Merge pull request #236 from Renumics/layout_templates

Layout templates

renumics/spotlight/app.py

@@ -50,7 +50,7 @@
 from renumics.spotlight.backend.middlewares.timing import add_timing_middleware
 from renumics.spotlight.app_config import AppConfig
 from renumics.spotlight.data_source import DataSource, create_datasource
-from renumics.spotlight.layout.default import DEFAULT_LAYOUT
+from renumics.spotlight import layouts
 
 from renumics.spotlight.data_store import DataStore
 
@@ -86,7 +86,7 @@ class SpotlightApp(FastAPI):
 
     task_manager: TaskManager
     websocket_manager: Optional[WebsocketManager]
-    _layout: Optional[Layout]
+    _layout: Layout
     config: Config
     username: str
     filebrowsing_allowed: bool
@@ -106,7 +106,7 @@ def __init__(self) -> None:
         self.task_manager = TaskManager()
         self.websocket_manager = None
         self.config = Config()
-        self._layout = None
+        self._layout = layouts.default()
         self.project_root = Path.cwd()
         self.vite_url = None
         self.username = ""
@@ -368,11 +368,11 @@ def layout(self) -> Layout:
         """
         Frontend layout
         """
-        return self._layout or DEFAULT_LAYOUT
+        return self._layout
 
     @layout.setter
     def layout(self, layout: Optional[Layout]) -> None:
-        self._layout = layout
+        self._layout = layout or layouts.default()
         self._broadcast(ResetLayoutMessage())
 
     async def get_current_layout_dict(self, user_id: str) -> Optional[Dict]:

renumics/spotlight/layouts/__init__.py

@@ -0,0 +1,5 @@
+from .default import default
+from .model_debug import debug_classification
+from .model_compare import compare_classification
+
+__all__ = ["default", "debug_classification", "compare_classification"]

renumics/spotlight/layouts/default.py

@@ -0,0 +1,26 @@
+from renumics.spotlight.layout import (
+    histogram,
+    inspector,
+    layout,
+    scatterplot,
+    similaritymap,
+    split,
+    tab,
+    table,
+)
+from renumics.spotlight.layout.nodes import Layout
+
+
+def default() -> Layout:
+    """
+    Default layout for spotlight.
+    """
+
+    return layout(
+        split(
+            tab(table(), weight=60),
+            tab(similaritymap(), scatterplot(), histogram(), weight=40),
+            weight=60,
+        ),
+        tab(inspector(), weight=40),
+    )

renumics/spotlight/layouts/model_compare.py

@@ -0,0 +1,163 @@
+from typing import Optional, Union, Dict, Any
+from renumics.spotlight import dtypes
+from renumics.spotlight import layout
+from renumics.spotlight.layout import (
+    Layout,
+    Tab,
+    Split,
+    lenses,
+    table,
+    similaritymap,
+    inspector,
+    split,
+    tab,
+    metric,
+    issues,
+    confusion_matrix,
+)
+
+
+def compare_classification(
+    label: str = "label",
+    model1_prediction: str = "m1_prediction",
+    model1_embedding: str = "",
+    model1_correct: str = "",
+    model2_prediction: str = "m2_prediction",
+    model2_embedding: str = "",
+    model2_correct: str = "",
+    inspect: Optional[Dict[str, Any]] = None,
+) -> Layout:
+    """This function generates a Spotlight layout for comparing two different machine learning classification models.
+
+    Args:
+        label: Name of the column that contains the label.
+        model1_prediction: Name of the column that contains the prediction for model 1.
+        model1_embedding: Name of the column that contains thee embedding for model 1.
+        model1_correct: Name of the column that contains a flag if the data sample is predicted correctly by model 1.
+        model2_prediction: Name of the column that contains the prediction for model 2.
+        model2_embedding: Name of the column that contains thee embedding for model 2.
+        model2_correct: Name of the column that contains a flag if the data sample is predicted correctly by model 2.
+        inspect: Name and type of the columns that are displayed in the inspector, e.g. {'audio': spotlight.dtypes.audio_dtype}.
+
+    Returns:
+        The configured layout for `spotlight.show`.
+    """
+
+    # first column: table + issues
+    metrics = split(
+        [
+            tab(
+                metric(
+                    name="Accuracy model 1",
+                    metric="accuracy",
+                    columns=[label, model1_prediction],
+                )
+            ),
+            tab(
+                metric(
+                    name="Accuracy model 2",
+                    metric="accuracy",
+                    columns=[label, model2_prediction],
+                )
+            ),
+        ],
+        orientation="vertical",
+        weight=15,
+    )
+    column1 = split(
+        [metrics, tab(table(), weight=65)], weight=80, orientation="horizontal"
+    )
+    column1 = split(
+        [column1, tab(issues(), weight=40)], weight=80, orientation="horizontal"
+    )
+
+    column2_list = []
+    column2_list.append(
+        tab(
+            confusion_matrix(
+                name="Model 1 confusion matrix",
+                x_column=label,
+                y_column=model1_prediction,
+            ),
+            confusion_matrix(
+                name="Model 2 confusion matrix",
+                x_column=label,
+                y_column=model2_prediction,
+            ),
+            weight=40,
+        )
+    )
+
+    # third column: similarity maps
+    if model1_correct != "":
+        if model2_correct != "":
+            row2 = tab(
+                confusion_matrix(
+                    name="Model1 vs. Model2 - binned scatterplot",
+                    x_column=model1_correct,
+                    y_column=model2_correct,
+                ),
+                weight=40,
+            )
+            column2_list.append(row2)
+
+    if model1_embedding != "":
+        if model2_embedding != "":
+            row3 = tab(
+                similaritymap(
+                    name="Model 1 embedding",
+                    columns=[model1_embedding],
+                    color_by_column=label,
+                ),
+                similaritymap(
+                    name="Model 2 embedding",
+                    columns=[model2_embedding],
+                    color_by_column=label,
+                ),
+                weight=40,
+            )
+            column2_list.append(row3)
+
+    column2: Union[Tab, Split]
+
+    if len(column2_list) == 1:
+        column2 = column2_list[0]
+    elif len(column2_list) == 2:
+        column2 = split(column2_list, orientation="horizontal")
+    else:
+        column2 = split(
+            [column2_list[0], column2_list[1]], weight=80, orientation="horizontal"
+        )
+        column2 = split([column2, column2_list[2]], orientation="horizontal")
+
+    # fourth column: inspector
+    inspector_fields = []
+    if inspect:
+        for item, dtype_like in inspect.items():
+            dtype = dtypes.create_dtype(dtype_like)
+            if dtypes.is_audio_dtype(dtype):
+                inspector_fields.append(lenses.audio(item))
+            elif dtypes.is_image_dtype(dtype):
+                inspector_fields.append(lenses.image(item))
+            else:
+                print(f"Type {dtype} not supported by this layout.")
+
+        inspector_fields.append(lenses.scalar(label))
+        inspector_fields.append(lenses.scalar(model1_prediction))
+        inspector_fields.append(lenses.scalar(model2_prediction))
+
+        inspector_view = inspector("Inspector", lenses=inspector_fields, num_columns=4)
+
+    else:
+        inspector_view = inspector("Inspector", num_columns=4)
+
+    # build everything together
+    column2.weight = 40
+    half1 = split([column1, column2], weight=80, orientation="vertical")
+    half2 = tab(inspector_view, weight=40)
+
+    nodes = [half1, half2]
+
+    the_layout = layout.layout(nodes)
+
+    return the_layout

renumics/spotlight/layouts/model_debug.py

@@ -0,0 +1,127 @@
+from typing import Optional, Union, Dict, List, Any
+from renumics.spotlight import layout
+from renumics.spotlight.layout import (
+    Layout,
+    Tab,
+    Split,
+    lenses,
+    table,
+    similaritymap,
+    inspector,
+    split,
+    tab,
+    metric,
+    issues,
+    confusion_matrix,
+    histogram,
+)
+from renumics.spotlight.dtypes import create_dtype, is_audio_dtype, is_image_dtype
+
+
+def debug_classification(
+    label: str = "label",
+    prediction: str = "prediction",
+    embedding: str = "",
+    inspect: Optional[Dict[str, Any]] = None,
+    features: Optional[List[str]] = None,
+) -> Layout:
+    """This function generates a Spotlight layout for debugging a machine learning classification model.
+
+    Args:
+        label: Name of the column that contains the label.
+        prediction: Name of the column that contains the prediction.
+        embedding: Name of the column that contains the embedding.
+        inspect: Name and type of the columns that are displayed in the inspector, e.g. {'audio': spotlight.dtypes.audio_dtype}.
+        features: Names of the columns that contain useful metadata and features.
+
+    Returns:
+        The configured layout for `spotlight.show`.
+    """
+
+    # first column: table + issues
+    metrics = tab(
+        metric(name="Accuracy", metric="accuracy", columns=[label, prediction]),
+        weight=15,
+    )
+    column1 = split(
+        [metrics, tab(table(), weight=65)], weight=80, orientation="horizontal"
+    )
+    column1 = split(
+        [column1, tab(issues(), weight=40)], weight=80, orientation="horizontal"
+    )
+
+    column2_list = []
+    column2_list.append(
+        tab(
+            confusion_matrix(
+                name="Confusion matrix", x_column=label, y_column=prediction
+            ),
+            weight=40,
+        )
+    )
+
+    # second column: confusion matric, feature histograms (optional), embedding (optional)
+    if features is not None:
+        histogram_list = []
+        for idx, feature in enumerate(features):
+            if idx > 2:
+                break
+            h = histogram(
+                name="Histogram {}".format(feature),
+                column=feature,
+                stack_by_column=label,
+            )
+            histogram_list.append(h)
+
+        row2 = tab(*histogram_list, weight=40)
+        column2_list.append(row2)
+
+    if embedding != "":
+        row3 = tab(
+            similaritymap(name="Embedding", columns=[embedding], color_by_column=label),
+            weight=40,
+        )
+        column2_list.append(row3)
+
+    column2: Union[Tab, Split]
+
+    if len(column2_list) == 1:
+        column2 = column2_list[0]
+    elif len(column2_list) == 2:
+        column2 = split(column2_list, orientation="horizontal")
+    else:
+        column2 = split(
+            [column2_list[0], column2_list[1]], weight=80, orientation="horizontal"
+        )
+        column2 = split([column2, column2_list[2]], orientation="horizontal")
+
+    # fourth column: inspector
+    inspector_fields = []
+    if inspect:
+        for item, dtype_like in inspect.items():
+            dtype = create_dtype(dtype_like)
+            if is_audio_dtype(dtype):
+                inspector_fields.append(lenses.audio(item))
+            elif is_image_dtype(dtype):
+                inspector_fields.append(lenses.image(item))
+            else:
+                print("Type {} not supported by this layout.".format(dtype))
+
+        inspector_fields.append(lenses.scalar(label))
+        inspector_fields.append(lenses.scalar(prediction))
+
+        inspector_view = inspector("Inspector", lenses=inspector_fields, num_columns=4)
+
+    else:
+        inspector_view = inspector("Inspector", num_columns=4)
+
+    # build everything together
+    column2.weight = 40
+    half1 = split([column1, column2], weight=80, orientation="vertical")
+    half2 = tab(inspector_view, weight=40)
+
+    nodes = [half1, half2]
+
+    the_layout = layout.layout(nodes)
+
+    return the_layout