allow skipping cache

Dockers/puppeteer-api/app/routes/browser.py

@@ -13,7 +13,7 @@ from app.services.browser import (
 router = APIRouter()
 
 @router.get("/")
-async def visit_url(url: str, x_api_key: Optional[str] = Header(None)):
+async def visit_url(url: str, skipCache: bool = False, x_api_key: Optional[str] = Header(None)):
     # Validate API key
     if not x_api_key or x_api_key != API_KEY:
         raise HTTPException(status_code=401, detail="Invalid API key")
@@ -21,16 +21,17 @@ async def visit_url(url: str, x_api_key: Optional[str] = Header(None)):
     # Decode URL if it's encoded
     decoded_url = unquote(url)
 
-    # Check cache first
-    cached_result = get_cached_data(decoded_url, "visit")
-    if cached_result:
-        return cached_result
+    # Check cache first (unless skipCache is True)
+    if not skipCache:
+        cached_result = get_cached_data(decoded_url, "visit")
+        if cached_result:
+            return cached_result
 
     try:
         # Call the service function
         result = await visit_url_service(decoded_url)
 
-        # Save to cache
+        # Save to cache (even if we skipped reading from it)
         save_to_cache(decoded_url, "visit", result)
 
         return result
@@ -39,7 +40,7 @@ async def visit_url(url: str, x_api_key: Optional[str] = Header(None)):
         raise HTTPException(status_code=500, detail=str(e))
 
 @router.get("/seo")
-async def extract_seo(url: str, x_api_key: Optional[str] = Header(None)):
+async def extract_seo(url: str, skipCache: bool = False, x_api_key: Optional[str] = Header(None)):
     """Extract SEO information from a website"""
     # Validate API key
     if not x_api_key or x_api_key != API_KEY:
@@ -48,10 +49,11 @@ async def extract_seo(url: str, x_api_key: Optional[str] = Header(None)):
     # Decode URL if it's encoded
     decoded_url = unquote(url)
 
-    # Check cache first
-    cached_result = get_cached_data(decoded_url, "seo")
-    if cached_result:
-        return cached_result
+    # Check cache first (unless skipCache is True)
+    if not skipCache:
+        cached_result = get_cached_data(decoded_url, "seo")
+        if cached_result:
+            return cached_result
 
     try:
         # Call the service function
@@ -65,7 +67,7 @@ async def extract_seo(url: str, x_api_key: Optional[str] = Header(None)):
         raise HTTPException(status_code=500, detail=str(e))
 
 @router.get("/meta")
-async def extract_meta_tags(url: str, x_api_key: Optional[str] = Header(None)):
+async def extract_meta_tags(url: str, skipCache: bool = False, x_api_key: Optional[str] = Header(None)):
     """Extract meta tags from a website"""
     # Validate API key
     if not x_api_key or x_api_key != API_KEY:
@@ -74,10 +76,11 @@ async def extract_meta_tags(url: str, x_api_key: Optional[str] = Header(None)):
     # Decode URL if it's encoded
     decoded_url = unquote(url)
 
-    # Check cache first
-    cached_result = get_cached_data(decoded_url, "meta")
-    if cached_result:
-        return cached_result
+    # Check cache first (unless skipCache is True)
+    if not skipCache:
+        cached_result = get_cached_data(decoded_url, "meta")
+        if cached_result:
+            return cached_result
 
     try:
         # Call the service function
@@ -91,7 +94,7 @@ async def extract_meta_tags(url: str, x_api_key: Optional[str] = Header(None)):
         raise HTTPException(status_code=500, detail=str(e))
 
 @router.get("/pagination")
-async def detect_pagination(url: str, x_api_key: Optional[str] = Header(None)):
+async def detect_pagination(url: str, skipCache: bool = False, x_api_key: Optional[str] = Header(None)):
     """Detect pagination on a website and determine the pagination pattern"""
     # Validate API key
     if not x_api_key or x_api_key != API_KEY:
@@ -100,10 +103,11 @@ async def detect_pagination(url: str, x_api_key: Optional[str] = Header(None)):
     # Decode URL if it's encoded
     decoded_url = unquote(url)
 
-    # Check cache first
-    cached_result = get_cached_data(decoded_url, "pagination")
-    if cached_result:
-        return cached_result
+    # Check cache first (unless skipCache is True)
+    if not skipCache:
+        cached_result = get_cached_data(decoded_url, "pagination")
+        if cached_result:
+            return cached_result
 
     try:
         # Call the service function