sitemap and its test methods

Author: bhatia2akshit

reflex/app.py

@@ -20,7 +20,7 @@
 from types import SimpleNamespace
 from typing import TYPE_CHECKING, Any, BinaryIO, get_args, get_type_hints
 
-from fastapi import FastAPI, HTTPException, Request
+from fastapi import FastAPI, HTTPException, Request, Response
 from fastapi import UploadFile as FastAPIUploadFile
 from fastapi.middleware import cors
 from fastapi.responses import JSONResponse, StreamingResponse
@@ -80,6 +80,7 @@
     replace_brackets_with_keywords,
     verify_route_validity,
 )
+from reflex.sitemap import generate_sitemaps, read_sitemap_file
 from reflex.state import (
     BaseState,
     RouterData,
@@ -423,6 +424,8 @@ class App(MiddlewareMixin, LifespanMixin):
     # Put the toast provider in the app wrap.
     toaster: Component | None = dataclasses.field(default_factory=toast.provider)
 
+    _sitemap_properties: dict[str, dict] = dataclasses.field(default_factory=dict)
+
     @property
     def api(self) -> FastAPI | None:
         """Get the backend api.
@@ -617,6 +620,19 @@ def _add_default_endpoints(self):
 
         self.api.get(str(constants.Endpoint.PING))(ping)
         self.api.get(str(constants.Endpoint.HEALTH))(health)
+        self.api.get(str(constants.Endpoint.SITEMAP))(self.serve_sitemap)
+
+    async def serve_sitemap(self) -> Response:
+        """Asynchronously serve the sitemap as an XML response.
+
+        This function checks if a sitemap.xml file exists in the root directory of the app. If so, this file is served
+        as a Response. Otherwise, a new sitemap is generated and saved to sitemap.xml before being served.
+
+        Returns:
+            Response: An HTTP response with the XML sitemap content and the media type set to "application/xml".
+        """
+        sitemaps = read_sitemap_file()
+        return Response(content=sitemaps, media_type="application/xml")
 
     def _add_optional_endpoints(self):
         """Add optional api endpoints (_upload)."""
@@ -695,6 +711,8 @@ def add_page(
         on_load: EventType[()] | None = None,
         meta: list[dict[str, str]] = constants.DefaultPage.META_LIST,
         context: dict[str, Any] | None = None,
+        sitemap_priority: float = constants.DefaultPage.SITEMAP_PRIORITY,
+        sitemap_changefreq: str = constants.DefaultPage.SITEMAP_CHANGEFREQ,
     ):
         """Add a page to the app.
 
@@ -710,6 +728,9 @@ def add_page(
             on_load: The event handler(s) that will be called each time the page load.
             meta: The metadata of the page.
             context: Values passed to page for custom page-specific logic.
+            sitemap_priority: The priority of the page in the sitemap. If None, the priority is calculated based on the
+            depth of the route.
+            sitemap_changefreq: The change frequency of the page in the sitemap. Default to 'weekly'
 
         Raises:
             PageValueError: When the component is not set for a non-404 page.
@@ -775,7 +796,7 @@ def add_page(
                 )
 
         # Setup dynamic args for the route.
-        # this state assignment is only required for tests using the deprecated state kwarg for App
+        # This state assignment is only required for tests using the deprecated state kwarg for App
         state = self._state if self._state else State
         state.setup_dynamic_args(get_route_args(route))
 
@@ -784,6 +805,11 @@ def add_page(
                 on_load if isinstance(on_load, list) else [on_load]
             )
 
+        self._sitemap_properties[route] = {
+            "priority": sitemap_priority,
+            "changefreq": sitemap_changefreq,
+        }
+
         self._unevaluated_pages[route] = unevaluated_page
 
     def _compile_page(self, route: str, save_page: bool = True):
@@ -811,6 +837,10 @@ def _compile_page(self, route: str, save_page: bool = True):
         if save_page:
             self._pages[route] = component
 
+    def get_sitemap_properties(self) -> dict[str, dict]:
+        """Get the sitemap properties."""
+        return self._sitemap_properties
+
     def get_load_events(self, route: str) -> list[IndividualEventType[()]]:
         """Get the load events for a route.
 
@@ -1103,6 +1133,9 @@ def _compile(self, export: bool = False):
 
         self._pages = {}
 
+        # generate sitemaps from sitemap properties
+        generate_sitemaps(self._sitemap_properties)
+
         def get_compilation_time() -> str:
             return str(datetime.now().time()).split(".")[0]
 

reflex/constants/event.py

@@ -13,6 +13,7 @@ class Endpoint(Enum):
     AUTH_CODESPACE = "auth-codespace"
     HEALTH = "_health"
     ALL_ROUTES = "_all_routes"
+    SITEMAP = "sitemap.xml"
 
     def __str__(self) -> str:
         """Get the string representation of the endpoint.

reflex/constants/route.py

@@ -60,6 +60,10 @@ class DefaultPage(SimpleNamespace):
     IMAGE = "favicon.ico"
     # The default meta list to show for Reflex apps.
     META_LIST = []
+    # The default changefrequency for sitemap generation.
+    SITEMAP_CHANGEFREQ = "weekly"
+    # The default priority for sitemap generation.
+    SITEMAP_PRIORITY = 10.0
 
 
 # 404 variables

reflex/sitemap.py

@@ -0,0 +1,146 @@
+"""This module contains functions to generate and manage the sitemap.xml file."""
+
+from pathlib import Path
+from typing import Dict, List
+from xml.dom import minidom
+from xml.etree.ElementTree import Element, SubElement, tostring
+
+from reflex import constants
+from reflex.config import get_config
+from reflex.utils import prerequisites
+
+# _static folder in the .web directory containing the sitemap.xml file.
+_sitemap_folder_path: Path = (
+    Path.cwd() / prerequisites.get_web_dir() / constants.Dirs.STATIC
+)
+
+# sitemap file path
+_sitemap_file_path: Path = _sitemap_folder_path / "sitemap.xml"
+
+
+def check_sitemap_file_exists() -> bool:
+    """Check if the sitemap file exists.
+
+    Returns:
+        bool: True if the sitemap file exists in the .web/_static folder.
+    """
+    return _sitemap_folder_path.exists() & _sitemap_file_path.exists()
+
+
+def read_sitemap_file() -> str:
+    """Read the sitemap file.
+
+    Returns:
+        str: The contents of the sitemap file.
+    """
+    with _sitemap_file_path.open("r") as f:
+        return f.read()
+
+
+def generate_xml(links: List[Dict[str, str]]) -> str:
+    """Generate an XML sitemap from a list of links.
+
+    Args:
+        links (List[Dict[str, Any]]): A list of dictionaries where each dictionary contains
+            'loc' (URL of the page), 'changefreq' (frequency of changes), and 'priority' (priority of the page).
+
+    Returns:
+        str: A pretty-printed XML string representing the sitemap.
+    """
+    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_sitemaps(sitemap_config: Dict[str, Dict[str, str]]) -> None:
+    """Generate the sitemap.xml file.
+
+    This function generates the sitemap.xml file by crawling through the available pages in the app and generating a list
+    of links with their respective sitemap properties such as location (URL), change frequency, and priority. Dynamic
+    routes and the 404 page are excluded from the sitemap.
+
+    Args:
+        sitemap_config: A dictionary containing the sitemap properties for each route.
+    """
+    links = generate_links_for_sitemap(sitemap_config)
+    generate_static_sitemap(links)
+
+
+def generate_links_for_sitemap(
+    sitemap_config: Dict[str, Dict[str, str]],
+) -> List[dict[str, str]]:
+    """Generate a list of links for which sitemaps are generated.
+
+    This function loops through sitemap_config and generates a list of links with their respective sitemap properties
+    such as location (URL), change frequency, and priority. Dynamic routes and the 404 page are excluded from the
+    sitemap.
+
+    Args:
+        sitemap_config: A dictionary containing the sitemap properties for each route.
+
+    Returns:
+        List: A list of dictionaries where each dictionary contains the 'loc' (URL of the page), 'priority' and
+        'changefreq' of each route.
+    """
+    links = []
+
+    # find link of pages that are not dynamically created.
+    for route in sitemap_config:
+        # Ignore dynamic routes and 404
+        if ("[" in route and "]" in route) or route == "404":
+            continue
+
+        sitemap_changefreq = sitemap_config[route]["changefreq"]
+        sitemap_priority = sitemap_config[route]["priority"]
+
+        # Handle the index route
+        if route == "index":
+            route = "/"
+
+        if not route.startswith("/"):
+            route = f"/{route}"
+
+        if (
+            sitemap_priority == constants.DefaultPage.SITEMAP_PRIORITY
+        ):  # indicates that user didn't set priority
+            depth = route.count("/")
+            sitemap_priority = max(0.5, 1.0 - (depth * 0.1))
+
+        deploy_url = get_config().deploy_url  # pick domain url from the config file.
+
+        links.append(
+            {
+                "loc": f"{deploy_url}{route}",
+                "changefreq": sitemap_changefreq,
+                "priority": sitemap_priority,
+            }
+        )
+    return links
+
+
+def generate_static_sitemap(links: List[Dict[str, str]]) -> None:
+    """Generates the sitemaps for the pages stored in _pages. Store it in sitemap.xml.
+
+    This method is called from two methods:
+     1. Every time the web app is deployed onto the server.
+     2. When the user (or crawler) requests for the sitemap.xml file.
+
+    Args:
+        links: The list of urls for which the sitemap is to be generated.
+    """
+    sitemap = generate_xml(links)
+    Path(_sitemap_folder_path).mkdir(parents=True, exist_ok=True)
+
+    # this method is only called when old sitemap.xml is not retrieved. So we can safely replace an already existing xml
+    # file.
+    with _sitemap_file_path.open("w") as f:
+        f.write(sitemap)

reflex/utils/build.py

@@ -203,13 +203,6 @@ def build(
         "Collecting build traces",
     ]
 
-    # Generate a sitemap if a deploy URL is provided.
-    if deploy_url is not None:
-        generate_sitemap_config(deploy_url, export=for_export)
-        command = "export-sitemap"
-
-        checkpoints.extend(["Loading next-sitemap", "Generation completed"])
-
     # Start the subprocess with the progress bar.
     process = processes.new_process(
         [*prerequisites.get_js_package_executor(raise_on_none=True)[0], "run", command],

tests/units/test_app.py

@@ -372,6 +372,23 @@ def test_add_duplicate_page_route_error(app: App, first_page, second_page, route
         app.add_page(second_page, route="/" + route.strip("/") if route else None)
 
 
+def test_add_page_with_sitemap_properties(app):
+    """Test if the sitemap properties of the app instance is set properly or not."""
+    # check with given values.
+    app.add_page(
+        page1, route="/page1", sitemap_priority=0.9, sitemap_changefreq="daily"
+    )
+    assert app._sitemap_properties["page1"] == {"priority": 0.9, "changefreq": "daily"}
+
+    # check default values added.
+    app.add_page(page2, route="/page2")
+    print(app._sitemap_properties)
+    assert app._sitemap_properties["page2"] == {
+        "priority": 10.0,
+        "changefreq": "weekly",
+    }
+
+
 def test_initialize_with_admin_dashboard(test_model):
     """Test setting the admin dashboard of an app.