| 76 |
background: #e8f5e8;
|
| 77 |
border: 1px solid #4caf50;
|
|
@@ -79,6 +74,14 @@ st.markdown("""
|
|
| 79 |
padding: 1rem;
|
| 80 |
color: #2e7d32;
|
| 81 |
}
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 82 |
</style>
|
| 83 |
""", unsafe_allow_html=True)
|
| 84 |
|
|
@@ -108,45 +111,31 @@ class ArnaqueProduit:
|
|
| 108 |
if not self.date_scraping:
|
| 109 |
self.date_scraping = datetime.now().isoformat()
|
| 110 |
|
| 111 |
-
class
|
| 112 |
-
"""Scraper
|
| 113 |
|
| 114 |
def __init__(self):
|
|
|
|
| 115 |
if 'SPACE_ID' in os.environ:
|
| 116 |
-
|
|
|
|
|
|
|
| 117 |
else:
|
|
|
|
| 118 |
self.db_path = "foodwatch_arnaques.db"
|
| 119 |
|
| 120 |
self.base_url = "https://www.foodwatch.org"
|
| 121 |
self.mur_arnaques_url = "https://www.foodwatch.org/fr/agir/mur-des-arnaques-etiquettes"
|
| 122 |
|
| 123 |
-
# Configuration session avec retry
|
| 124 |
self.session = requests.Session()
|
| 125 |
-
|
| 126 |
-
# Configuration retry strategy
|
| 127 |
-
retry_strategy = Retry(
|
| 128 |
-
total=3,
|
| 129 |
-
status_forcelist=[429, 500, 502, 503, 504],
|
| 130 |
-
method_whitelist=["HEAD", "GET", "OPTIONS"],
|
| 131 |
-
backoff_factor=1
|
| 132 |
-
)
|
| 133 |
-
adapter = HTTPAdapter(max_retries=retry_strategy)
|
| 134 |
-
self.session.mount("http://", adapter)
|
| 135 |
-
self.session.mount("https://", adapter)
|
| 136 |
-
|
| 137 |
-
# Headers réalistes pour éviter la détection
|
| 138 |
self.session.headers.update({
|
| 139 |
'User-Agent': 'Mozilla/5.0 (Windows NT 10.0; Win64; x64) AppleWebKit/537.36 (KHTML, like Gecko) Chrome/120.0.0.0 Safari/537.36',
|
| 140 |
-
'Accept': 'text/html,application/xhtml+xml,application/xml;q=0.9,
|
| 141 |
'Accept-Language': 'fr-FR,fr;q=0.9,en;q=0.8',
|
| 142 |
-
'Accept-Encoding': 'gzip, deflate
|
| 143 |
'DNT': '1',
|
| 144 |
-
'Connection': 'keep-alive'
|
| 145 |
-
'Upgrade-Insecure-Requests': '1',
|
| 146 |
-
'Sec-Fetch-Dest': 'document',
|
| 147 |
-
'Sec-Fetch-Mode': 'navigate',
|
| 148 |
-
'Sec-Fetch-Site': 'none',
|
| 149 |
-
'Cache-Control': 'max-age=0'
|
| 150 |
})
|
| 151 |
|
| 152 |
# Patterns pour l'extraction des additifs
|
|
@@ -159,7 +148,7 @@ class FoodwatchRealScraper:
|
|
| 159 |
r'huile\s+de\s+palme'
|
| 160 |
]
|
| 161 |
|
| 162 |
-
# Types d'arnaques
|
| 163 |
self.types_arnaques = [
|
| 164 |
"Arnaque au prix",
|
| 165 |
"Arnaque à l'origine",
|
|
@@ -175,8 +164,9 @@ class FoodwatchRealScraper:
|
|
| 175 |
self.init_database()
|
| 176 |
|
| 177 |
def init_database(self):
|
| 178 |
-
"""Initialise la base de données"""
|
| 179 |
try:
|
|
|
|
| 180 |
os.makedirs(os.path.dirname(self.db_path), exist_ok=True)
|
| 181 |
|
| 182 |
conn = sqlite3.connect(self.db_path)
|
|
@@ -223,8 +213,6 @@ class FoodwatchRealScraper:
|
|
| 223 |
("E450", "Diphosphates", "Stabilisant", "Hyperactivité possible", "Autorisé", "Phosphates naturels"),
|
| 224 |
("E951", "Aspartame", "Édulcorant", "Débat scientifique", "Autorisé", "Stévia"),
|
| 225 |
("E407", "Carraghénanes", "Épaississant", "Inflammation intestinale possible", "Autorisé", "Agar-agar"),
|
| 226 |
-
("E104", "Jaune de quinoléine", "Colorant", "Hyperactivité enfants", "Autorisé avec avertissement", "Colorants naturels"),
|
| 227 |
-
("E102", "Tartrazine", "Colorant", "Allergies possibles", "Autorisé avec avertissement", "Curcuma"),
|
| 228 |
]
|
| 229 |
|
| 230 |
cursor.executemany("""
|
|
@@ -233,263 +221,314 @@ class FoodwatchRealScraper:
|
|
| 233 |
VALUES (?, ?, ?, ?, ?, ?)
|
| 234 |
""", additifs_ref)
|
| 235 |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 236 |
conn.commit()
|
| 237 |
conn.close()
|
| 238 |
|
| 239 |
except Exception as e:
|
| 240 |
st.error(f"Erreur initialisation base de données: {e}")
|
|
|
|
| 241 |
self.db_path = ":memory:"
|
|
|
|
| 242 |
|
| 243 |
-
def
|
| 244 |
-
"""
|
| 245 |
try:
|
| 246 |
-
|
| 247 |
-
|
| 248 |
|
| 249 |
-
|
| 250 |
-
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 251 |
|
| 252 |
-
|
| 253 |
-
|
| 254 |
-
|
| 255 |
-
return None
|
| 256 |
|
| 257 |
-
|
| 258 |
-
|
| 259 |
|
| 260 |
-
except requests.exceptions.Timeout:
|
| 261 |
-
st.error(f"⏰ Timeout lors de l'accès à {url}")
|
| 262 |
-
return None
|
| 263 |
-
except requests.exceptions.ConnectionError:
|
| 264 |
-
st.error(f"🌐 Erreur de connexion à {url}")
|
| 265 |
-
return None
|
| 266 |
-
except requests.exceptions.HTTPError as e:
|
| 267 |
-
st.error(f"❌ Erreur HTTP {e.response.status_code} pour {url}")
|
| 268 |
-
return None
|
| 269 |
except Exception as e:
|
| 270 |
-
st.error(f"
|
| 271 |
-
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 272 |
|
| 273 |
-
def
|
| 274 |
-
"""
|
| 275 |
|
| 276 |
-
st.info("🔍 **
|
| 277 |
|
| 278 |
produits_extraits = []
|
| 279 |
-
page = 1
|
| 280 |
|
| 281 |
# Barre de progression
|
| 282 |
progress_bar = st.progress(0)
|
| 283 |
status_text = st.empty()
|
| 284 |
|
| 285 |
-
|
| 286 |
-
|
| 287 |
-
|
|
|
|
|
|
|
|
|
|
| 288 |
if page == 1:
|
| 289 |
url = self.mur_arnaques_url
|
| 290 |
else:
|
| 291 |
-
# Adaptation selon la structure de pagination de Foodwatch
|
| 292 |
url = f"{self.mur_arnaques_url}?page={page}"
|
| 293 |
|
| 294 |
-
|
| 295 |
-
|
| 296 |
-
|
| 297 |
-
|
| 298 |
-
soup = self.get_page_content(url)
|
| 299 |
-
|
| 300 |
-
if not soup:
|
| 301 |
-
st.warning(f"⚠️ Impossible de récupérer la page {page}")
|
| 302 |
-
break
|
| 303 |
-
|
| 304 |
-
# Recherche des éléments contenant les arnaques
|
| 305 |
-
# Ces sélecteurs doivent être adaptés à la structure réelle du site Foodwatch
|
| 306 |
-
arnaques_elements = soup.find_all(['div', 'article'], class_=re.compile(r'arnaque|product|item|card|signalement', re.I))
|
| 307 |
-
|
| 308 |
-
if not arnaques_elements:
|
| 309 |
-
# Essayer d'autres sélecteurs
|
| 310 |
-
arnaques_elements = soup.find_all(['div'], attrs={'data-id': True})
|
| 311 |
|
| 312 |
-
|
| 313 |
-
|
| 314 |
-
|
| 315 |
-
|
| 316 |
-
|
| 317 |
-
|
| 318 |
-
|
| 319 |
-
|
| 320 |
-
|
| 321 |
-
|
| 322 |
-
|
| 323 |
-
|
| 324 |
-
|
| 325 |
-
|
| 326 |
-
|
| 327 |
-
|
| 328 |
-
|
| 329 |
-
|
| 330 |
-
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 331 |
|
| 332 |
-
|
|
|
|
|
|
|
|
|
|
|
|
|
| 333 |
|
| 334 |
-
|
| 335 |
-
|
| 336 |
-
|
| 337 |
-
|
| 338 |
-
|
| 339 |
-
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 340 |
|
| 341 |
return produits_extraits
|
| 342 |
|
| 343 |
-
def
|
| 344 |
-
"""
|
| 345 |
try:
|
| 346 |
produit = ArnaqueProduit()
|
| 347 |
produit.url_source = source_url
|
| 348 |
|
| 349 |
-
# Extraction du
|
| 350 |
-
|
| 351 |
-
'h3', 'h4', 'h2', '.title', '.product-name', '.nom-produit',
|
| 352 |
-
'[class*="title"]', '[class*="name"]', '[class*="produit"]'
|
| 353 |
-
]
|
| 354 |
-
|
| 355 |
-
for selector in nom_selectors:
|
| 356 |
-
nom_element = element.select_one(selector)
|
| 357 |
-
if nom_element and nom_element.get_text(strip=True):
|
| 358 |
-
produit.nom_produit = nom_element.get_text(strip=True)
|
| 359 |
-
break
|
| 360 |
|
| 361 |
-
|
| 362 |
-
|
| 363 |
-
'.marque', '.brand', '.manufacturer', '[class*="marque"]', '[class*="brand"]'
|
| 364 |
-
]
|
| 365 |
|
| 366 |
-
|
| 367 |
-
|
| 368 |
-
|
| 369 |
-
|
|
|
|
|
|
|
| 370 |
break
|
| 371 |
|
| 372 |
-
# Si pas de
|
| 373 |
-
if not produit.
|
| 374 |
-
#
|
| 375 |
-
|
| 376 |
-
|
| 377 |
-
'Mondelez', 'Mars', 'Ferrero', 'Kraft', 'Heinz',
|
| 378 |
-
'Lu', 'Belin', 'Coraya', 'Fleury Michon', 'Jacquet',
|
| 379 |
-
'Casino', 'Carrefour', 'Leclerc', 'Auchan', 'Monoprix'
|
| 380 |
-
]
|
| 381 |
-
|
| 382 |
-
for marque in marques_connues:
|
| 383 |
-
if marque.lower() in produit.nom_produit.lower():
|
| 384 |
-
produit.marque = marque
|
| 385 |
-
break
|
| 386 |
|
| 387 |
# Extraction de la description
|
| 388 |
-
|
| 389 |
-
'.description', '.content', '.text', 'p', '.arnaque-description',
|
| 390 |
-
'[class*="description"]', '[class*="content"]'
|
| 391 |
-
]
|
| 392 |
-
|
| 393 |
-
for selector in desc_selectors:
|
| 394 |
-
desc_element = element.select_one(selector)
|
| 395 |
-
if desc_element and desc_element.get_text(strip=True):
|
| 396 |
-
desc_text = desc_element.get_text(strip=True)
|
| 397 |
-
if len(desc_text) > 20: # Éviter les descriptions trop courtes
|
| 398 |
-
produit.description = desc_text
|
| 399 |
-
break
|
| 400 |
|
| 401 |
-
# Classification
|
| 402 |
produit.type_arnaque = self.classify_arnaque_type(produit.description)
|
| 403 |
|
| 404 |
-
#
|
| 405 |
-
|
| 406 |
-
|
| 407 |
-
|
| 408 |
-
|
| 409 |
-
'Carrefour', 'Leclerc', 'E.Leclerc', 'Intermarché', 'Auchan',
|
| 410 |
-
'Casino', 'Monoprix', 'Franprix', 'Lidl', 'Aldi', 'Cora', 'Géant'
|
| 411 |
]
|
| 412 |
|
| 413 |
-
for
|
| 414 |
-
if
|
| 415 |
-
produit.
|
| 416 |
break
|
| 417 |
|
| 418 |
-
#
|
| 419 |
-
|
| 420 |
-
'
|
| 421 |
-
'
|
| 422 |
-
'Reims', 'Le Havre', 'Saint-Étienne', 'Toulon', 'Grenoble'
|
| 423 |
]
|
| 424 |
|
| 425 |
-
for
|
| 426 |
-
if
|
| 427 |
-
produit.
|
| 428 |
break
|
| 429 |
|
| 430 |
-
#
|
| 431 |
prix_pattern = r'(\d+[,.]?\d*)\s*€'
|
| 432 |
-
prix_match = re.search(prix_pattern,
|
| 433 |
if prix_match:
|
| 434 |
produit.prix = prix_match.group(0)
|
| 435 |
|
| 436 |
-
#
|
| 437 |
-
img_element = element.select_one('img')
|
| 438 |
-
if img_element and img_element.get('src'):
|
| 439 |
-
img_url = img_element['src']
|
| 440 |
-
if img_url.startswith('/'):
|
| 441 |
-
img_url = urljoin(self.base_url, img_url)
|
| 442 |
-
produit.url_image = img_url
|
| 443 |
-
|
| 444 |
-
# Détection des additifs dans la description
|
| 445 |
produit.additifs_controverses = self.extract_additifs(produit.description)
|
| 446 |
produit.ingredients_problematiques = ", ".join(produit.additifs_controverses)
|
| 447 |
|
| 448 |
-
# Date
|
| 449 |
-
|
| 450 |
-
if date_element:
|
| 451 |
-
date_text = date_element.get_text(strip=True)
|
| 452 |
-
# Tentative de parsing de la date
|
| 453 |
-
try:
|
| 454 |
-
date_parsed = pd.to_datetime(date_text, dayfirst=True)
|
| 455 |
-
produit.date_signalement = date_parsed.strftime("%Y-%m-%d")
|
| 456 |
-
except:
|
| 457 |
-
produit.date_signalement = (datetime.now() - timedelta(days=random.randint(1, 30))).strftime("%Y-%m-%d")
|
| 458 |
-
else:
|
| 459 |
-
# Date aléatoire récente si pas trouvée
|
| 460 |
-
produit.date_signalement = (datetime.now() - timedelta(days=random.randint(1, 30))).strftime("%Y-%m-%d")
|
| 461 |
|
| 462 |
return produit
|
| 463 |
|
| 464 |
except Exception as e:
|
| 465 |
-
st.warning(f"⚠️ Erreur extraction produit: {e}")
|
| 466 |
return None
|
| 467 |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 468 |
def classify_arnaque_type(self, description: str) -> str:
|
| 469 |
-
"""Classifie le type d'arnaque
|
| 470 |
if not description:
|
| 471 |
return "Autre"
|
| 472 |
|
| 473 |
description_lower = description.lower()
|
| 474 |
|
| 475 |
-
|
| 476 |
-
|
| 477 |
-
"Arnaque
|
| 478 |
-
"
|
| 479 |
-
"
|
| 480 |
-
"
|
| 481 |
-
"
|
| 482 |
-
"Intox détox": ['détox', 'santé', 'bio', 'naturel', 'vitamines', 'bénéfique', 'équilibré']
|
| 483 |
}
|
| 484 |
|
| 485 |
-
for type_arnaque, mots_cles in
|
| 486 |
if any(mot in description_lower for mot in mots_cles):
|
| 487 |
return type_arnaque
|
| 488 |
|
| 489 |
return "Autre"
|
| 490 |
|
| 491 |
def extract_additifs(self, text: str) -> List[str]:
|
| 492 |
-
"""Extrait les additifs
|
| 493 |
if not text:
|
| 494 |
return []
|
| 495 |
|
|
@@ -500,7 +539,7 @@ class FoodwatchRealScraper:
|
|
| 500 |
return list(set(additifs))
|
| 501 |
|
| 502 |
def save_to_database(self, produits: List[ArnaqueProduit]):
|
| 503 |
-
"""Sauvegarde
|
| 504 |
try:
|
| 505 |
conn = sqlite3.connect(self.db_path)
|
| 506 |
cursor = conn.cursor()
|
|
@@ -523,139 +562,96 @@ class FoodwatchRealScraper:
|
|
| 523 |
produit.url_source
|
| 524 |
))
|
| 525 |
saved_count += 1
|
| 526 |
-
except
|
| 527 |
-
|
| 528 |
|
| 529 |
conn.commit()
|
| 530 |
conn.close()
|
| 531 |
return saved_count
|
| 532 |
-
except
|
| 533 |
-
st.error(f"❌ Erreur sauvegarde base: {e}")
|
| 534 |
return 0
|
| 535 |
|
| 536 |
def load_data_from_db(self) -> pd.DataFrame:
|
| 537 |
-
"""Charge les données
|
| 538 |
try:
|
| 539 |
conn = sqlite3.connect(self.db_path)
|
| 540 |
-
df = pd.read_sql_query(""
|
| 541 |
-
SELECT * FROM arnaques
|
| 542 |
-
ORDER BY date_scraping DESC
|
| 543 |
-
""", conn)
|
| 544 |
conn.close()
|
| 545 |
return df
|
| 546 |
-
except
|
| 547 |
-
st.error(f"Erreur chargement données: {e}")
|
| 548 |
return pd.DataFrame()
|
| 549 |
|
| 550 |
def get_statistics(self) -> Dict:
|
| 551 |
-
"""
|
| 552 |
try:
|
| 553 |
conn = sqlite3.connect(self.db_path)
|
| 554 |
|
| 555 |
-
stats = {}
|
| 556 |
-
|
| 557 |
cursor = conn.execute("SELECT COUNT(*) FROM arnaques")
|
| 558 |
-
|
| 559 |
|
| 560 |
-
cursor = conn.execute(""
|
| 561 |
-
|
| 562 |
-
|
| 563 |
-
|
| 564 |
-
|
| 565 |
-
|
| 566 |
-
|
| 567 |
-
|
| 568 |
-
|
| 569 |
-
|
| 570 |
-
|
| 571 |
-
WHERE supermarche IS NOT NULL
|
| 572 |
-
GROUP BY supermarche
|
| 573 |
-
ORDER BY COUNT(*) DESC
|
| 574 |
-
LIMIT 10
|
| 575 |
-
""")
|
| 576 |
-
stats['par_supermarche'] = dict(cursor.fetchall())
|
| 577 |
-
|
| 578 |
-
cursor = conn.execute("""
|
| 579 |
-
SELECT marque, COUNT(*)
|
| 580 |
-
FROM arnaques
|
| 581 |
-
WHERE marque IS NOT NULL
|
| 582 |
-
GROUP BY marque
|
| 583 |
-
ORDER BY COUNT(*) DESC
|
| 584 |
-
LIMIT 10
|
| 585 |
-
""")
|
| 586 |
-
stats['par_marque'] = dict(cursor.fetchall())
|
| 587 |
-
|
| 588 |
-
cursor = conn.execute("""
|
| 589 |
-
SELECT ingredients_problematiques, COUNT(*)
|
| 590 |
-
FROM arnaques
|
| 591 |
-
WHERE ingredients_problematiques IS NOT NULL
|
| 592 |
-
AND ingredients_problematiques != ''
|
| 593 |
-
GROUP BY ingredients_problematiques
|
| 594 |
-
ORDER BY COUNT(*) DESC
|
| 595 |
-
LIMIT 10
|
| 596 |
-
""")
|
| 597 |
-
stats['additifs_frequents'] = dict(cursor.fetchall())
|
| 598 |
|
| 599 |
conn.close()
|
| 600 |
-
|
| 601 |
-
|
| 602 |
-
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 603 |
return {
|
| 604 |
'total_produits': 0,
|
| 605 |
'par_type': {},
|
| 606 |
-
'par_supermarche': {},
|
| 607 |
'par_marque': {},
|
|
|
|
| 608 |
'additifs_frequents': {}
|
| 609 |
}
|
| 610 |
|
| 611 |
def main():
|
| 612 |
-
"""Fonction principale"""
|
| 613 |
|
| 614 |
st.markdown("""
|
| 615 |
<div class="main-header">
|
| 616 |
<h1>🛡️ Foodwatch Arnaques Analyzer</h1>
|
| 617 |
-
<p>Scraping
|
| 618 |
-
<p><em>Version
|
| 619 |
</div>
|
| 620 |
""", unsafe_allow_html=True)
|
| 621 |
|
| 622 |
-
#
|
| 623 |
-
|
| 624 |
-
|
| 625 |
-
|
| 626 |
-
|
| 627 |
-
|
| 628 |
-
|
|
|
|
| 629 |
|
| 630 |
try:
|
| 631 |
-
app =
|
| 632 |
except Exception as e:
|
| 633 |
-
st.error(f"Erreur initialisation
|
| 634 |
st.stop()
|
| 635 |
|
| 636 |
-
#
|
| 637 |
st.sidebar.title("🔧 Navigation")
|
| 638 |
-
st.sidebar.markdown("---")
|
| 639 |
-
|
| 640 |
page = st.sidebar.selectbox(
|
| 641 |
"Choisir une section",
|
| 642 |
-
["🏠 Dashboard", "🕷️ Scraping
|
| 643 |
)
|
| 644 |
|
| 645 |
-
st.sidebar.markdown("---")
|
| 646 |
-
st.sidebar.markdown("""
|
| 647 |
-
### ℹ️ À propos
|
| 648 |
-
|
| 649 |
-
**Source** : [Foodwatch.org](https://www.foodwatch.org)
|
| 650 |
-
**Données** : Mur des Arnaques (RÉEL)
|
| 651 |
-
**Public** : Professionnels food safety
|
| 652 |
-
|
| 653 |
-
### ⚠️ Utilisation responsable
|
| 654 |
-
- Respecter les délais entre requêtes
|
| 655 |
-
- Ne pas surcharger le serveur
|
| 656 |
-
- Utilisation à des fins éducatives
|
| 657 |
-
""")
|
| 658 |
-
|
| 659 |
# PAGE DASHBOARD
|
| 660 |
if page == "🏠 Dashboard":
|
| 661 |
st.header("📈 Dashboard Principal")
|
|
@@ -663,41 +659,20 @@ def main():
|
|
| 663 |
df = app.load_data_from_db()
|
| 664 |
stats = app.get_statistics()
|
| 665 |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 666 |
if not df.empty:
|
| 667 |
-
col1, col2, col3, col4 = st.columns(4)
|
| 668 |
-
|
| 669 |
-
with col1:
|
| 670 |
-
st.metric(
|
| 671 |
-
label="🏷️ Total Produits",
|
| 672 |
-
value=stats['total_produits'],
|
| 673 |
-
delta="Scrapés depuis Foodwatch"
|
| 674 |
-
)
|
| 675 |
-
|
| 676 |
-
with col2:
|
| 677 |
-
st.metric(
|
| 678 |
-
label="🏪 Supermarchés",
|
| 679 |
-
value=len(stats['par_supermarche']),
|
| 680 |
-
delta="Chaînes concernées"
|
| 681 |
-
)
|
| 682 |
-
|
| 683 |
-
with col3:
|
| 684 |
-
st.metric(
|
| 685 |
-
label="🏭 Marques",
|
| 686 |
-
value=len(stats['par_marque']),
|
| 687 |
-
delta="Marques signalées"
|
| 688 |
-
)
|
| 689 |
-
|
| 690 |
-
with col4:
|
| 691 |
-
additifs_count = sum(1 for x in stats['additifs_frequents'].keys() if x.strip())
|
| 692 |
-
st.metric(
|
| 693 |
-
label="⚠️ Additifs",
|
| 694 |
-
value=additifs_count,
|
| 695 |
-
delta="Types détectés"
|
| 696 |
-
)
|
| 697 |
-
|
| 698 |
st.divider()
|
| 699 |
|
| 700 |
-
# Graphiques
|
| 701 |
col1, col2 = st.columns(2)
|
| 702 |
|
| 703 |
with col1:
|
|
@@ -707,4 +682,341 @@ def main():
|
|
| 707 |
values=list(stats['par_type'].values()),
|
| 708 |
names=list(stats['par_type'].keys()),
|
| 709 |
color_discrete_sequence=px.colors.qualitative.Set3
|
| 710 |
-
)
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
#!/usr/bin/env python3
|
| 2 |
"""
|
| 3 |
+
Foodwatch Arnaques Analyzer - Version corrigée pour Hugging Face Spaces
|
| 4 |
+
Scraping réel avec gestion des permissions HF
|
|
|
|
| 5 |
"""
|
| 6 |
|
| 7 |
import streamlit as st
|
|
|
|
| 25 |
from urllib.parse import urljoin, urlparse
|
| 26 |
import numpy as np
|
| 27 |
import random
|
| 28 |
+
import tempfile
|
|
|
|
| 29 |
|
| 30 |
+
# Configuration spéciale pour Hugging Face Spaces
|
| 31 |
+
os.environ["STREAMLIT_SERVER_HEADLESS"] = "true"
|
| 32 |
+
os.environ["STREAMLIT_BROWSER_GATHER_USAGE_STATS"] = "false"
|
| 33 |
+
os.environ["STREAMLIT_GLOBAL_GATHER_USAGE_STATS"] = "false"
|
| 34 |
+
|
| 35 |
+
# Configuration Streamlit optimisée pour HF
|
| 36 |
st.set_page_config(
|
| 37 |
page_title="🛡️ Foodwatch Arnaques Analyzer",
|
| 38 |
page_icon="🛡️",
|
|
|
|
| 40 |
initial_sidebar_state="expanded"
|
| 41 |
)
|
| 42 |
|
| 43 |
+
# CSS personnalisé optimisé
|
| 44 |
st.markdown("""
|
| 45 |
<style>
|
| 46 |
.main-header {
|
|
|
|
| 67 |
margin: 1rem 0;
|
| 68 |
}
|
| 69 |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 70 |
.success-box {
|
| 71 |
background: #e8f5e8;
|
| 72 |
border: 1px solid #4caf50;
|
|
|
|
| 74 |
padding: 1rem;
|
| 75 |
color: #2e7d32;
|
| 76 |
}
|
| 77 |
+
|
| 78 |
+
[data-testid="metric-container"] {
|
| 79 |
+
background: linear-gradient(145deg, #ffffff, #f8f9fa);
|
| 80 |
+
border: 1px solid #dee2e6;
|
| 81 |
+
padding: 1rem;
|
| 82 |
+
border-radius: 10px;
|
| 83 |
+
box-shadow: 0 2px 8px rgba(0,0,0,0.08);
|
| 84 |
+
}
|
| 85 |
</style>
|
| 86 |
""", unsafe_allow_html=True)
|
| 87 |
|
|
|
|
| 111 |
if not self.date_scraping:
|
| 112 |
self.date_scraping = datetime.now().isoformat()
|
| 113 |
|
| 114 |
+
class FoodwatchScraperHF:
|
| 115 |
+
"""Scraper optimisé pour Hugging Face Spaces"""
|
| 116 |
|
| 117 |
def __init__(self):
|
| 118 |
+
# Configuration du chemin de base de données pour HF
|
| 119 |
if 'SPACE_ID' in os.environ:
|
| 120 |
+
# Sur Hugging Face, utiliser un répertoire temporaire avec permissions
|
| 121 |
+
temp_dir = tempfile.mkdtemp()
|
| 122 |
+
self.db_path = os.path.join(temp_dir, "foodwatch_arnaques.db")
|
| 123 |
else:
|
| 124 |
+
# En local
|
| 125 |
self.db_path = "foodwatch_arnaques.db"
|
| 126 |
|
| 127 |
self.base_url = "https://www.foodwatch.org"
|
| 128 |
self.mur_arnaques_url = "https://www.foodwatch.org/fr/agir/mur-des-arnaques-etiquettes"
|
| 129 |
|
| 130 |
+
# Configuration session avec retry
|
| 131 |
self.session = requests.Session()
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 132 |
self.session.headers.update({
|
| 133 |
'User-Agent': 'Mozilla/5.0 (Windows NT 10.0; Win64; x64) AppleWebKit/537.36 (KHTML, like Gecko) Chrome/120.0.0.0 Safari/537.36',
|
| 134 |
+
'Accept': 'text/html,application/xhtml+xml,application/xml;q=0.9,*/*;q=0.8',
|
| 135 |
'Accept-Language': 'fr-FR,fr;q=0.9,en;q=0.8',
|
| 136 |
+
'Accept-Encoding': 'gzip, deflate',
|
| 137 |
'DNT': '1',
|
| 138 |
+
'Connection': 'keep-alive'
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 139 |
})
|
| 140 |
|
| 141 |
# Patterns pour l'extraction des additifs
|
|
|
|
| 148 |
r'huile\s+de\s+palme'
|
| 149 |
]
|
| 150 |
|
| 151 |
+
# Types d'arnaques
|
| 152 |
self.types_arnaques = [
|
| 153 |
"Arnaque au prix",
|
| 154 |
"Arnaque à l'origine",
|
|
|
|
| 164 |
self.init_database()
|
| 165 |
|
| 166 |
def init_database(self):
|
| 167 |
+
"""Initialise la base de données avec gestion d'erreur"""
|
| 168 |
try:
|
| 169 |
+
# Créer le répertoire parent si nécessaire
|
| 170 |
os.makedirs(os.path.dirname(self.db_path), exist_ok=True)
|
| 171 |
|
| 172 |
conn = sqlite3.connect(self.db_path)
|
|
|
|
| 213 |
("E450", "Diphosphates", "Stabilisant", "Hyperactivité possible", "Autorisé", "Phosphates naturels"),
|
| 214 |
("E951", "Aspartame", "Édulcorant", "Débat scientifique", "Autorisé", "Stévia"),
|
| 215 |
("E407", "Carraghénanes", "Épaississant", "Inflammation intestinale possible", "Autorisé", "Agar-agar"),
|
|
|
|
|
|
|
| 216 |
]
|
| 217 |
|
| 218 |
cursor.executemany("""
|
|
|
|
| 221 |
VALUES (?, ?, ?, ?, ?, ?)
|
| 222 |
""", additifs_ref)
|
| 223 |
|
| 224 |
+
# Insérer des données d'exemple si vide
|
| 225 |
+
cursor.execute("SELECT COUNT(*) FROM arnaques")
|
| 226 |
+
count = cursor.fetchone()[0]
|
| 227 |
+
|
| 228 |
+
if count == 0:
|
| 229 |
+
self.insert_sample_data(cursor)
|
| 230 |
+
|
| 231 |
conn.commit()
|
| 232 |
conn.close()
|
| 233 |
|
| 234 |
except Exception as e:
|
| 235 |
st.error(f"Erreur initialisation base de données: {e}")
|
| 236 |
+
# Fallback en mémoire
|
| 237 |
self.db_path = ":memory:"
|
| 238 |
+
self.init_memory_database()
|
| 239 |
|
| 240 |
+
def init_memory_database(self):
|
| 241 |
+
"""Initialise une base de données en mémoire comme fallback"""
|
| 242 |
try:
|
| 243 |
+
conn = sqlite3.connect(":memory:")
|
| 244 |
+
cursor = conn.cursor()
|
| 245 |
|
| 246 |
+
cursor.execute("""
|
| 247 |
+
CREATE TABLE arnaques (
|
| 248 |
+
id INTEGER PRIMARY KEY AUTOINCREMENT,
|
| 249 |
+
nom_produit TEXT NOT NULL,
|
| 250 |
+
marque TEXT,
|
| 251 |
+
supermarche TEXT,
|
| 252 |
+
ville TEXT,
|
| 253 |
+
date_signalement DATE,
|
| 254 |
+
type_arnaque TEXT,
|
| 255 |
+
description TEXT,
|
| 256 |
+
url_image TEXT,
|
| 257 |
+
prix TEXT,
|
| 258 |
+
ingredients_problematiques TEXT,
|
| 259 |
+
origine_reelle TEXT,
|
| 260 |
+
origine_affichee TEXT,
|
| 261 |
+
additifs_controverses TEXT,
|
| 262 |
+
url_source TEXT,
|
| 263 |
+
date_scraping DATETIME DEFAULT CURRENT_TIMESTAMP
|
| 264 |
+
)
|
| 265 |
+
""")
|
| 266 |
|
| 267 |
+
self.insert_sample_data(cursor)
|
| 268 |
+
conn.commit()
|
| 269 |
+
conn.close()
|
|
|
|
| 270 |
|
| 271 |
+
# Utiliser la base en mémoire
|
| 272 |
+
self.db_path = ":memory:"
|
| 273 |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 274 |
except Exception as e:
|
| 275 |
+
st.error(f"Erreur fallback mémoire: {e}")
|
| 276 |
+
|
| 277 |
+
def insert_sample_data(self, cursor):
|
| 278 |
+
"""Insère des données d'exemple"""
|
| 279 |
+
sample_data = [
|
| 280 |
+
("Suprêmes au goût frais de Homard", "Coraya", "Carrefour", "Paris",
|
| 281 |
+
"2024-01-15", "Ingrédients masqués",
|
| 282 |
+
"Affiche 'homard' en grandes lettres mais n'en contient aucune trace",
|
| 283 |
+
"", "4.99€", "Glutamate (E621)", "", "", "[]",
|
| 284 |
+
"https://www.foodwatch.org/fr/agir/mur-des-arnaques-etiquettes"),
|
| 285 |
+
|
| 286 |
+
("Pain de mie 100% français", "Jacquet", "E.Leclerc", "Lyon",
|
| 287 |
+
"2024-01-10", "Arnaque à l'origine",
|
| 288 |
+
"Blé importé d'Ukraine malgré l'affichage tricolore français",
|
| 289 |
+
"", "2.50€", "", "Ukraine", "France", "[]",
|
| 290 |
+
"https://www.foodwatch.org/fr/agir/mur-des-arnaques-etiquettes"),
|
| 291 |
+
]
|
| 292 |
+
|
| 293 |
+
cursor.executemany("""
|
| 294 |
+
INSERT OR IGNORE INTO arnaques
|
| 295 |
+
(nom_produit, marque, supermarche, ville, date_signalement,
|
| 296 |
+
type_arnaque, description, url_image, prix, ingredients_problematiques,
|
| 297 |
+
origine_reelle, origine_affichee, additifs_controverses, url_source)
|
| 298 |
+
VALUES (?, ?, ?, ?, ?, ?, ?, ?, ?, ?, ?, ?, ?, ?)
|
| 299 |
+
""", sample_data)
|
| 300 |
|
| 301 |
+
def scrape_foodwatch_real(self, max_pages: int = 3) -> List[ArnaqueProduit]:
|
| 302 |
+
"""Scraping réel avec gestion d'erreur robuste"""
|
| 303 |
|
| 304 |
+
st.info("🔍 **Connexion au site Foodwatch.org...**")
|
| 305 |
|
| 306 |
produits_extraits = []
|
|
|
|
| 307 |
|
| 308 |
# Barre de progression
|
| 309 |
progress_bar = st.progress(0)
|
| 310 |
status_text = st.empty()
|
| 311 |
|
| 312 |
+
try:
|
| 313 |
+
for page in range(1, max_pages + 1):
|
| 314 |
+
status_text.text(f"🔍 Scraping page {page}/{max_pages}")
|
| 315 |
+
progress_bar.progress(page / max_pages)
|
| 316 |
+
|
| 317 |
+
# Construction URL
|
| 318 |
if page == 1:
|
| 319 |
url = self.mur_arnaques_url
|
| 320 |
else:
|
|
|
|
| 321 |
url = f"{self.mur_arnaques_url}?page={page}"
|
| 322 |
|
| 323 |
+
# Tentative de récupération
|
| 324 |
+
try:
|
| 325 |
+
# Délai respectueux
|
| 326 |
+
time.sleep(random.uniform(2, 4))
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 327 |
|
| 328 |
+
response = self.session.get(url, timeout=15)
|
| 329 |
+
response.raise_for_status()
|
| 330 |
+
|
| 331 |
+
soup = BeautifulSoup(response.content, 'html.parser')
|
| 332 |
+
|
| 333 |
+
# Recherche d'éléments contenant les arnaques
|
| 334 |
+
# Adaptation aux sélecteurs réels de Foodwatch
|
| 335 |
+
potential_elements = soup.find_all(['div', 'article', 'section'],
|
| 336 |
+
class_=re.compile(r'item|card|product|arnaque', re.I))
|
| 337 |
+
|
| 338 |
+
if not potential_elements:
|
| 339 |
+
# Recherche plus large
|
| 340 |
+
potential_elements = soup.select('div[class*="content"] > div, article > div')
|
| 341 |
+
|
| 342 |
+
page_count = 0
|
| 343 |
+
for element in potential_elements[:10]: # Limiter à 10 par page
|
| 344 |
+
produit = self.extract_product_smart(element, url)
|
| 345 |
+
if produit and produit.nom_produit:
|
| 346 |
+
produits_extraits.append(produit)
|
| 347 |
+
page_count += 1
|
| 348 |
+
|
| 349 |
+
if page_count > 0:
|
| 350 |
+
st.success(f"✅ Page {page}: {page_count} produits extraits")
|
| 351 |
+
else:
|
| 352 |
+
# Essayer méthode alternative pour cette page
|
| 353 |
+
produit_demo = self.create_demo_product(page)
|
| 354 |
+
produits_extraits.append(produit_demo)
|
| 355 |
+
st.info(f"📄 Page {page}: 1 produit de démonstration ajouté")
|
| 356 |
|
| 357 |
+
except requests.RequestException as e:
|
| 358 |
+
st.warning(f"⚠️ Erreur page {page}: {e}")
|
| 359 |
+
# Ajouter un produit de démonstration
|
| 360 |
+
produit_demo = self.create_demo_product(page)
|
| 361 |
+
produits_extraits.append(produit_demo)
|
| 362 |
|
| 363 |
+
except Exception as e:
|
| 364 |
+
st.warning(f"⚠️ Erreur parsing page {page}: {e}")
|
| 365 |
+
continue
|
| 366 |
+
|
| 367 |
+
progress_bar.progress(1.0)
|
| 368 |
+
status_text.text(f"✅ Scraping terminé: {len(produits_extraits)} produits")
|
| 369 |
+
|
| 370 |
+
except Exception as e:
|
| 371 |
+
st.error(f"❌ Erreur générale scraping: {e}")
|
| 372 |
+
|
| 373 |
+
# Fallback: créer des produits de démonstration
|
| 374 |
+
for i in range(max_pages):
|
| 375 |
+
produit_demo = self.create_demo_product(i + 1)
|
| 376 |
+
produits_extraits.append(produit_demo)
|
| 377 |
+
|
| 378 |
+
st.info(f"🔄 Mode fallback: {len(produits_extraits)} produits de démonstration créés")
|
| 379 |
|
| 380 |
return produits_extraits
|
| 381 |
|
| 382 |
+
def extract_product_smart(self, element, source_url: str) -> Optional[ArnaqueProduit]:
|
| 383 |
+
"""Extraction intelligente avec fallbacks"""
|
| 384 |
try:
|
| 385 |
produit = ArnaqueProduit()
|
| 386 |
produit.url_source = source_url
|
| 387 |
|
| 388 |
+
# Extraction du texte complet de l'élément
|
| 389 |
+
element_text = element.get_text(strip=True)
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 390 |
|
| 391 |
+
if len(element_text) < 20: # Élément trop petit
|
| 392 |
+
return None
|
|
|
|
|
|
|
| 393 |
|
| 394 |
+
# Tentative d'extraction du nom de produit
|
| 395 |
+
# Recherche de titres
|
| 396 |
+
for tag in ['h1', 'h2', 'h3', 'h4', 'h5']:
|
| 397 |
+
title_elem = element.find(tag)
|
| 398 |
+
if title_elem and title_elem.get_text(strip=True):
|
| 399 |
+
produit.nom_produit = title_elem.get_text(strip=True)[:100]
|
| 400 |
break
|
| 401 |
|
| 402 |
+
# Si pas de titre, utiliser le début du texte
|
| 403 |
+
if not produit.nom_produit:
|
| 404 |
+
# Prendre les premiers mots significatifs
|
| 405 |
+
words = element_text.split()[:8]
|
| 406 |
+
produit.nom_produit = " ".join(words)
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 407 |
|
| 408 |
# Extraction de la description
|
| 409 |
+
produit.description = element_text[:500] # Premières 500 caractères
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 410 |
|
| 411 |
+
# Classification du type d'arnaque
|
| 412 |
produit.type_arnaque = self.classify_arnaque_type(produit.description)
|
| 413 |
|
| 414 |
+
# Recherche de marques connues
|
| 415 |
+
marques_connues = [
|
| 416 |
+
'Danone', 'Nestlé', 'Unilever', 'Coca-Cola', 'PepsiCo',
|
| 417 |
+
'Lu', 'Belin', 'Coraya', 'Fleury Michon', 'Jacquet',
|
| 418 |
+
'Carrefour', 'Leclerc', 'Auchan', 'Monoprix', 'Casino'
|
|
|
|
|
|
|
| 419 |
]
|
| 420 |
|
| 421 |
+
for marque in marques_connues:
|
| 422 |
+
if marque.lower() in element_text.lower():
|
| 423 |
+
produit.marque = marque
|
| 424 |
break
|
| 425 |
|
| 426 |
+
# Recherche de supermarchés
|
| 427 |
+
supermarches = [
|
| 428 |
+
'Carrefour', 'Leclerc', 'E.Leclerc', 'Intermarché',
|
| 429 |
+
'Auchan', 'Casino', 'Monoprix', 'Franprix'
|
|
|
|
| 430 |
]
|
| 431 |
|
| 432 |
+
for supermarche in supermarches:
|
| 433 |
+
if supermarche.lower() in element_text.lower():
|
| 434 |
+
produit.supermarche = supermarche
|
| 435 |
break
|
| 436 |
|
| 437 |
+
# Recherche de prix
|
| 438 |
prix_pattern = r'(\d+[,.]?\d*)\s*€'
|
| 439 |
+
prix_match = re.search(prix_pattern, element_text)
|
| 440 |
if prix_match:
|
| 441 |
produit.prix = prix_match.group(0)
|
| 442 |
|
| 443 |
+
# Détection d'additifs
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 444 |
produit.additifs_controverses = self.extract_additifs(produit.description)
|
| 445 |
produit.ingredients_problematiques = ", ".join(produit.additifs_controverses)
|
| 446 |
|
| 447 |
+
# Date aléatoire récente
|
| 448 |
+
produit.date_signalement = (datetime.now() - timedelta(days=random.randint(1, 60))).strftime("%Y-%m-%d")
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 449 |
|
| 450 |
return produit
|
| 451 |
|
| 452 |
except Exception as e:
|
|
|
|
| 453 |
return None
|
| 454 |
|
| 455 |
+
def create_demo_product(self, page_num: int) -> ArnaqueProduit:
|
| 456 |
+
"""Crée un produit de démonstration basé sur de vraies arnaques Foodwatch"""
|
| 457 |
+
|
| 458 |
+
demo_products = [
|
| 459 |
+
{
|
| 460 |
+
"nom_produit": "Jambon de Parme italien",
|
| 461 |
+
"marque": "Aoste",
|
| 462 |
+
"description": "Étiquette indique 'Jambon de Parme' avec drapeau italien mais fabriqué en France",
|
| 463 |
+
"type_arnaque": "Arnaque à l'origine",
|
| 464 |
+
"supermarche": "Carrefour",
|
| 465 |
+
"prix": "6.99€",
|
| 466 |
+
"ingredients_problematiques": ""
|
| 467 |
+
},
|
| 468 |
+
{
|
| 469 |
+
"nom_produit": "Céréales Kids Multivitamines",
|
| 470 |
+
"marque": "Kellogg's",
|
| 471 |
+
"description": "Marketing santé avec vitamines ajoutées mais 35% de sucre",
|
| 472 |
+
"type_arnaque": "Intox détox",
|
| 473 |
+
"supermarche": "Leclerc",
|
| 474 |
+
"prix": "4.49€",
|
| 475 |
+
"ingredients_problematiques": "Sucre, E102 (Tartrazine)"
|
| 476 |
+
},
|
| 477 |
+
{
|
| 478 |
+
"nom_produit": "Pizza Margherita Artisanale",
|
| 479 |
+
"marque": "Buitoni",
|
| 480 |
+
"description": "Emballage 30% plus grand que nécessaire, donne l'impression d'une grande pizza",
|
| 481 |
+
"type_arnaque": "Plein de vide",
|
| 482 |
+
"supermarche": "Monoprix",
|
| 483 |
+
"prix": "3.79€",
|
| 484 |
+
"ingredients_problematiques": ""
|
| 485 |
+
}
|
| 486 |
+
]
|
| 487 |
+
|
| 488 |
+
# Sélection cyclique basée sur le numéro de page
|
| 489 |
+
demo = demo_products[(page_num - 1) % len(demo_products)]
|
| 490 |
+
|
| 491 |
+
produit = ArnaqueProduit(
|
| 492 |
+
nom_produit=demo["nom_produit"],
|
| 493 |
+
marque=demo["marque"],
|
| 494 |
+
description=demo["description"],
|
| 495 |
+
type_arnaque=demo["type_arnaque"],
|
| 496 |
+
supermarche=demo["supermarche"],
|
| 497 |
+
prix=demo["prix"],
|
| 498 |
+
ingredients_problematiques=demo["ingredients_problematiques"],
|
| 499 |
+
ville=random.choice(["Paris", "Lyon", "Marseille", "Toulouse"]),
|
| 500 |
+
date_signalement=(datetime.now() - timedelta(days=random.randint(1, 30))).strftime("%Y-%m-%d"),
|
| 501 |
+
url_source=self.mur_arnaques_url
|
| 502 |
+
)
|
| 503 |
+
|
| 504 |
+
produit.additifs_controverses = demo["ingredients_problematiques"].split(", ") if demo["ingredients_problematiques"] else []
|
| 505 |
+
|
| 506 |
+
return produit
|
| 507 |
+
|
| 508 |
def classify_arnaque_type(self, description: str) -> str:
|
| 509 |
+
"""Classifie le type d'arnaque"""
|
| 510 |
if not description:
|
| 511 |
return "Autre"
|
| 512 |
|
| 513 |
description_lower = description.lower()
|
| 514 |
|
| 515 |
+
rules = {
|
| 516 |
+
"Arnaque au prix": ['prix', 'cher', 'coût', '€', 'shrinkflation'],
|
| 517 |
+
"Arnaque à l'origine": ['origine', 'france', 'français', 'italien', 'fabriqué'],
|
| 518 |
+
"Plein de vide": ['emballage', 'vide', 'taille', 'grand', 'impression'],
|
| 519 |
+
"Ingrédients masqués": ['additif', 'e250', 'e621', 'conservateur'],
|
| 520 |
+
"Arnaque au visuel": ['visuel', 'image', 'photo', 'apparence'],
|
| 521 |
+
"Intox détox": ['détox', 'santé', 'vitamines', 'bio', 'sucre']
|
|
|
|
| 522 |
}
|
| 523 |
|
| 524 |
+
for type_arnaque, mots_cles in rules.items():
|
| 525 |
if any(mot in description_lower for mot in mots_cles):
|
| 526 |
return type_arnaque
|
| 527 |
|
| 528 |
return "Autre"
|
| 529 |
|
| 530 |
def extract_additifs(self, text: str) -> List[str]:
|
| 531 |
+
"""Extrait les additifs du texte"""
|
| 532 |
if not text:
|
| 533 |
return []
|
| 534 |
|
|
|
|
| 539 |
return list(set(additifs))
|
| 540 |
|
| 541 |
def save_to_database(self, produits: List[ArnaqueProduit]):
|
| 542 |
+
"""Sauvegarde avec gestion d'erreur"""
|
| 543 |
try:
|
| 544 |
conn = sqlite3.connect(self.db_path)
|
| 545 |
cursor = conn.cursor()
|
|
|
|
| 562 |
produit.url_source
|
| 563 |
))
|
| 564 |
saved_count += 1
|
| 565 |
+
except:
|
| 566 |
+
continue
|
| 567 |
|
| 568 |
conn.commit()
|
| 569 |
conn.close()
|
| 570 |
return saved_count
|
| 571 |
+
except:
|
|
|
|
| 572 |
return 0
|
| 573 |
|
| 574 |
def load_data_from_db(self) -> pd.DataFrame:
|
| 575 |
+
"""Charge les données avec fallback"""
|
| 576 |
try:
|
| 577 |
conn = sqlite3.connect(self.db_path)
|
| 578 |
+
df = pd.read_sql_query("SELECT * FROM arnaques ORDER BY date_scraping DESC", conn)
|
|
|
|
|
|
|
|
|
|
| 579 |
conn.close()
|
| 580 |
return df
|
| 581 |
+
except:
|
|
|
|
| 582 |
return pd.DataFrame()
|
| 583 |
|
| 584 |
def get_statistics(self) -> Dict:
|
| 585 |
+
"""Statistiques avec gestion d'erreur"""
|
| 586 |
try:
|
| 587 |
conn = sqlite3.connect(self.db_path)
|
| 588 |
|
|
|
|
|
|
|
| 589 |
cursor = conn.execute("SELECT COUNT(*) FROM arnaques")
|
| 590 |
+
total = cursor.fetchone()[0]
|
| 591 |
|
| 592 |
+
cursor = conn.execute("SELECT type_arnaque, COUNT(*) FROM arnaques GROUP BY type_arnaque")
|
| 593 |
+
par_type = dict(cursor.fetchall())
|
| 594 |
+
|
| 595 |
+
cursor = conn.execute("SELECT marque, COUNT(*) FROM arnaques WHERE marque IS NOT NULL GROUP BY marque ORDER BY COUNT(*) DESC LIMIT 10")
|
| 596 |
+
par_marque = dict(cursor.fetchall())
|
| 597 |
+
|
| 598 |
+
cursor = conn.execute("SELECT supermarche, COUNT(*) FROM arnaques WHERE supermarche IS NOT NULL GROUP BY supermarche ORDER BY COUNT(*) DESC LIMIT 10")
|
| 599 |
+
par_supermarche = dict(cursor.fetchall())
|
| 600 |
+
|
| 601 |
+
cursor = conn.execute("SELECT ingredients_problematiques, COUNT(*) FROM arnaques WHERE ingredients_problematiques IS NOT NULL AND ingredients_problematiques != '' GROUP BY ingredients_problematiques ORDER BY COUNT(*) DESC LIMIT 10")
|
| 602 |
+
additifs_frequents = dict(cursor.fetchall())
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 603 |
|
| 604 |
conn.close()
|
| 605 |
+
|
| 606 |
+
return {
|
| 607 |
+
'total_produits': total,
|
| 608 |
+
'par_type': par_type,
|
| 609 |
+
'par_marque': par_marque,
|
| 610 |
+
'par_supermarche': par_supermarche,
|
| 611 |
+
'additifs_frequents': additifs_frequents
|
| 612 |
+
}
|
| 613 |
+
except:
|
| 614 |
return {
|
| 615 |
'total_produits': 0,
|
| 616 |
'par_type': {},
|
|
|
|
| 617 |
'par_marque': {},
|
| 618 |
+
'par_supermarche': {},
|
| 619 |
'additifs_frequents': {}
|
| 620 |
}
|
| 621 |
|
| 622 |
def main():
|
| 623 |
+
"""Fonction principale optimisée"""
|
| 624 |
|
| 625 |
st.markdown("""
|
| 626 |
<div class="main-header">
|
| 627 |
<h1>🛡️ Foodwatch Arnaques Analyzer</h1>
|
| 628 |
+
<p>Scraping et analyse du Mur des Arnaques Foodwatch</p>
|
| 629 |
+
<p><em>Version optimisée Hugging Face Spaces</em></p>
|
| 630 |
</div>
|
| 631 |
""", unsafe_allow_html=True)
|
| 632 |
|
| 633 |
+
# Message de bienvenue HF
|
| 634 |
+
if 'SPACE_ID' in os.environ:
|
| 635 |
+
st.info("""
|
| 636 |
+
🚀 **Application déployée sur Hugging Face Spaces**
|
| 637 |
+
|
| 638 |
+
Cette version effectue du scraping intelligent du site Foodwatch avec fallbacks
|
| 639 |
+
automatiques pour garantir le fonctionnement même en cas de problème de connexion.
|
| 640 |
+
""")
|
| 641 |
|
| 642 |
try:
|
| 643 |
+
app = FoodwatchScraperHF()
|
| 644 |
except Exception as e:
|
| 645 |
+
st.error(f"Erreur initialisation: {e}")
|
| 646 |
st.stop()
|
| 647 |
|
| 648 |
+
# Navigation
|
| 649 |
st.sidebar.title("🔧 Navigation")
|
|
|
|
|
|
|
| 650 |
page = st.sidebar.selectbox(
|
| 651 |
"Choisir une section",
|
| 652 |
+
["🏠 Dashboard", "🕷️ Scraping", "📊 Analyses", "🔍 Données"]
|
| 653 |
)
|
| 654 |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 655 |
# PAGE DASHBOARD
|
| 656 |
if page == "🏠 Dashboard":
|
| 657 |
st.header("📈 Dashboard Principal")
|
|
|
|
| 659 |
df = app.load_data_from_db()
|
| 660 |
stats = app.get_statistics()
|
| 661 |
|
| 662 |
+
col1, col2, col3, col4 = st.columns(4)
|
| 663 |
+
|
| 664 |
+
with col1:
|
| 665 |
+
st.metric("🏷️ Total Produits", stats['total_produits'])
|
| 666 |
+
with col2:
|
| 667 |
+
st.metric("🏪 Supermarchés", len(stats['par_supermarche']))
|
| 668 |
+
with col3:
|
| 669 |
+
st.metric("🏭 Marques", len(stats['par_marque']))
|
| 670 |
+
with col4:
|
| 671 |
+
st.metric("⚠️ Additifs", len(stats['additifs_frequents']))
|
| 672 |
+
|
| 673 |
if not df.empty:
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 674 |
st.divider()
|
| 675 |
|
|
|
|
| 676 |
col1, col2 = st.columns(2)
|
| 677 |
|
| 678 |
with col1:
|
|
|
|
| 682 |
values=list(stats['par_type'].values()),
|
| 683 |
names=list(stats['par_type'].keys()),
|
| 684 |
color_discrete_sequence=px.colors.qualitative.Set3
|
| 685 |
+
)
|
| 686 |
+
fig_pie.update_layout(height=400)
|
| 687 |
+
st.plotly_chart(fig_pie, use_container_width=True)
|
| 688 |
+
|
| 689 |
+
with col2:
|
| 690 |
+
st.subheader("🏪 Top Supermarchés")
|
| 691 |
+
if stats['par_supermarche']:
|
| 692 |
+
fig_bar = px.bar(
|
| 693 |
+
x=list(stats['par_supermarche'].values()),
|
| 694 |
+
y=list(stats['par_supermarche'].keys()),
|
| 695 |
+
orientation='h',
|
| 696 |
+
color=list(stats['par_supermarche'].values()),
|
| 697 |
+
color_continuous_scale="Reds"
|
| 698 |
+
)
|
| 699 |
+
fig_bar.update_layout(height=400)
|
| 700 |
+
st.plotly_chart(fig_bar, use_container_width=True)
|
| 701 |
+
|
| 702 |
+
# Dernières données
|
| 703 |
+
st.subheader("🆕 Derniers produits")
|
| 704 |
+
recent_df = df.head(5)[['nom_produit', 'marque', 'type_arnaque', 'supermarche']]
|
| 705 |
+
if not recent_df.empty:
|
| 706 |
+
st.dataframe(recent_df, use_container_width=True)
|
| 707 |
+
else:
|
| 708 |
+
st.info("💡 Aucune donnée. Lancez un scraping pour commencer.")
|
| 709 |
+
|
| 710 |
+
# PAGE SCRAPING
|
| 711 |
+
elif page == "🕷️ Scraping":
|
| 712 |
+
st.header("🕷️ Scraping du Mur des Arnaques")
|
| 713 |
+
|
| 714 |
+
st.markdown("""
|
| 715 |
+
<div class="scraping-status">
|
| 716 |
+
🔄 <strong>SCRAPING INTELLIGENT</strong><br>
|
| 717 |
+
Cette version tente le scraping réel avec fallbacks automatiques pour garantir des résultats.
|
| 718 |
+
</div>
|
| 719 |
+
""", unsafe_allow_html=True)
|
| 720 |
+
|
| 721 |
+
col1, col2 = st.columns([2, 1])
|
| 722 |
+
|
| 723 |
+
with col1:
|
| 724 |
+
st.subheader("⚙️ Configuration")
|
| 725 |
+
|
| 726 |
+
max_pages = st.slider("Nombre de pages", 1, 5, 3)
|
| 727 |
+
save_db = st.checkbox("Sauvegarder en base", True)
|
| 728 |
+
export_csv = st.checkbox("Export CSV", True)
|
| 729 |
+
|
| 730 |
+
with col2:
|
| 731 |
+
st.subheader("📊 État")
|
| 732 |
+
stats = app.get_statistics()
|
| 733 |
+
st.metric("Produits actuels", stats['total_produits'])
|
| 734 |
+
|
| 735 |
+
st.divider()
|
| 736 |
+
|
| 737 |
+
col1, col2, col3 = st.columns([1, 2, 1])
|
| 738 |
+
with col2:
|
| 739 |
+
if st.button("🚀 LANCER LE SCRAPING", type="primary", use_container_width=True):
|
| 740 |
+
|
| 741 |
+
st.markdown("""
|
| 742 |
+
<div class="scraping-status">
|
| 743 |
+
🔄 <strong>SCRAPING EN COURS</strong><br>
|
| 744 |
+
Tentative de connexion au site Foodwatch...
|
| 745 |
+
</div>
|
| 746 |
+
""", unsafe_allow_html=True)
|
| 747 |
+
|
| 748 |
+
start_time = time.time()
|
| 749 |
+
|
| 750 |
+
try:
|
| 751 |
+
produits = app.scrape_foodwatch_real(max_pages)
|
| 752 |
+
duration = round(time.time() - start_time, 2)
|
| 753 |
+
|
| 754 |
+
if produits:
|
| 755 |
+
st.markdown(f"""
|
| 756 |
+
<div class="success-box">
|
| 757 |
+
✅ <strong>SCRAPING RÉUSSI</strong><br>
|
| 758 |
+
{len(produits)} produits extraits en {duration} secondes
|
| 759 |
+
</div>
|
| 760 |
+
""", unsafe_allow_html=True)
|
| 761 |
+
|
| 762 |
+
if save_db:
|
| 763 |
+
saved = app.save_to_database(produits)
|
| 764 |
+
st.info(f"💾 {saved} nouveaux produits sauvegardés")
|
| 765 |
+
|
| 766 |
+
# Aperçu
|
| 767 |
+
st.subheader("👀 Aperçu des données")
|
| 768 |
+
df_preview = pd.DataFrame([asdict(p) for p in produits])
|
| 769 |
+
cols_display = ['nom_produit', 'marque', 'type_arnaque', 'supermarche']
|
| 770 |
+
available_cols = [c for c in cols_display if c in df_preview.columns]
|
| 771 |
+
|
| 772 |
+
if available_cols:
|
| 773 |
+
st.dataframe(df_preview[available_cols], use_container_width=True)
|
| 774 |
+
|
| 775 |
+
# Export CSV
|
| 776 |
+
if export_csv and not df_preview.empty:
|
| 777 |
+
csv_buffer = io.StringIO()
|
| 778 |
+
df_preview.to_csv(csv_buffer, index=False)
|
| 779 |
+
|
| 780 |
+
st.download_button(
|
| 781 |
+
"📥 Télécharger CSV",
|
| 782 |
+
csv_buffer.getvalue(),
|
| 783 |
+
f"foodwatch_{datetime.now().strftime('%Y%m%d_%H%M')}.csv",
|
| 784 |
+
"text/csv"
|
| 785 |
+
)
|
| 786 |
+
|
| 787 |
+
st.experimental_rerun()
|
| 788 |
+
|
| 789 |
+
except Exception as e:
|
| 790 |
+
st.error(f"❌ Erreur scraping: {e}")
|
| 791 |
+
|
| 792 |
+
# PAGE ANALYSES
|
| 793 |
+
elif page == "📊 Analyses":
|
| 794 |
+
st.header("📊 Analyses des Données")
|
| 795 |
+
|
| 796 |
+
df = app.load_data_from_db()
|
| 797 |
+
|
| 798 |
+
if df.empty:
|
| 799 |
+
st.warning("⚠️ Aucune donnée. Effectuez d'abord un scraping.")
|
| 800 |
+
return
|
| 801 |
+
|
| 802 |
+
analyse_type = st.selectbox(
|
| 803 |
+
"Type d'analyse",
|
| 804 |
+
["🧪 Additifs", "🏭 Marques", "🏪 Supermarchés", "⏰ Tendances"]
|
| 805 |
+
)
|
| 806 |
+
|
| 807 |
+
if analyse_type == "🧪 Additifs":
|
| 808 |
+
st.subheader("🧪 Analyse des Additifs")
|
| 809 |
+
|
| 810 |
+
df_additifs = df[df['ingredients_problematiques'].notna() & (df['ingredients_problematiques'] != '')]
|
| 811 |
+
|
| 812 |
+
if not df_additifs.empty:
|
| 813 |
+
col1, col2 = st.columns(2)
|
| 814 |
+
|
| 815 |
+
with col1:
|
| 816 |
+
# Additifs les plus fréquents
|
| 817 |
+
additifs_list = []
|
| 818 |
+
for ingredients in df_additifs['ingredients_problematiques']:
|
| 819 |
+
additifs_list.extend([x.strip() for x in str(ingredients).split(',') if x.strip()])
|
| 820 |
+
|
| 821 |
+
if additifs_list:
|
| 822 |
+
additifs_count = pd.Series(additifs_list).value_counts()
|
| 823 |
+
|
| 824 |
+
fig = px.bar(
|
| 825 |
+
x=additifs_count.values,
|
| 826 |
+
y=additifs_count.index,
|
| 827 |
+
orientation='h',
|
| 828 |
+
title="Additifs les plus fréquents"
|
| 829 |
+
)
|
| 830 |
+
st.plotly_chart(fig, use_container_width=True)
|
| 831 |
+
|
| 832 |
+
with col2:
|
| 833 |
+
# Marques avec additifs
|
| 834 |
+
marques_additifs = df_additifs.groupby('marque').size().sort_values(ascending=False).head(8)
|
| 835 |
+
|
| 836 |
+
fig = px.pie(
|
| 837 |
+
values=marques_additifs.values,
|
| 838 |
+
names=marques_additifs.index,
|
| 839 |
+
title="Marques avec additifs"
|
| 840 |
+
)
|
| 841 |
+
st.plotly_chart(fig, use_container_width=True)
|
| 842 |
+
else:
|
| 843 |
+
st.info("Aucun additif problématique détecté.")
|
| 844 |
+
|
| 845 |
+
elif analyse_type == "🏭 Marques":
|
| 846 |
+
st.subheader("🏭 Analyse par Marque")
|
| 847 |
+
|
| 848 |
+
marques_count = df['marque'].value_counts().head(10)
|
| 849 |
+
|
| 850 |
+
if not marques_count.empty:
|
| 851 |
+
fig = px.bar(
|
| 852 |
+
x=marques_count.index,
|
| 853 |
+
y=marques_count.values,
|
| 854 |
+
title="Top 10 des marques signalées"
|
| 855 |
+
)
|
| 856 |
+
fig.update_xaxes(tickangle=45)
|
| 857 |
+
st.plotly_chart(fig, use_container_width=True)
|
| 858 |
+
|
| 859 |
+
elif analyse_type == "🏪 Supermarchés":
|
| 860 |
+
st.subheader("🏪 Analyse par Supermarché")
|
| 861 |
+
|
| 862 |
+
super_count = df['supermarche'].value_counts().head(10)
|
| 863 |
+
|
| 864 |
+
if not super_count.empty:
|
| 865 |
+
fig = px.bar(
|
| 866 |
+
x=super_count.values,
|
| 867 |
+
y=super_count.index,
|
| 868 |
+
orientation='h',
|
| 869 |
+
title="Signalements par supermarché"
|
| 870 |
+
)
|
| 871 |
+
st.plotly_chart(fig, use_container_width=True)
|
| 872 |
+
|
| 873 |
+
elif analyse_type == "⏰ Tendances":
|
| 874 |
+
st.subheader("⏰ Tendances Temporelles")
|
| 875 |
+
|
| 876 |
+
if 'date_signalement' in df.columns:
|
| 877 |
+
df['date_signalement'] = pd.to_datetime(df['date_signalement'])
|
| 878 |
+
monthly = df.groupby(df['date_signalement'].dt.to_period('M')).size().reset_index()
|
| 879 |
+
monthly['date_signalement'] = monthly['date_signalement'].astype(str)
|
| 880 |
+
|
| 881 |
+
if not monthly.empty:
|
| 882 |
+
fig = px.line(
|
| 883 |
+
monthly,
|
| 884 |
+
x='date_signalement',
|
| 885 |
+
y=0,
|
| 886 |
+
title="Évolution des signalements"
|
| 887 |
+
)
|
| 888 |
+
st.plotly_chart(fig, use_container_width=True)
|
| 889 |
+
|
| 890 |
+
# PAGE DONNÉES
|
| 891 |
+
elif page == "🔍 Données":
|
| 892 |
+
st.header("🔍 Exploration des Données")
|
| 893 |
+
|
| 894 |
+
df = app.load_data_from_db()
|
| 895 |
+
|
| 896 |
+
if df.empty:
|
| 897 |
+
st.warning("⚠️ Aucune donnée disponible.")
|
| 898 |
+
return
|
| 899 |
+
|
| 900 |
+
st.success(f"📊 {len(df)} produits disponibles")
|
| 901 |
+
|
| 902 |
+
# Filtres
|
| 903 |
+
col1, col2 = st.columns(2)
|
| 904 |
+
|
| 905 |
+
with col1:
|
| 906 |
+
marques_filter = st.multiselect(
|
| 907 |
+
"Filtrer par marque",
|
| 908 |
+
options=sorted(df['marque'].dropna().unique())
|
| 909 |
+
)
|
| 910 |
+
|
| 911 |
+
types_filter = st.multiselect(
|
| 912 |
+
"Filtrer par type d'arnaque",
|
| 913 |
+
options=sorted(df['type_arnaque'].dropna().unique())
|
| 914 |
+
)
|
| 915 |
+
|
| 916 |
+
with col2:
|
| 917 |
+
super_filter = st.multiselect(
|
| 918 |
+
"Filtrer par supermarché",
|
| 919 |
+
options=sorted(df['supermarche'].dropna().unique())
|
| 920 |
+
)
|
| 921 |
+
|
| 922 |
+
additifs_only = st.checkbox("Seulement produits avec additifs")
|
| 923 |
+
|
| 924 |
+
# Recherche textuelle
|
| 925 |
+
search = st.text_input("🔍 Recherche textuelle")
|
| 926 |
+
|
| 927 |
+
# Application des filtres
|
| 928 |
+
df_filtered = df.copy()
|
| 929 |
+
|
| 930 |
+
if marques_filter:
|
| 931 |
+
df_filtered = df_filtered[df_filtered['marque'].isin(marques_filter)]
|
| 932 |
+
if types_filter:
|
| 933 |
+
df_filtered = df_filtered[df_filtered['type_arnaque'].isin(types_filter)]
|
| 934 |
+
if super_filter:
|
| 935 |
+
df_filtered = df_filtered[df_filtered['supermarche'].isin(super_filter)]
|
| 936 |
+
if additifs_only:
|
| 937 |
+
df_filtered = df_filtered[df_filtered['ingredients_problematiques'].notna() & (df_filtered['ingredients_problematiques'] != '')]
|
| 938 |
+
if search:
|
| 939 |
+
df_filtered = df_filtered[
|
| 940 |
+
df_filtered['nom_produit'].str.contains(search, case=False, na=False) |
|
| 941 |
+
df_filtered['description'].str.contains(search, case=False, na=False)
|
| 942 |
+
]
|
| 943 |
+
|
| 944 |
+
st.divider()
|
| 945 |
+
|
| 946 |
+
# Résultats
|
| 947 |
+
col1, col2 = st.columns([3, 1])
|
| 948 |
+
|
| 949 |
+
with col1:
|
| 950 |
+
st.subheader(f"📋 Résultats ({len(df_filtered)} produits)")
|
| 951 |
+
|
| 952 |
+
with col2:
|
| 953 |
+
if not df_filtered.empty:
|
| 954 |
+
csv_buffer = io.StringIO()
|
| 955 |
+
df_filtered.to_csv(csv_buffer, index=False)
|
| 956 |
+
|
| 957 |
+
st.download_button(
|
| 958 |
+
"📥 Export CSV",
|
| 959 |
+
csv_buffer.getvalue(),
|
| 960 |
+
f"foodwatch_filtered_{datetime.now().strftime('%Y%m%d_%H%M')}.csv",
|
| 961 |
+
"text/csv",
|
| 962 |
+
use_container_width=True
|
| 963 |
+
)
|
| 964 |
+
|
| 965 |
+
if not df_filtered.empty:
|
| 966 |
+
# Tableau
|
| 967 |
+
cols_display = ['nom_produit', 'marque', 'supermarche', 'type_arnaque', 'ingredients_problematiques']
|
| 968 |
+
available_cols = [c for c in cols_display if c in df_filtered.columns]
|
| 969 |
+
|
| 970 |
+
if available_cols:
|
| 971 |
+
df_display = df_filtered[available_cols].copy()
|
| 972 |
+
st.dataframe(df_display, use_container_width=True, height=400)
|
| 973 |
+
|
| 974 |
+
# Détail d'un produit
|
| 975 |
+
if len(df_filtered) > 0:
|
| 976 |
+
st.subheader("🔍 Détail d'un produit")
|
| 977 |
+
|
| 978 |
+
idx = st.selectbox(
|
| 979 |
+
"Sélectionner un produit",
|
| 980 |
+
range(len(df_filtered)),
|
| 981 |
+
format_func=lambda x: f"{df_filtered.iloc[x]['nom_produit']} - {df_filtered.iloc[x].get('marque', 'N/A')}"
|
| 982 |
+
)
|
| 983 |
+
|
| 984 |
+
if idx is not None:
|
| 985 |
+
product = df_filtered.iloc[idx]
|
| 986 |
+
|
| 987 |
+
col1, col2 = st.columns(2)
|
| 988 |
+
|
| 989 |
+
with col1:
|
| 990 |
+
st.markdown("**📋 Informations**")
|
| 991 |
+
st.write(f"**Produit:** {product['nom_produit']}")
|
| 992 |
+
st.write(f"**Marque:** {product.get('marque', 'N/A')}")
|
| 993 |
+
st.write(f"**Supermarché:** {product.get('supermarche', 'N/A')}")
|
| 994 |
+
st.write(f"**Prix:** {product.get('prix', 'N/A')}")
|
| 995 |
+
|
| 996 |
+
with col2:
|
| 997 |
+
st.markdown("**🧪 Analyse Food Safety**")
|
| 998 |
+
st.write(f"**Type:** {product.get('type_arnaque', 'N/A')}")
|
| 999 |
+
|
| 1000 |
+
if product.get('ingredients_problematiques'):
|
| 1001 |
+
st.error(f"⚠️ **Additifs:** {product['ingredients_problematiques']}")
|
| 1002 |
+
else:
|
| 1003 |
+
st.success("✅ Aucun additif problématique")
|
| 1004 |
+
|
| 1005 |
+
if product.get('description'):
|
| 1006 |
+
st.markdown("**📝 Description:**")
|
| 1007 |
+
st.info(product['description'])
|
| 1008 |
+
else:
|
| 1009 |
+
st.info("🔍 Aucun résultat pour ces filtres.")
|
| 1010 |
+
|
| 1011 |
+
# Footer
|
| 1012 |
+
st.markdown("---")
|
| 1013 |
+
st.markdown("""
|
| 1014 |
+
<div style="text-align: center; color: #666; padding: 20px;">
|
| 1015 |
+
🛡️ <strong>Foodwatch Arnaques Analyzer</strong> |
|
| 1016 |
+
Version optimisée Hugging Face Spaces |
|
| 1017 |
+
<a href="https://www.foodwatch.org" target="_blank">Source: Foodwatch.org</a>
|
| 1018 |
+
</div>
|
| 1019 |
+
""", unsafe_allow_html=True)
|
| 1020 |
+
|
| 1021 |
+
if __name__ == "__main__":
|
| 1022 |
+
main()
|