feat(eda): convert all plot functions to new UI

dataprep/eda/container.py

@@ -19,6 +19,25 @@
     loader=PackageLoader("dataprep", "eda/templates"),
 )
 
+TAB_VISUAL_TYPES = {
+    "missing_impact_1v1",
+    "missing_impact",
+    "categorical_column",
+    "numerical_column",
+    "datetime_column",
+    "cat_and_num_cols",
+    "two_num_cols",
+    "two_cat_cols",
+    "dt_and_num_cols",
+    "dt_and_cat_cols",
+    "dt_cat_num_cols",
+    "correlation_impact",
+    "correlation_single_heatmaps",
+    "correlation_scatter",
+}
+
+GRID_VISUAL_TYPES = {"distribution_grid", "missing_impact_1vn"}
+
 
 class Container:
     """
@@ -30,38 +49,13 @@ def __init__(
         to_render: Dict[str, Any],
         visual_type: str,
     ) -> None:
-        if visual_type in ("distribution_grid", "missing_impact_1vn"):
-            self.context = {
-                "resources": INLINE.render(),
-                "components": components(to_render.get("layout")),
-                "tabledata": to_render.get("tabledata"),
-                "overview_insights": to_render.get("overview_insights"),
-                "column_insights": to_render.get("column_insights"),
-                "meta": to_render.get("meta"),
-                "title": "DataPrep.EDA Report",
-                "rnd": random.randint(0, 99),  # for multiple cells running in the same notebook
-                "container_width": f"{to_render['fig_width']*3}px",
-                "legend_labels": to_render.get("legend_labels"),
-            }
+        self.context = Context(**to_render)
+        setattr(self.context, "rnd", random.randint(0, 9999))
+        if visual_type in GRID_VISUAL_TYPES:
             self.template_base = ENV_LOADER.get_template("grid_base.html")
-
-        elif "_column" in visual_type or visual_type in (
-            "missing_impact",
-            "missing_impact_1v1",
-            "correlation_impact",
-        ):
-            # todo: param management
+        elif visual_type in TAB_VISUAL_TYPES:
             if to_render.get("tabledata"):
-                to_render["meta"].insert(0, "Stats")
-            self.context = {
-                "resources": INLINE.render(),
-                "tabledata": to_render.get("tabledata"),
-                "insights": to_render.get("insights"),
-                "components": components(to_render.get("layout")),
-                "meta": to_render.get("meta"),
-                "title": "DataPrep.EDA Report",
-                "rnd": random.randint(100, 999),  # for multiple cells running in the same notebook
-            }
+                self.context.meta.insert(0, "Stats")  # type: ignore
             if visual_type == "correlation_impact":
                 self.template_base = ENV_LOADER.get_template("tab_base_corr.html")
             else:
@@ -116,3 +110,34 @@ def show_browser(self) -> None:
         with open(tmpf.name, "w") as file:
             file.write(self.template_base.render(context=self.context))
         webbrowser.open_new_tab(f"file://{tmpf.name}")
+
+
+class Context:
+    """
+    Define the context class that stores all the parameters needed by template engine.
+    The instance is read-only.
+    """
+
+    _title = "DataPrep.EDA Report"
+    _resources = INLINE.render()
+    _container_width = 650  # default width just in case nothing got passed in
+
+    def __init__(self, **param: Any) -> None:
+        self.title = self._title
+        self.resources = self._resources
+        self.container_width = self._container_width
+
+        for attr, value in param.items():
+            if attr == "layout":
+                setattr(self, "components", components(value))
+            else:
+                setattr(self, attr, value)
+
+    def __getitem__(self, key: str) -> Any:
+        try:
+            return getattr(self, key)
+        except KeyError:
+            return None
+
+    def __getattr__(self, attr: str) -> None:
+        return None

dataprep/eda/correlation/__init__.py

@@ -8,7 +8,6 @@
 import pandas as pd
 
 from ..progress_bar import ProgressBar
-from ..report import Report
 from .compute import compute_correlation
 from .render import render_correlation
 from ..container import Container
@@ -24,7 +23,7 @@ def plot_correlation(
     value_range: Optional[Tuple[float, float]] = None,
     k: Optional[int] = None,
     progress: bool = True,
-) -> Union[Report, Container]:
+) -> Container:
     """
     This function is designed to calculate the correlation between columns
     There are three functions: plot_correlation(df), plot_correlation(df, x)
@@ -67,9 +66,6 @@ def plot_correlation(
     """
     with ProgressBar(minimum=1, disable=not progress):
         itmdt = compute_correlation(df, x=x, y=y, value_range=value_range, k=k)
-    fig = render_correlation(itmdt)
+    to_render = render_correlation(itmdt)
 
-    if itmdt.visual_type == "correlation_impact" or "_column" in itmdt.visual_type:
-        return Container(fig, itmdt.visual_type)
-    else:
-        return Report(fig)
+    return Container(to_render, itmdt.visual_type)

dataprep/eda/correlation/render.py

@@ -33,8 +33,8 @@
 
 def render_correlation(
     itmdt: Intermediate,
-    plot_width: int = 500,
-    plot_height: int = 500,
+    plot_width: int = 400,
+    plot_height: int = 400,
     palette: Optional[Sequence[str]] = None,
 ) -> Figure:
     """
@@ -154,6 +154,7 @@ def render_correlation_impact(
             toolbar_location=None,
             tooltips=tooltips,
             background_fill_color="#fafafa",
+            title=" ",
         )
 
         tweak_figure(fig)
@@ -167,8 +168,8 @@ def render_correlation_impact(
             fill_color={"field": "correlation", "transform": mapper},
             line_color=None,
         )
-
-        fig.add_layout(color_bar, "right")
+        fig.frame_width = plot_width
+        fig.add_layout(color_bar, "left")
         tab = Panel(child=fig, title=method)
         tabs.append(tab)
 
@@ -177,6 +178,7 @@ def render_correlation_impact(
         "tabledata": itmdt["tabledata"],
         "layout": [panel.child for panel in tabs],
         "meta": [panel.title for panel in tabs],
+        "container_width": plot_width + 150,
     }
 
 
@@ -233,7 +235,7 @@ def render_correlation_heatmaps(
 
 def render_correlation_single_heatmaps(
     itmdt: Intermediate, plot_width: int, plot_height: int, palette: Sequence[str]
-) -> Tabs:
+) -> Dict[str, Any]:
     """
     Render correlation heatmaps, but with single column
     """
@@ -272,8 +274,11 @@ def render_correlation_single_heatmaps(
         tab = Panel(child=fig, title=method)
         tabs.append(tab)
 
-    tabs = Tabs(tabs=tabs)
-    return tabs
+    return {
+        "layout": [panel.child for panel in tabs],
+        "meta": [panel.title for panel in tabs],
+        "container_width": plot_width,
+    }
 
 
 def create_color_mapper(palette: Sequence[str]) -> Tuple[LinearColorMapper, ColorBar]:
@@ -296,7 +301,7 @@ def create_color_mapper(palette: Sequence[str]) -> Tuple[LinearColorMapper, Colo
 ######### Scatter #########
 def render_scatter(
     itmdt: Intermediate, plot_width: int, plot_height: int, palette: Sequence[str]
-) -> Figure:
+) -> Dict[str, Any]:
     """
     Render scatter plot with a regression line and possible most influencial points
     """
@@ -312,7 +317,6 @@ def render_scatter(
         plot_width=plot_width,
         plot_height=plot_height,
         toolbar_location=None,
-        title=Title(text="Scatter Plot & Regression Line", align="center"),
         tools=[],
         x_axis_label=xcol,
         y_axis_label=ycol,
@@ -350,7 +354,11 @@ def render_scatter(
         )
 
         fig.add_layout(legend, place="right")
-    return fig
+    return {
+        "layout": [fig],
+        "meta": ["Scatter Plot & Regression Line"],
+        "container_width": plot_width,
+    }
 
 
 ######### Interactions for report #########

dataprep/eda/create_report/__init__.py

@@ -8,7 +8,7 @@
 from bokeh.resources import INLINE
 from jinja2 import Environment, PackageLoader
 from .formatter import format_report
-from .io import Report
+from .report import Report
 
 __all__ = ["create_report"]
 

dataprep/eda/create_report/formatter.py

@@ -173,11 +173,13 @@ def format_basic(df: dd.DataFrame) -> Dict[str, Any]:
     res["has_missing"] = True
     itmdt = completions["miss"](data["miss"])
 
-    rndrd = render_missing(itmdt)["layout"]
+    rndrd = render_missing(itmdt)
     figs.clear()
-    for fig in rndrd:
+    for fig in rndrd["layout"]:
         fig.sizing_mode = "stretch_width"
-        fig.title = Title(text=tab.title, align="center", text_font_size="20px")
+        fig.title = Title(
+            text=rndrd["meta"][rndrd["layout"].index(fig)], align="center", text_font_size="20px"
+        )
         figs.append(fig)
     res["missing"] = components(figs)
 

dataprep/eda/distribution/__init__.py

@@ -10,7 +10,6 @@
 from ..container import Container
 from ..dtypes import DTypeDef
 from ..progress_bar import ProgressBar
-from ..report import Report
 from .compute import compute
 from .render import render
 
@@ -39,7 +38,7 @@ def plot(
     tile_size: Optional[float] = None,
     dtype: Optional[DTypeDef] = None,
     progress: bool = True,
-) -> Union[Report, Container]:
+) -> Container:
     """Generates plots for exploratory data analysis.
 
     If no columns are specified, the distribution of
@@ -168,8 +167,5 @@ def plot(
             value_range=value_range,
             dtype=dtype,
         )
-    figure = render(intermediate, yscale=yscale, tile_size=tile_size)
-    if intermediate.visual_type == "distribution_grid" or "_column" in intermediate.visual_type:
-        return Container(figure, intermediate.visual_type)
-    else:
-        return Report(figure)
+    to_render = render(intermediate, yscale=yscale, tile_size=tile_size)
+    return Container(to_render, intermediate.visual_type)