Michael/Priceflow
1
0 Fork 0
mirror of https://github.com/R0m1k3/Priceflow.git synced 2026-10-24 15:48:11 +02:00
Code Issues Packages Projects Releases Wiki Activity

feat: Implement new product parsers for Stokomani, Auchan, Carrefour, and Gifi, and add a centralized search configuration module.

Browse Source
This commit is contained in:
Michael committed 2025-11-30 20:26:19 +01:00
1 parent e0d73f4d0e
commit b44082aeb4
5 files changed
+75 -50
Download Patch File Download Diff File Expand all files Collapse all files

No files matched your search

app/services/parsers/carrefour_parser.py
+13 -18
View File
@@ -14,16 +14,14 @@ class CarrefourParser(BaseParser):
# Carrefour products
# New container: div containing both image and title link
cards = soup.select("div.product-list-card-plp-grid-new, article")
if not cards:
cards = soup.select("div[class*='product-list-card']") # Fallback
cards = soup.select("div.product-list-card-plp-grid-new, article, div[class*='product-card']")
# If still no cards, try finding by link
# If no cards, try finding by link
if not cards:
links = soup.select("a.c-link.product-card-click-wrapper")
links = soup.select("a.c-link.product-card-click-wrapper, a[href*='/p/'], a[href*='/produit']")
cards = []
for link in links:
parent = link.find_parent("div", class_=lambda x: x and "product-list-card" in x)
parent = link.find_parent("div", class_=lambda x: x and "product" in x)
if parent:
cards.append(parent)
else:
@@ -53,21 +51,18 @@ class CarrefourParser(BaseParser):
price = None
# Price is often text node near h3 or in a specific price element
price_el = card.select_one("[class*='price'], .product-card-price")
price_el = card.select_one("[class*='price'], .product-card-price, span[class*='amount']")
if price_el:
price = self.parse_price_text(price_el.get_text())
if not price and title_el:
# Fallback: check siblings of title for text price
price_text = ""
for sibling in title_el.find_next_siblings():
if sibling.name is None: # Text node
price_text += sibling.strip() + " "
elif sibling.name in ['div', 'span']:
price_text += sibling.get_text(strip=True) + " "
if "€" in price_text:
break
price = self.parse_price_text(price_text)
if not price:
# Fallback: check all text in the card for a price pattern
text = card.get_text(" ", strip=True)
# Simple regex for price like 12,99 € or 12.99€
import re
match = re.search(r'(\d+[.,]\d{2})\s*€?', text)
if match:
price = self.parse_price_text(match.group(0))
results.append(ProductResult(
title=title,