Update examples

davidnabergoj · Aug 18, 2022 · ab0f364 · ab0f364
1 parent 652e12b
commit ab0f364
Show file tree

Hide file tree

Showing 2 changed files with 16 additions and 27 deletions.
diff --git a/docs/source/quickstart.rst b/docs/source/quickstart.rst
@@ -82,14 +82,14 @@ We simply move the plotting code into a function and pass this function to ``boo
         test_x = np.linspace(-2, 3).reshape(-1, 1)
         ax.scatter(data_full[:, 0], data_full[:, 1])
         ax.plot(test_x, lr.predict(test_x), c='r')
+        ax.set_xlim(-2, 3)
+        ax.set_ylim(-20, 40)
 
     bootplot(
         plot_regression,
         data=np.column_stack([x, y]),
         output_image_path='quickstart_regression.png',
-        output_animation_path='quickstart_regression.gif',
-        xlim=(-2, 3),
-        ylim=(-20, 40)
+        output_animation_path='quickstart_regression.gif'
     )
 
 The result is an image and an animation that both display regression line uncertainty:
@@ -103,5 +103,3 @@ The result is an image and an animation that both display regression line uncert
     It is often essential to manually specify axis limits.
     This is to ensure all bootstrapped plots cover the same area and hence use the same axis ticks.
     If axis limits are not provided, some image elements such as ticks may appear blurry depending on the plotting function.
-    Axis limits may be specified in the plotting function or passed to ``bootplot`` (bootplot).
-    The user will be warned and encouraged to provide limits if none are passed to ``bootplot``.
diff --git a/examples/notebooks/categorical_data.ipynb b/examples/notebooks/categorical_data.ipynb
@@ -149,7 +149,9 @@
    "outputs": [],
    "source": [
     "def target_distribution_bar_chart(data_subset, data_full, ax):\n",
-    "    ax.bar(*np.unique(data_subset, return_counts=True))"
+    "    ax.bar(*np.unique(data_subset, return_counts=True))\n",
+    "    ax.set_xlim(-1, n_classes)\n",
+    "    ax.set_ylim(0, 25)"
    ],
    "metadata": {
     "collapsed": false,
@@ -182,7 +184,10 @@
     "        bbox_to_anchor=(1.1, 1.1, 0, 0),\n",
     "        shadow=True,\n",
     "        title='Class'\n",
-    "    )"
+    "    )\n",
+    "\n",
+    "    ax.set_xlim(-3, 3)\n",
+    "    ax.set_ylim(-3, 3)"
    ],
    "metadata": {
     "collapsed": false,
@@ -196,7 +201,7 @@
    "source": [
     "## Running bootplot\n",
     "\n",
-    "Having defined the plotting functions, we simply pass them to `bootplot` along with the data and axis limits. In this notebook, we will only work with the returned static images, but we could also save these to disk by specifying `output_image_path` for images and `output_animation_path` for animations. We manually specify `m = 1000` bootstrap samples to make the final result a bit smoother."
+    "Having defined the plotting functions, we simply pass them to `bootplot` along with the data. In this notebook, we will only work with the returned static images, but we could also save these to disk by specifying `output_image_path` for images and `output_animation_path` for animations. We manually specify `m = 1000` bootstrap samples to make the final result a bit smoother."
    ],
    "metadata": {
     "collapsed": false,
@@ -210,12 +215,7 @@
    "execution_count": 6,
    "outputs": [],
    "source": [
-    "bar_chart_image = bootplot(\n",
-    "    target_distribution_bar_chart,\n",
-    "    data,\n",
-    "    xlim=(-1, n_classes),\n",
-    "    ylim=(0, 25)\n",
-    ")"
+    "bar_chart_image = bootplot(target_distribution_bar_chart, data)"
    ],
    "metadata": {
     "collapsed": false,
@@ -229,13 +229,7 @@
    "execution_count": 7,
    "outputs": [],
    "source": [
-    "pie_chart_image = bootplot(\n",
-    "    target_distribution_pie_chart,\n",
-    "    data,\n",
-    "    m=1000,\n",
-    "    xlim=(-3, 3),\n",
-    "    ylim=(-3, 3)\n",
-    ")"
+    "pie_chart_image = bootplot(target_distribution_pie_chart, data, m=1000)"
    ],
    "metadata": {
     "collapsed": false,
@@ -310,13 +304,10 @@
     "def target_distribution_bar_chart_with_ground_truth(data_subset, data_full, ax):\n",
     "    ax.bar(*np.unique(data_subset, return_counts=True))\n",
     "    ax.scatter(np.arange(n_classes), binomial_pmf(np.arange(n_classes), n_classes - 1, event_probability) * len(data), color='tab:orange', edgecolor='white', s=100)\n",
+    "    ax.set_xlim(-1, n_classes)\n",
+    "    ax.set_ylim(0, 25)\n",
     "\n",
-    "bar_chart_image_with_ground_truth = bootplot(\n",
-    "    target_distribution_bar_chart_with_ground_truth,\n",
-    "    data,\n",
-    "    xlim=(-1, n_classes),\n",
-    "    ylim=(0, 25)\n",
-    ")\n",
+    "bar_chart_image_with_ground_truth = bootplot(target_distribution_bar_chart_with_ground_truth, data)\n",
     "\n",
     "plt.figure(figsize=(9, 9))\n",
     "plt.imshow(bar_chart_image_with_ground_truth)\n",