Updated eval script for monkey comparison with c model

Author: pdeubel

evaluation/plot_monkey_tester_c_model.ipynb

@@ -8,13 +8,18 @@
    "outputs": [],
    "source": [
     "import numpy as np\n",
-    "from matplotlib import pyplot as plt"
+    "from matplotlib import pyplot as plt\n",
+    "import os\n",
+    "\n",
+    "os.chdir(\"../\")"
    ]
   },
   {
    "cell_type": "markdown",
    "id": "a373b877-d487-4ccf-afd8-cefa968d8a0f",
-   "metadata": {},
+   "metadata": {
+    "tags": []
+   },
    "source": [
     "# Plot of three monkey testers compared with the Approach\n",
     "\n",
@@ -33,12 +38,12 @@
    "metadata": {},
    "outputs": [],
    "source": [
-    "random_monkey_dir = \"\"\n",
-    "random_widget_monkey_dir = \"\"\n",
-    "random_widget_monkey_zero_random_click_prob_dir = \"\"\n",
+    "random_monkey_dir = \"datasets/gui_env/random-clicks/2022-04-20_14-02-28\"\n",
+    "random_widget_monkey_dir = \"datasets/gui_env/random-widgets/2022-04-20_14-03-07\"\n",
+    "random_widget_monkey_zero_random_click_prob_dir = \"datasets/gui_env/random-widgets/2022-04-20_14-04-12\"\n",
     "\n",
-    "first_controller_model_file = \"\"\n",
-    "second_controller_model_file = \"\""
+    "first_controller_model_file = \"controller_v_103_eval.npz\"\n",
+    "second_controller_model_file = \"controller_v_76_eval.npz\""
    ]
   },
   {
@@ -52,7 +57,7 @@
     "random_widget_monkey_rewards = []\n",
     "random_widget_monkey_zero_random_click_prob_rewards = []\n",
     "\n",
-    "for monkey_dir in monkey_tester_dirs:\n",
+    "for i in range(3):\n",
     "    \n",
     "    if i == 0:\n",
     "        monkey_dir = random_monkey_dir\n",
@@ -72,13 +77,13 @@
     "        sequence_data = np.load(os.path.join(monkey_dir, seq_dir, \"data.npz\"))\n",
     "        reward_list.append(sequence_data[\"rewards\"])\n",
     "        \n",
-    "random_monkey_mean = np.mean(random_monkey_rewards, axis=0)\n",
+    "random_monkey_mean = np.mean(random_monkey_rewards, axis=0).cumsum()\n",
     "random_monkey_std = np.std(random_monkey_rewards, axis=0)\n",
     "\n",
-    "random_widget_monkey_mean = np.mean(random_widget_monkey_rewards, axis=0)\n",
+    "random_widget_monkey_mean = np.mean(random_widget_monkey_rewards, axis=0).cumsum()\n",
     "random_widget_monkey_std = np.std(random_widget_monkey_rewards, axis=0)\n",
     "\n",
-    "random_widget_monkey_zero_random_click_prob_mean = np.mean(random_widget_monkey_zero_random_click_prob_rewards, axis=0)\n",
+    "random_widget_monkey_zero_random_click_prob_mean = np.mean(random_widget_monkey_zero_random_click_prob_rewards, axis=0).cumsum()\n",
     "random_widget_monkey_zero_random_click_prob_std = np.std(random_widget_monkey_zero_random_click_prob_rewards, axis=0)"
    ]
   },
@@ -93,13 +98,13 @@
     "second_controller_rewards = []\n",
     "\n",
     "\n",
-    "first_controller_rewards.append(np.load(first_controller_model_file)[\"all_rewards\"])\n",
-    "second_controller_rewards.append(np.load(second_controller_model_file)[\"all_rewards\"])\n",
+    "first_controller_rewards = np.load(first_controller_model_file)[\"all_rewards\"]\n",
+    "second_controller_rewards = np.load(second_controller_model_file)[\"all_rewards\"]\n",
     "\n",
     "first_controller_mean = np.mean(first_controller_rewards, axis=0)\n",
     "second_controller_mean = np.mean(second_controller_rewards, axis=0)\n",
     "\n",
-    "controller_mean = np.mean([first_controller_mean, second_controller_mean], axis=0)\n",
+    "controller_mean = np.mean([first_controller_mean, second_controller_mean], axis=0).cumsum()\n",
     "controller_std = np.std([first_controller_mean, second_controller_mean], axis=0)"
    ]
   },
@@ -112,36 +117,34 @@
    "source": [
     "data_length_x = random_monkey_mean.shape\n",
     "\n",
-    "assert data_length_x = random_monkey_std.shape\n",
-    "assert data_length_x = random_widget_monkey_mean.shape\n",
-    "assert data_length_x = random_widget_monkey_std.shape\n",
-    "assert data_length_x = random_widget_monkey_zero_random_click_prob_mean.shape\n",
-    "assert data_length_x = random_widget_monkey_zero_random_click_prob_std.shape\n",
-    "\n",
-    "assert data_length_x = controller_mean.shape\n",
-    "assert data_length_x = random_monkey_std.shape\n",
+    "assert data_length_x == random_monkey_std.shape\n",
+    "assert data_length_x == random_widget_monkey_mean.shape\n",
+    "assert data_length_x == random_widget_monkey_std.shape\n",
+    "assert data_length_x == random_widget_monkey_zero_random_click_prob_mean.shape\n",
+    "assert data_length_x == random_widget_monkey_zero_random_click_prob_std.shape\n",
     "\n",
-    "x_values = np.arange(data_length_x[0])\n",
+    "assert data_length_x == controller_mean.shape\n",
+    "assert data_length_x == random_monkey_std.shape\n",
     "\n",
     "fig, ax = plt.subplots()\n",
+    "x_data = np.arange(data_length_x[0])\n",
     "\n",
-    "ax.plot(x_values, random_monkey_mean, label=\"Random Monkey Tester\")\n",
-    "ax.fill_between(x_values, random_monkey_mean - random_monkey_std, random_monkey_mean + random_monkey_std, alpha=0.25)\n",
+    "ax.plot(x_data, random_monkey_mean, label=\"Random Monkey Tester\")\n",
+    "ax.fill_between(x_data, random_monkey_mean - random_monkey_std, random_monkey_mean + random_monkey_std, alpha=0.25)\n",
     "\n",
-    "ax.plot(x_values, random_widget_monkey_mean, label=\"Random Widget Monkey Tester\")\n",
-    "ax.fill_between(x_values, random_widget_monkey_mean - random_widget_monkey_std, random_widget_monkey_mean + random_widget_monkey_std, alpha=0.25)\n",
+    "ax.plot(x_data, random_widget_monkey_mean, label=\"Random Widget Monkey Tester (0.125)\")\n",
+    "ax.fill_between(x_data, random_widget_monkey_mean - random_widget_monkey_std, random_widget_monkey_mean + random_widget_monkey_std, alpha=0.25)\n",
     "\n",
-    "ax.plot(x_values, random_widget_monkey_zero_random_click_prob_mean, label=\"Random Widget Monkey Tester (0.0)\")\n",
-    "ax.fill_between(x_values,\n",
-    "                 random_widget_monkey_zero_random_click_prob_mean - random_widget_monkey_zero_random_click_prob_std,\n",
+    "ax.plot(x_data, random_widget_monkey_zero_random_click_prob_mean, label=\"Random Widget Monkey Tester (0.0)\")\n",
+    "ax.fill_between(x_data, random_widget_monkey_zero_random_click_prob_mean - random_widget_monkey_zero_random_click_prob_std,\n",
     "                 random_widget_monkey_zero_random_click_prob_mean + random_widget_monkey_zero_random_click_prob_std,\n",
     "                 alpha=0.25)\n",
     "\n",
-    "ax.plot(x_values, controller_mean, label=\"Best Approach\")\n",
-    "ax.fill_between(x_values, controller_mean - controller_std, controller_mean + controller_std, alpha=0.25)\n",
+    "ax.plot(x_data, controller_mean, label=\"Investigated Approach (v_275, m_187)\")\n",
+    "ax.fill_between(x_data, controller_mean - controller_std, controller_mean + controller_std, alpha=0.25)\n",
     "\n",
-    "ax.set(ylabel=\"Code Coverage (\\%)\", xlabel='Time Steps')\n",
-    "ax.legend(loc='upper left')\n",
+    "ax.set(ylabel=\"Code Coverage (%)\", xlabel='Time Steps')\n",
+    "ax.legend(loc='lower right')\n",
     "\n",
     "fig.savefig(\"c_model_comparison_monkey_tester.pdf\", bbox_inches='tight')\n",
     "plt.show()"
@@ -150,7 +153,7 @@
  ],
  "metadata": {
   "kernelspec": {
-   "display_name": "Python 3",
+   "display_name": "Python 3 (ipykernel)",
    "language": "python",
    "name": "python3"
   },
@@ -164,7 +167,7 @@
    "name": "python",
    "nbconvert_exporter": "python",
    "pygments_lexer": "ipython3",
-   "version": "3.8.8"
+   "version": "3.9.6"
   }
  },
  "nbformat": 4,