feat: Add support for small (1 subfigure) graphs

Author: Petr Laštovička

benchmarks/NAC_benchmark.ipynb

@@ -8,10 +8,18 @@
     "\n",
     "In this notebook we provide utils to run benchmarks and experiment with our code.\n",
     "\n",
-    "In the first section we start with utility functions, in the second part we load/generate benchmark data. After we run individual benchmarks on selected graph classes with selected algorithms. The algorithms are described in that section.\n",
-    "\n",
-    "```bash\n",
-    "tensorboard --logdir benchmarks/logs/nac\n",
+    "In the first section we start with utility functions, in the second part we load/generate benchmark data. After we run individual benchmarks on selected graph classes with selected algorithms. The algorithms are described in that section."
+   ]
+  },
+  {
+   "cell_type": "markdown",
+   "metadata": {},
+   "source": [
+    "If you are using VScode, add this option to your `.vscode/settings.json` file.\n",
+    "```json\n",
+    "{\n",
+    "    \"jupyter.notebookFileRoot\": \"${workspaceFolder}\"\n",
+    "}\n",
     "```"
    ]
   },
@@ -1790,7 +1798,7 @@
    "name": "python",
    "nbconvert_exporter": "python",
    "pygments_lexer": "ipython3",
-   "version": "3.12.6"
+   "version": "3.12.8"
   }
  },
  "nbformat": 4,

benchmarks/NAC_latex_export.ipynb

@@ -19,9 +19,21 @@
     "The `nac` package itself is completely independent."
    ]
   },
+  {
+   "cell_type": "markdown",
+   "metadata": {},
+   "source": [
+    "If you are using VScode, add this option to your `.vscode/settings.json` file.\n",
+    "```json\n",
+    "{\n",
+    "    \"jupyter.notebookFileRoot\": \"${workspaceFolder}\"\n",
+    "}\n",
+    "```"
+   ]
+  },
   {
    "cell_type": "code",
-   "execution_count": null,
+   "execution_count": 119,
    "metadata": {},
    "outputs": [],
    "source": [
@@ -76,7 +88,7 @@
   },
   {
    "cell_type": "code",
-   "execution_count": null,
+   "execution_count": 120,
    "metadata": {},
    "outputs": [],
    "source": [
@@ -89,16 +101,17 @@
   },
   {
    "cell_type": "code",
-   "execution_count": null,
+   "execution_count": 121,
    "metadata": {},
    "outputs": [],
    "source": [
     "# https://jwalton.info/Embed-Publication-Matplotlib-Latex/\n",
     "import matplotlib.backends.backend_pgf\n",
     "\n",
+    "DEFAULT_FIG_WIDTH = 398.33858\n",
     "\n",
     "def fig_size(\n",
-    "        width: float = 398.33858,\n",
+    "        width: float = DEFAULT_FIG_WIDTH,\n",
     "        fraction: float = 2,\n",
     "        subplots: Tuple[int,int] = (1, 1),\n",
     "    ):\n",
@@ -117,6 +130,7 @@
     "    fig_dim: tuple\n",
     "            Dimensions of figure in inches\n",
     "    \"\"\"\n",
+    "\n",
     "    # Width of figure (in pts)\n",
     "    fig_width_pt = width * fraction\n",
     "    # Convert from pt to inches\n",
@@ -131,6 +145,10 @@
     "    # Figure height in inches\n",
     "    fig_height_in = fig_width_in * golden_ratio * (subplots[0] / subplots[1])\n",
     "\n",
+    "    if subplots == (1, 1):\n",
+    "        fig_width_in *= 3/4\n",
+    "        fig_height_in *= 2/5\n",
+    "\n",
     "    return (fig_width_in, fig_height_in)\n",
     "\n",
     "\n",
@@ -173,7 +191,7 @@
   },
   {
    "cell_type": "code",
-   "execution_count": null,
+   "execution_count": 122,
    "metadata": {},
    "outputs": [],
    "source": [
@@ -249,19 +267,20 @@
   },
   {
    "cell_type": "code",
-   "execution_count": null,
+   "execution_count": 124,
    "metadata": {},
    "outputs": [],
    "source": [
+    "\n",
     "def _group_and_plot(\n",
     "    df: pd.DataFrame,\n",
     "    with_log: bool,\n",
     "    axs: List[plt.Axes],\n",
+    "    aggregations: List[Literal[\"mean\", \"median\", \"3rd quartile\"]],\n",
     "    x_column: Literal[\"vertex_no\", \"monochromatic_classes_no\"],\n",
     "    based_on: Literal[\"relabel\", \"split\", \"merging\"],\n",
     "    value_columns: List[Literal[\"nac_first_mean_time\", \"nac_all_mean_time\"]],\n",
     "):\n",
-    "    aggregations = [\"mean\", \"median\", \"3rd quartile\"]\n",
     "    df = df.loc[:, [x_column, based_on, *value_columns]]\n",
     "    groupped = df.groupby([x_column, based_on])\n",
     "\n",
@@ -295,7 +314,18 @@
     "            ax.set_yscale(\"log\")\n",
     "        ax.xaxis.set_major_locator(MaxNLocator(integer=True))\n",
     "        ax.set_xlabel(rename_based_on[x_column])\n",
-    "        ax.legend(loc='upper left')\n",
+    "\n",
+    "\n",
+    "\n",
+    "        if len(aggregations) == 1:\n",
+    "            ax.legend(\n",
+    "                bbox_to_anchor=(1.0, 1.0),\n",
+    "                loc='upper left',\n",
+    "            )\n",
+    "        else:\n",
+    "            ax.legend(\n",
+    "                loc='upper left',\n",
+    "            )\n",
     "\n",
     "def plot_frame(\n",
     "    title: str,\n",
@@ -314,7 +344,8 @@
     "        # \"merging\",\n",
     "        \"split_merging\",\n",
     "    ],\n",
-    "    ops_aggregation = [\"mean\", \"median\",], #  \"3rd quartile\",\n",
+    "    # ops_aggregation = [\"mean\", \"median\",], #  \"3rd quartile\",\n",
+    "    ops_aggregation = [\"mean\"], #  \"3rd quartile\",\n",
     ") -> List[Figure]:\n",
     "    print(f\"Plotting {df.shape[0]} records...\")\n",
     "    figs = []\n",
@@ -337,7 +368,10 @@
     "            ncols = len(ops_aggregation)\n",
     "            fig = figure(\n",
     "                nrows * ncols,\n",
-    "                figsize=fig_size(subplots=(nrows, ncols)),\n",
+    "                figsize=fig_size(\n",
+    "                    # width=DEFAULT_FIG_WIDTH if len(ops_aggregation) > 1 else DEFAULT_FIG_WIDTH * 3/4,\n",
+    "                    subplots=(nrows, ncols),\n",
+    "                ),\n",
     "                layout='constrained',\n",
     "            )\n",
     "            title_detail = \" | \".join(title_rename[value_column] for value_column in value_columns)\n",
@@ -349,7 +383,7 @@
     "                axs = [\n",
     "                    fig.add_subplot(nrows, ncols, i+ncols*row+1)\n",
     "                    for i in range(len(ops_aggregation))]\n",
-    "                _group_and_plot(local_df, with_log, axs, x_column, based_on, value_columns)\n",
+    "                _group_and_plot(local_df, with_log, axs, ops_aggregation, x_column, based_on, value_columns)\n",
     "                row += 1\n",
     "    return figs\n",
     "\n",
@@ -372,7 +406,7 @@
   },
   {
    "cell_type": "code",
-   "execution_count": null,
+   "execution_count": 126,
    "metadata": {},
    "outputs": [],
    "source": [
@@ -386,7 +420,7 @@
   },
   {
    "cell_type": "code",
-   "execution_count": null,
+   "execution_count": 127,
    "metadata": {},
    "outputs": [],
    "source": [
@@ -459,7 +493,7 @@
   },
   {
    "cell_type": "code",
-   "execution_count": null,
+   "execution_count": 130,
    "metadata": {},
    "outputs": [],
    "source": [
@@ -471,6 +505,7 @@
     "    value_columns: List[Literal[\"nac_first_mean_time\", \"nac_all_mean_time\"]],\n",
     "    aggregation: Literal[\"mean\", \"median\", \"3rd quartile\"],\n",
     "    legend_rename_dict: Dict[str, str] = {},\n",
+    "    legend_outside: bool = False,\n",
     "):\n",
     "    df = df.loc[:, [x_column, *value_columns]]\n",
     "    groupped = df.groupby([x_column])\n",
@@ -496,11 +531,19 @@
     "    ax.xaxis.set_major_locator(MaxNLocator(integer=True))\n",
     "    ax.set_xlabel(rename_based_on[x_column])\n",
     "    handles, labels = ax.get_legend_handles_labels()\n",
-    "    ax.legend(\n",
+    "    if legend_outside:\n",
+    "        ax.legend(\n",
     "        handles,\n",
     "        [legend_rename_dict[l] for l in labels],\n",
-    "        # loc = 'upper left',\n",
-    "    )\n",
+    "            bbox_to_anchor=(1.0, 1.0),\n",
+    "            loc='upper left',\n",
+    "        )\n",
+    "    else:\n",
+    "        ax.legend(\n",
+    "            handles,\n",
+    "            [legend_rename_dict[l] for l in labels],\n",
+    "            # loc = 'upper left',\n",
+    "        )\n",
     "\n",
     "def plot_is_NAC_coloring_calls(\n",
     "    df: pd.DataFrame,\n",
@@ -567,7 +610,8 @@
     "        # [\"inv_edge_no\",    \"inv_triangle_component_no\",    \"inv_monochromatic_class_no\",    \"inv_nac_all_check_cycle_mask\",    \"inv_nac_all_check_is_NAC\", ],\n",
     "        # [\"new_edge_no\",    \"new_triangle_component_no\",    \"new_monochromatic_class_no\",    \"new_nac_all_check_cycle_mask\" ],\n",
     "    ]\n",
-    "    ops_aggregation = [\"mean\", \"median\", ] # \"3rd quartile\",\n",
+    "    # ops_aggregation = [\"mean\", \"median\", ] # \"3rd quartile\",\n",
+    "    ops_aggregation = [\"mean\" ] # \"3rd quartile\",\n",
     "\n",
     "    nrows = len(ops_value_groups)\n",
     "    ncols = len(ops_aggregation)\n",
@@ -595,7 +639,7 @@
     "                fig.add_subplot(nrows, ncols, i+ncols*row+1)\n",
     "                for i in range(len(ops_aggregation))]\n",
     "            for ax, aggregation in zip(axs,ops_aggregation):\n",
-    "                _plot_is_NAC_coloring_calls_groups(title, df, ax, x_column, value_columns, aggregation, legend_rename_dict=rename_dict)\n",
+    "                _plot_is_NAC_coloring_calls_groups(title, df, ax, x_column, value_columns, aggregation, legend_rename_dict=rename_dict, legend_outside=len(ops_aggregation) <= 1)\n",
     "            row += 1\n",
     "\n",
     "    return figs"
@@ -633,7 +677,7 @@
    "name": "python",
    "nbconvert_exporter": "python",
    "pygments_lexer": "ipython3",
-   "version": "3.12.6"
+   "version": "3.12.8"
   }
  },
  "nbformat": 4,