only drop original if there is lazy data

vizro-core/examples/default/app.py

@@ -14,14 +14,24 @@ def retrieve_gapminder():
 
 
 data_manager["gapminder"] = retrieve_gapminder
+data_manager["gapminder2"] = retrieve_gapminder
+
+df_gapminder = px.data.gapminder()
+df_gapminder2 = px.data.gapminder()
 
 
 @capture("action")
 def delete_memoized_cache(delete_button_id_n_clicks):
-    """Delete memoized cache."""
+    """Delete one memoized cache."""
     if delete_button_id_n_clicks:
         data_manager._cache.delete_memoized(data_manager._get_original_data, data_manager, "gapminder")
-        return None
+
+
+@capture("action")
+def empty_cache(empty_button_id_n_clicks):
+    """Empty the entire cache."""
+    if empty_button_id_n_clicks:
+        data_manager._cache.cache.clear()
 
 
 page = vm.Page(
@@ -30,38 +40,66 @@ def delete_memoized_cache(delete_button_id_n_clicks):
         vm.Graph(
             figure=px.box(
                 "gapminder",
+                # df_gapminder,
                 x="continent",
                 y="lifeExp",
                 color="continent",
                 title="Distribution per continent",
             ),
             id="the_graph",
         ),
-        # vm.Graph(
-        #     figure=px.box(
-        #         "gapminder",
-        #         x="continent",
-        #         y="lifeExp",
-        #         color="continent",
-        #         title="Distribution per continent",
-        #     ),
-        #     id="the_graph2",
-        # ),
+        vm.Graph(
+            id="the_graph2",
+            figure=px.scatter(
+                "gapminder2",
+                # df_gapminder2,
+                x="gdpPercap",
+                y="lifeExp",
+                size="pop",
+                color="continent",
+                hover_name="country",
+                facet_col="continent",
+                labels={
+                    "gdpPercap": "GDP per capita",
+                    "pop": "Population",
+                    "lifeExp": "Life expectancy",
+                    "continent": "Continent",
+                },
+                range_y=[25, 90],
+                color_discrete_map={
+                    "Africa": "#00b4ff",
+                    "Americas": "#ff9222",
+                    "Asia": "#3949ab",
+                    "Europe": "#ff5267",
+                    "Oceania": "#08bdba",
+                },
+            ),
+        ),
         vm.Button(
             id="delete_button_id",
             text="delete_memoized_cache",
             actions=[
                 vm.Action(
                     function=delete_memoized_cache(),
                     inputs=["delete_button_id.n_clicks"],
-                    outputs=[],
+                )
+            ]
+        ),
+        vm.Button(
+            id="empty_button_id",
+            text="empty_cache",
+            actions=[
+                vm.Action(
+                    function=empty_cache(),
+                    inputs=["empty_button_id.n_clicks"],
                 )
             ]
         )
     ],
     controls=[
         vm.Filter(
             column="year",
+            targets=["the_graph"],
             selector=vm.RangeSlider(
                 title="Select timeframe",
             ),
@@ -93,7 +131,7 @@ def delete_memoized_cache(delete_button_id_n_clicks):
     #     "CACHE_THRESHOLD": 20,  # The maximum number of items the cache can hold
     #     "CACHE_DEFAULT_TIMEOUT": 3000,  # Unit of time is seconds
     # }
-    data_manager._cache.config = {"CACHE_TYPE": "RedisCache", "CACHE_REDIS_URL": "redis://localhost:6379/0", "CACHE_DEFAULT_TIMEOUT": 3000,}
+    data_manager._cache.config = {"CACHE_TYPE": "RedisCache", "CACHE_REDIS_URL": "redis://localhost:6379/0", "CACHE_DEFAULT_TIMEOUT": 60,}
     Vizro().build(dashboard).run()
     # Vizro().build(dashboard).run(
     #     threaded=False,

vizro-core/src/vizro/_vizro.py

@@ -47,10 +47,10 @@ def build(self, dashboard: Dashboard):
         data_manager._cache.init_app(self.dash.server, config=data_manager._cache.config)
         # Note that model instantiation and pre_build are independent of Dash.
         self._pre_build()
-        # to clear original data if the cache type is not NullCache
-        data_manager._drop_original_data()
 
         self.dash.layout = dashboard.build()
+        # to clear original data if the cache type is not NullCache
+        data_manager._clean_original_data()
 
         return self
 

vizro-core/src/vizro/managers/_data_manager.py

@@ -89,7 +89,12 @@ def _get_component_data(self, component_id: ComponentID) -> pd.DataFrame:
             raise KeyError(f"Component {component_id} does not exist. You need to call add_component first.")
         dataset_name = self.__component_to_original[component_id]
 
-        return self._get_original_data(dataset_name)
+        component_data = self._get_original_data(dataset_name)
+
+        # clean up original data if the cache type is not NullCache
+        self._clean_original_data()
+
+        return component_data
 
     def _has_registered_data(self, component_id: ComponentID) -> bool:
         try:
@@ -102,13 +107,17 @@ def _clear(self):
         self.__init__()  # type: ignore[misc]
 
     # to clear original data if the cache type is not NullCache
-    def _drop_original_data(self):
-        logger.debug(f"__original_data: {self.__original_data}")
-        logger.debug(f"config: {self._cache.config}")
-        logger.debug("drop original data")
-        if self._cache.config["CACHE_TYPE"] != "NullCache":
-            self.__original_data = {}
-        logger.debug(f"__original_data: {self.__original_data}")
+    def _clean_original_data(self):
+        logger.debug(f"__original_data before cleaning: {self.__original_data.keys()}")
+        # logger.debug(f"config: {self._cache.config}")
+        logger.debug("clean original data")
+        if self._cache.config["CACHE_TYPE"] == "NullCache":
+            return
+        for dataset_name in list(self.__original_data.keys()):
+            if dataset_name in self.__lazy_data:
+                logger.debug(f"drop --> {dataset_name}")
+                del self.__original_data[dataset_name]
+        logger.debug(f"__original_data after cleaning: {self.__original_data.keys()}")
 
 
 data_manager = DataManager()