From ccc9c0080eba782a0c951dac029dae528ab30cf5 Mon Sep 17 00:00:00 2001
From: Claude <noreply@anthropic.com>
Date: Sat, 22 Nov 2025 15:09:50 +0000
Subject: [PATCH] fix: Fix AI service call and domain URL building

- Use AIService.analyze_image() class method instead of module function
- Handle AIExtractionResponse tuple return type properly
- Add _clean_domain() to strip protocol/www from stored domains
- Fixes "https://https://" URL duplication issue
---
 app/services/direct_search_service.py | 28 ++++++++++++++++++++-
 app/services/search_service.py        | 35 +++++++++++++++------------
 2 files changed, 46 insertions(+), 17 deletions(-)

diff --git a/app/services/direct_search_service.py b/app/services/direct_search_service.py
index 0af587a..044b64c 100644
--- a/app/services/direct_search_service.py
+++ b/app/services/direct_search_service.py
@@ -118,7 +118,10 @@ async def _search_site_browserless(
     timeout: float,
 ) -> list[SearchResult]:
     """Recherche sur un site via Browserless (Playwright)"""
-    domain = site.get("domain", "").lower()
+    raw_domain = site.get("domain", "").lower()
+    # Nettoyer le domaine (enlever protocole, www, slash final)
+    domain = _clean_domain(raw_domain)
+
     search_url = site.get("search_url") or _get_default_search_url(domain)
     product_selector = site.get("product_link_selector") or _get_default_product_selector(domain)
     wait_selector = _get_default_wait_selector(domain)
@@ -358,6 +361,29 @@ def _extract_domain(url: str) -> str:
         return ""
 
 
+def _clean_domain(domain: str) -> str:
+    """Nettoie un domaine (enlève protocole, www, slash final)"""
+    if not domain:
+        return ""
+
+    domain = domain.strip().lower()
+
+    # Enlever le protocole
+    if domain.startswith("https://"):
+        domain = domain[8:]
+    elif domain.startswith("http://"):
+        domain = domain[7:]
+
+    # Enlever www.
+    if domain.startswith("www."):
+        domain = domain[4:]
+
+    # Enlever le slash final
+    domain = domain.rstrip("/")
+
+    return domain
+
+
 def _is_non_product_url(url: str) -> bool:
     """Vérifie si l'URL n'est pas une page produit"""
     non_product_patterns = [
diff --git a/app/services/search_service.py b/app/services/search_service.py
index 0833ffb..6a68e86 100644
--- a/app/services/search_service.py
+++ b/app/services/search_service.py
@@ -12,7 +12,8 @@ from sqlalchemy.orm import Session
 
 from app.models import SearchSite
 from app.schemas import SearchProgress, SearchResultItem
-from app.services import ai_service, direct_search_service, light_scraper_service
+from app.services import direct_search_service, light_scraper_service
+from app.services.ai_service import AIService
 from app.services.scraper_service import ScraperService
 from app.services.settings_service import SettingsService
 
@@ -227,23 +228,25 @@ async def _scrape_with_browserless(
             )
 
         # Extraction IA
-        ai_result = await ai_service.analyze_image(
-            screenshot_path=screenshot_path,
-            webpage_text=page_text,
+        ai_result = await AIService.analyze_image(
+            image_path=screenshot_path,
+            page_text=page_text,
         )
 
-        if ai_result and ai_result.get("price"):
-            return SearchResultItem(
-                url=url,
-                title=fallback_title,
-                price=ai_result.get("price"),
-                currency=ai_result.get("currency", "EUR"),
-                in_stock=ai_result.get("in_stock"),
-                image_url=None,  # Pas d'image depuis le screenshot
-                site_name=site_name,
-                site_domain=domain,
-                confidence=ai_result.get("price_confidence", 0.5),
-            )
+        if ai_result:
+            extraction, metadata = ai_result
+            if extraction and extraction.price is not None:
+                return SearchResultItem(
+                    url=url,
+                    title=fallback_title,
+                    price=extraction.price,
+                    currency=extraction.currency or "EUR",
+                    in_stock=extraction.in_stock,
+                    image_url=None,
+                    site_name=site_name,
+                    site_domain=domain,
+                    confidence=extraction.price_confidence or 0.5,
+                )
 
         return SearchResultItem(
             url=url,