mirror of
https://github.com/R0m1k3/Priceflow.git
synced 2026-10-12 01:39:25 +02:00
Claude
66d9046de2
**Problem:** L'Incroyable product showing "3€99" was detected as 399.00 EUR instead of 3.99 EUR. URL: https://www.lincroyable.fr/p95361-porte-magique-lutin/ **Root Cause:** Previous normalization only handled "3,99 €" (with comma separator). Some French sites use alternative format: "3€99" (euro symbol as separator). Existing regex: ```python re.sub(r'(\d+),(\d{2})\s*€', r'\1.\2 €', content) # Only matches "3,99 €" ``` Missed format: ``` "3€99" → No match → AI sees "399" → Returns 399.00 ``` **Solution: Add Pattern for Euro-as-Separator Format** 1. **Text Normalization** (browserless_service.py:408-411) ```python # NEW Pattern 1: digits€XX → digits.XX € content = re.sub(r'(\d+)€(\d{2})\b', r'\1.\2 €', content) # "3€99" → "3.99 €" # "19€50" → "19.50 €" # Pattern 2: digits,XX € → digits.XX € (existing) content = re.sub(r'(\d+),(\d{2})\s*€', r'\1.\2 €', content) # "3,99 €" → "3.99 €" ``` 2. **Extracted Price Normalization** (browserless_service.py:438-441) - Same patterns applied to directly extracted prices - Ensures consistency regardless of extraction method **Regex Details:** - `(\d+)€(\d{2})\b` matches: * `\d+` = one or more digits (euros) * `€` = euro symbol as separator * `(\d{2})` = exactly 2 digits (cents) * `\b` = word boundary (prevents matching "10€999") - Replacement: `\1.\2 €` = "digits.cents €" **Supported Formats After Fix:** ``` "3,99 €" → "3.99 €" ✓ (existing) "3€99" → "3.99 €" ✓ (NEW) "19,50 €" → "19.50 €" ✓ (existing) "19€50" → "19.50 €" ✓ (NEW) "1 234,56" → "1234.56" ✓ (existing) ``` **Flow:** ``` Page text: "Porte Magique 3€99" ↓ Normalization: "Porte Magique 3.99 €" ↓ Extracted: "3€99" → Normalized: "3.99 €" ↓ Sent to AI: "PRIX DÉTECTÉ: 3.99 €\n\nPorte Magique 3.99 €" ↓ AI sees: "3.99" (NOT "399") ↓ Returns: 3.99 ✓ ``` **Expected Results:** - Before: "3€99" → AI returns 399.00 ❌ - After: "3€99" → Normalized to "3.99 €" → AI returns 3.99 ✅ Fixes L'Incroyable and any other sites using euro-as-separator format. Complements existing comma-separator normalization.