Add sitemap generation to reflex app

Author: Kellprog

reflex/app.py

@@ -86,6 +86,7 @@
 from reflex.utils import codespaces, console, exceptions, format, prerequisites, types
 from reflex.utils.exec import is_prod_mode, is_testing_env, should_skip_compile
 from reflex.utils.imports import ImportVar
+from sitemap import serve_sitemap
 
 # Define custom types.
 ComponentCallable = Callable[[], Component]
@@ -270,6 +271,7 @@ def __init__(self, **kwargs):
             )
         super().__init__(**kwargs)
         base_state_subclasses = BaseState.__subclasses__()
+        self.add_sitemap()
 
         # Special case to allow test cases have multiple subclasses of rx.BaseState.
         if not is_testing_env() and len(base_state_subclasses) > 1:
@@ -302,6 +304,11 @@ def __init__(self, **kwargs):
             from reflex.utils.compat import windows_hot_reload_lifespan_hack
 
             self.register_lifespan_task(windows_hot_reload_lifespan_hack)
+    
+    def add_sitemap(self):
+        @self.api.get("/sitemap.xml")
+        async def sitemap():
+            return await serve_sitemap(self)
 
     def _enable_state(self) -> None:
         """Enable state for the app."""
@@ -462,6 +469,8 @@ def add_page(
             EventHandler | EventSpec | list[EventHandler | EventSpec] | None
         ) = None,
         meta: list[dict[str, str]] = constants.DefaultPage.META_LIST,
+        sitemap_priority: Optional[float] = None,
+        sitemap_changefreq: Optional[str] = None,
     ):
         """Add a page to the app.
 
@@ -476,6 +485,8 @@ def add_page(
             image: The image to display on the page.
             on_load: The event handler(s) that will be called each time the page load.
             meta: The metadata of the page.
+            sitemap_priority: sitemap priority
+            sitemap_changefreq: sitemap change frequency
 
         Raises:
             ValueError: When the specified route name already exists.
@@ -522,6 +533,19 @@ def add_page(
         if isinstance(component, tuple):
             component = Fragment.create(*component)
 
+        # Set explicit sitemap attributes if provided
+        if sitemap_priority is not None:
+            component.sitemap_priority = sitemap_priority
+        if sitemap_changefreq is not None:
+            component.sitemap_changefreq = sitemap_changefreq
+
+        # Auto-detect priority if not explicitly set
+        if not hasattr(component, 'sitemap_priority'):
+            component.sitemap_priority = self._auto_detect_priority(route)
+        
+        # Set default changefreq if not explicitly set
+        if not hasattr(component, 'sitemap_changefreq'):
+            component.sitemap_changefreq = "weekly"
         # Ensure state is enabled if this page uses state.
         if self.state is None:
             if on_load or component._has_stateful_event_triggers():
@@ -567,6 +591,14 @@ def add_page(
                 on_load = [on_load]
             self.load_events[route] = on_load
 
+
+    def _auto_detect_priority(self, route: Optional[str]) -> float:
+        """Auto detect sitemap priority"""
+        if route is None or route == "/" or route == "index":
+            return 1.0
+        depth = route.count("/")
+        return max(0.1, 1.0 - (depth * 0.2))
+    
     def get_load_events(self, route: str) -> list[EventHandler | EventSpec]:
         """Get the load events for a route.
 

reflex/sitemap.py

@@ -0,0 +1,62 @@
+from typing import Dict, Any, List
+import reflex as rx
+from reflex.components.component import Component
+from reflex.config import get_config
+from fastapi import Response
+from xml.etree.ElementTree import Element, SubElement, tostring
+from xml.dom import minidom
+
+def generate_xml(links: List[Dict[str, Any]]) -> str:
+    urlset = Element("urlset", xmlns="https://www.sitemaps.org/schemas/sitemap/0.9")
+    for link in links:
+        url = SubElement(urlset, "url")
+        loc = SubElement(url, "loc")
+        loc.text = link["loc"]
+        changefreq = SubElement(url, "changefreq")
+        changefreq.text = link["changefreq"]
+        priority = SubElement(url, "priority")
+        priority.text = str(link["priority"])
+    rough_string = tostring(urlset, "utf-8")
+    reparsed = minidom.parseString(rough_string)
+    return reparsed.toprettyxml(indent="  ")
+
+def generate_sitemap(app: rx.App) -> str:
+    links = []
+    for route, component in app.pages.items():
+        # Ignore dynamic routes for now
+        if "[" in route and "]" in route:
+            continue
+        
+        # Handle the index route
+        if route == "index":
+            route = "/"
+        
+        
+        # Check for explicit sitemap settings
+        sitemap_priority = getattr(component, "sitemap_priority", None)
+        sitemap_changefreq = getattr(component, "sitemap_changefreq", "weekly")
+        
+        # Calculate priority if not explicitly set
+        if sitemap_priority is None:
+            depth = route.count("/")
+            sitemap_priority = max(0.5, 1.0 - (depth * 0.1))
+        
+        links.append(
+            {
+                "loc": f"{{{{ BASE_URL}}}}{route}",
+                "changefreq": sitemap_changefreq,
+                "priority": sitemap_priority,
+            }
+        )
+    return generate_xml(links)
+
+async def serve_sitemap(app: rx.App) -> Response:
+    sitemap_content = generate_sitemap(app)
+    domain = get_config().deploy_url or "http://localhost:3000"
+    modified_sitemap = sitemap_content.replace("{{ BASE_URL }}", domain)
+    return Response(content=modified_sitemap, media_type="application/xml")
+
+def add_sitemap_to_app(app: rx.App):
+    @app.api.get("/sitemap.xml")
+    async def sitemap():
+        return await serve_sitemap(app)