fix dataset update

.changeset/soft-worms-remain.md

@@ -0,0 +1,7 @@
+---
+"@gradio/dataset": patch
+"gradio": patch
+"website": patch
+---
+
+fix:fix dataset update

gradio/blocks.py

@@ -1723,12 +1723,12 @@ async def postprocess_data(
                 ) from err
 
             if block.stateful:
-                if not utils.is_update(predictions[i]):
+                if not utils.is_prop_update(predictions[i]):
                     state[block._id] = predictions[i]
                 output.append(None)
             else:
                 prediction_value = predictions[i]
-                if utils.is_update(
+                if utils.is_prop_update(
                     prediction_value
                 ):  # if update is passed directly (deprecated), remove Nones
                     prediction_value = utils.delete_none(
@@ -1738,7 +1738,7 @@ async def postprocess_data(
                 if isinstance(prediction_value, Block):
                     prediction_value = prediction_value.constructor_args.copy()
                     prediction_value["__type__"] = "update"
-                if utils.is_update(prediction_value):
+                if utils.is_prop_update(prediction_value):
                     kwargs = state[block._id].constructor_args.copy()
                     kwargs.update(prediction_value)
                     kwargs.pop("value", None)

gradio/components/dataset.py

@@ -2,6 +2,7 @@
 
 from __future__ import annotations
 
+import warnings
 from typing import Any, Literal
 
 from gradio_client.documentation import document
@@ -17,7 +18,8 @@
 @document()
 class Dataset(Component):
     """
-    Creates a gallery or table to display data samples. This component is designed for internal use to display examples.
+    Creates a gallery or table to display data samples. This component is primarily designed for internal use to display examples.
+    However, it can also be used directly to display a dataset and let users select examples.
     """
 
     EVENTS = [Events.click, Events.select]
@@ -26,7 +28,7 @@ def __init__(
         self,
         *,
         label: str | None = None,
-        components: list[Component] | list[str],
+        components: list[Component] | list[str] | None = None,
         component_props: list[dict[str, Any]] | None = None,
         samples: list[list[Any]] | None = None,
         headers: list[str] | None = None,
@@ -70,7 +72,7 @@ def __init__(
         self.container = container
         self.scale = scale
         self.min_width = min_width
-        self._components = [get_component_instance(c) for c in components]
+        self._components = [get_component_instance(c) for c in components or []]
         if component_props is None:
             self.component_props = [
                 component.recover_kwargs(
@@ -131,29 +133,39 @@ def get_config(self):
 
         return config
 
-    def preprocess(self, payload: int) -> int | list | None:
+    def preprocess(self, payload: int | None) -> int | list | None:
         """
         Parameters:
             payload: the index of the selected example in the dataset
         Returns:
             Passes the selected sample either as a `list` of data corresponding to each input component (if `type` is "value") or as an `int` index (if `type` is "index")
         """
+        if payload is None:
+            return None
         if self.type == "index":
             return payload
         elif self.type == "values":
             return self.samples[payload]
 
-    def postprocess(self, samples: list[list]) -> dict:
+    def postprocess(self, sample: int | list | None) -> int | None:
         """
         Parameters:
-            samples: Expects a `list[list]` corresponding to the dataset data, can be used to update the dataset.
+            sample: Expects an `int` index or `list` of sample data. Returns the index of the sample in the dataset or `None` if the sample is not found.
         Returns:
-            Returns the updated dataset data as a `dict` with the key "samples".
+            Returns the index of the sample in the dataset.
         """
-        return {
-            "samples": samples,
-            "__type__": "update",
-        }
+        if sample is None or isinstance(sample, int):
+            return sample
+        if isinstance(sample, list):
+            try:
+                index = self.samples.index(sample)
+            except ValueError:
+                index = None
+                warnings.warn(
+                    "The `Dataset` component does not support updating the dataset data by providing "
+                    "a set of list values. Instead, you should return a new Dataset(samples=...) object."
+                )
+            return index
 
     def example_payload(self) -> Any:
         return 0

gradio/flagging.py

@@ -164,7 +164,7 @@ def flag(
             ) / client_utils.strip_invalid_filename_characters(
                 getattr(component, "label", None) or f"component {idx}"
             )
-            if utils.is_update(sample):
+            if utils.is_prop_update(sample):
                 csv_data.append(str(sample))
             else:
                 data = (

gradio/helpers.py

@@ -544,7 +544,7 @@ def load_from_cache(self, example_id: int) -> list[Any]:
                     component, components.File
                 ):
                     value_to_use = value_as_dict
-                if not utils.is_update(value_as_dict):
+                if not utils.is_prop_update(value_as_dict):
                     raise TypeError("value wasn't an update")  # caught below
                 output.append(value_as_dict)
             except (ValueError, TypeError, SyntaxError):

gradio/utils.py

@@ -737,7 +737,7 @@ def validate_url(possible_url: str) -> bool:
         return False
 
 
-def is_update(val):
+def is_prop_update(val):
     return isinstance(val, dict) and "update" in val.get("__type__", "")
 
 

js/_website/src/lib/templates/gradio/03_components/dataset.svx

@@ -86,6 +86,40 @@ def predict(···) -> list[list]
 <DemosSection demos={obj.demos} />
 {/if}
 
+### Examples
+
+**Updating a Dataset**
+
+In this example, we display a text dataset using `gr.Dataset` and then update it when the user clicks a button:
+
+```py
+import gradio as gr
+
+philosophy_quotes = [
+    ["I think therefore I am."],
+    ["The unexamined life is not worth living."]
+]
+
+startup_quotes = [
+    ["Ideas are easy. Implementation is hard"],
+    ["Make mistakes faster."]
+]
+
+def show_startup_quotes():
+    return gr.Dataset(samples=startup_quotes)
+
+with gr.Blocks() as demo:
+    textbox = gr.Textbox()
+    dataset = gr.Dataset(components=[textbox], samples=philosophy_quotes)
+    button = gr.Button()
+
+    button.click(show_startup_quotes, None, dataset)
+
+demo.launch()
+```
+
+
+
 {#if obj.fns && obj.fns.length > 0}
 <!--- Event Listeners -->
 ### Event Listeners 
@@ -97,3 +131,4 @@ def predict(···) -> list[list]
 ### Guides
 <GuidesSection guides={obj.guides}/>
 {/if}
+

js/dataset/Index.svelte

@@ -12,7 +12,7 @@
 	>;
 	export let label = "Examples";
 	export let headers: string[];
-	export let samples: any[][];
+	export let samples: any[][] | null = null;
 	export let elem_id = "";
 	export let elem_classes: string[] = [];
 	export let visible = true;
@@ -34,7 +34,7 @@
 		: `${root}/file=`;
 	let page = 0;
 	$: gallery = components.length < 2;
-	let paginate = samples.length > samples_per_page;
+	let paginate = samples ? samples.length > samples_per_page : false;
 
 	let selected_samples: any[][];
 	let page_count: number;
@@ -51,6 +51,7 @@
 	}
 
 	$: {
+		samples = samples || [];
 		paginate = samples.length > samples_per_page;
 		if (paginate) {
 			visible_pages = [];

test/test_blocks.py

@@ -732,6 +732,33 @@ def infer(a, b):
         ):
             await demo.postprocess_data(demo.fns[0], predictions=(1, 2), state=None)
 
+    @pytest.mark.asyncio
+    async def test_dataset_is_updated(self):
+        def update(value):
+            return value, gr.Dataset(samples=[["New A"], ["New B"]])
+
+        with gr.Blocks() as demo:
+            with gr.Row():
+                textbox = gr.Textbox()
+                dataset = gr.Dataset(
+                    components=["text"], samples=[["Original"]], label="Saved Prompts"
+                )
+                dataset.click(update, inputs=[dataset], outputs=[textbox, dataset])
+        app, _, _ = demo.launch(prevent_thread_lock=True)
+
+        client = TestClient(app)
+
+        session_1 = client.post(
+            "/api/predict/",
+            json={"data": [0], "session_hash": "1", "fn_index": 0},
+        )
+        assert "Original" in session_1.json()["data"][0]
+        session_2 = client.post(
+            "/api/predict/",
+            json={"data": [0], "session_hash": "1", "fn_index": 0},
+        )
+        assert "New" in session_2.json()["data"][0]
+
 
 class TestStateHolder:
     @pytest.mark.asyncio