client: Add a client Notebook

Add a Notebook that demonstrates how to invoke the KServe Inference
Service.

NOTE: Do not forget to create an Authorization Policy for the Inference
Service. See here: kubeflow/manifests#2811

Signed-off-by: Dimitris Poulopoulos <dimitris.a.poulopoulos@gmail.com>

03.client.ipynb

@@ -0,0 +1,132 @@
+{
+ "cells": [
+  {
+   "cell_type": "code",
+   "execution_count": null,
+   "id": "67f190cf-9caf-4da7-adc7-048065698f6d",
+   "metadata": {},
+   "outputs": [],
+   "source": [
+    "import os\n",
+    "import h5py\n",
+    "import requests\n",
+    "import pandas as pd\n",
+    "\n",
+    "from PIL import Image\n",
+    "from io import BytesIO\n",
+    "from torchvision import transforms"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": null,
+   "id": "b1447be3-51e3-4900-ad6e-5c48ce79bae2",
+   "metadata": {},
+   "outputs": [],
+   "source": [
+    "data_path = \"data\"\n",
+    "\n",
+    "path_test_df = os.path.join(data_path, \"test-metadata.csv\")\n",
+    "path_test_hdf5 = os.path.join(data_path, \"test-image.hdf5\")"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": null,
+   "id": "01d6b115-c9ad-4ad4-a966-a58f09d5b9de",
+   "metadata": {},
+   "outputs": [],
+   "source": [
+    "test_df = pd.read_csv(path_test_df)\n",
+    "isic_id = test_df.isic_id.values.tolist()\n",
+    "\n",
+    "hdf5_img = h5py.File(path_test_hdf5, 'r')"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": null,
+   "id": "681ae630-d372-440d-a29e-5a68c209943d",
+   "metadata": {},
+   "outputs": [],
+   "source": [
+    "example_image = Image.open(BytesIO(hdf5_img[isic_id[0]][()]))\n",
+    "example_image.show()"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": null,
+   "id": "53e38ac0-bd5f-4c46-a098-f7735c275d8e",
+   "metadata": {},
+   "outputs": [],
+   "source": [
+    "def prepare_request_body(img: Image) -> dict:\n",
+    "    transformations = transforms.Compose([\n",
+    "        transforms.Resize((224, 224)),\n",
+    "        transforms.ToTensor(),\n",
+    "        transforms.Normalize(mean=[0.485, 0.456, 0.406], std=[0.229, 0.224, 0.225])\n",
+    "    ])\n",
+    "    \n",
+    "    image = transformations(img)[None, :]\n",
+    "    \n",
+    "    body = {\n",
+    "      \"inputs\": [\n",
+    "        {\n",
+    "          \"name\": \"input.1\",\n",
+    "          \"shape\": [1, 3, 224, 224],\n",
+    "          \"datatype\": \"FP32\",\n",
+    "          \"data\": image.tolist()\n",
+    "        }\n",
+    "      ]\n",
+    "    }\n",
+    "    \n",
+    "    return body"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": null,
+   "id": "eb1562d1-a944-4b18-b392-5217662e44c1",
+   "metadata": {},
+   "outputs": [],
+   "source": [
+    "URL = \"http://skin-cancer-detection.kubeflow-user-example-com.svc.cluster.local/v2/models/skin_cancer_detection/infer\"\n",
+    "response = requests.post(URL, json=prepare_request_body(example_image))"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": null,
+   "id": "0f2ad89c-74ad-478c-9fc0-612c6ae129e6",
+   "metadata": {},
+   "outputs": [],
+   "source": [
+    "if response.status_code == 200:\n",
+    "    prob = response.json()[\"outputs\"][0][\"data\"][0]\n",
+    "    print(f\"The probability that the lesion is malignant is {prob:.3f}\")"
+   ]
+  }
+ ],
+ "metadata": {
+  "kernelspec": {
+   "display_name": "Python 3 (ipykernel)",
+   "language": "python",
+   "name": "python3"
+  },
+  "language_info": {
+   "codemirror_mode": {
+    "name": "ipython",
+    "version": 3
+   },
+   "file_extension": ".py",
+   "mimetype": "text/x-python",
+   "name": "python",
+   "nbconvert_exporter": "python",
+   "pygments_lexer": "ipython3",
+   "version": "3.8.10"
+  }
+ },
+ "nbformat": 4,
+ "nbformat_minor": 5
+}