train_and_deploy_your_first_model_on_lambda.ipynb

{
 "cells": [
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# Train and Deploy Your First Machine Learning Model on Amazon SageMaker"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Create SageMaker session\n",
    "\n",
    "A SageMaker session needs to be initialized in order to start interacting the SageMaker service."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 15,
   "metadata": {},
   "outputs": [],
   "source": [
    "import boto3\n",
    "import re\n",
    "\n",
    "import os\n",
    "import numpy as np\n",
    "import pandas as pd\n",
    "import sagemaker as sage\n",
    "import json\n",
    "import pickle\n",
    "import sys\n",
    "import traceback\n",
    "\n",
    "from sklearn import ensemble\n",
    "from sklearn.metrics import precision_score, recall_score\n",
    "from sklearn.model_selection import train_test_split\n",
    "\n",
    "boto_session = boto3.Session(profile_name=\"up-sagemaker\")\n",
    "session = sage.Session(boto_session=boto_session)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Train and save the model locally"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 16,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Starting the training.\n",
      "Training complete.\n"
     ]
    }
   ],
   "source": [
    "def _save_model(clf):\n",
    "    with open(os.path.join(\"model/\", 'model.pkl'), 'wb') as out:\n",
    "        pickle.dump(clf, out)\n",
    "        \n",
    "def train():\n",
    "    print('Starting the training.')\n",
    "    try:\n",
    "        data_df = pd.read_csv(\"data/iris.csv\", header=None)\n",
    "\n",
    "        # labels are in the first column\n",
    "        labels = data_df.iloc[:, 0]\n",
    "        features = data_df.iloc[:, 1:]\n",
    "\n",
    "        train_features, test_features, labels_train, labels_test = \\\n",
    "            train_test_split(\n",
    "                features, labels, test_size=0.3, random_state=42\n",
    "            )\n",
    "\n",
    "        max_leaf_nodes = 100\n",
    "\n",
    "        clf = ensemble.RandomForestClassifier(\n",
    "            max_leaf_nodes=max_leaf_nodes\n",
    "        )\n",
    "        clf = clf.fit(train_features, labels_train)\n",
    "\n",
    "        test_predictions = clf.predict(test_features)\n",
    "\n",
    "        precision = precision_score(\n",
    "            labels_test, test_predictions, average='macro'\n",
    "        )\n",
    "        recall = recall_score(\n",
    "            labels_test, test_predictions, average='macro'\n",
    "        )\n",
    "\n",
    "        clf = clf.fit(features, labels)\n",
    "        _save_model(clf)\n",
    "\n",
    "        print('Training complete.')\n",
    "    except Exception as e:\n",
    "        print(\n",
    "            'An Exception during training: ' + str(e) + '\\n'\n",
    "        )\n",
    "train()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Upload the model to S3 bucket \n",
    "\n",
    "Using the SageMaker session, which was initialized earlier, the model will be upload to S3. "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 17,
   "metadata": {},
   "outputs": [],
   "source": [
    "local_model_directory = 'model'\n",
    "s3_prefix = 'up-lambda-iris-model'\n",
    "\n",
    "model_location = session.upload_data(local_model_directory, key_prefix=s3_prefix)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Deploy the model\n",
    "\n",
    "Using the trained `Estimator`, a RESTful service will be initialized on Amazon SageMaker service."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 8,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "-----------!"
     ]
    }
   ],
   "source": [
    "from sagemaker.predictor import json_serializer\n",
    "\n",
    "predictor = estimator.deploy(1, 'ml.m4.xlarge', serializer=json_serializer)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Sample some data and use it for a prediction"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 5,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "{'feat_1': [5.8, 5.4, 5.0, 6.4, 5.7, 5.8, 5.4, 4.8, 5.0, 5.5],\n",
       " 'feat_2': [2.7, 3.9, 3.4, 3.2, 2.5, 2.6, 3.9, 3.1, 3.0, 4.2],\n",
       " 'feat_3': [3.9, 1.7, 1.6, 5.3, 5.0, 4.0, 1.3, 1.6, 1.6, 1.4],\n",
       " 'feat_4': [1.2, 0.4, 0.4, 2.3, 2.0, 1.2, 0.4, 0.2, 0.2, 0.2]}"
      ]
     },
     "execution_count": 5,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "data_df = pd.read_csv(\"data/iris.csv\", header=None, names=[\"feat_1\", \"feat_2\", \"feat_3\", \"feat_4\"])\n",
    "\n",
    "sampled_df = data_df.sample(10)\n",
    "\n",
    "sampled_post_request_body = sampled_df.to_dict(orient='list')\n",
    "\n",
    "sampled_post_request_body"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 6,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "{\"predictions\": [\"versicolor\", \"setosa\", \"setosa\", \"virginica\", \"virginica\", \"versicolor\", \"setosa\", \"setosa\", \"setosa\", \"setosa\"]}\n"
     ]
    }
   ],
   "source": [
    "print(predictor.predict(sampled_post_request_body).decode('utf-8'))"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Delete the endpoint"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 7,
   "metadata": {},
   "outputs": [],
   "source": [
    "session.delete_endpoint(predictor.endpoint)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": []
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": []
  }
 ],
 "metadata": {
  "kernelspec": {
   "display_name": "Python 3",
   "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": "3.6.10"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 2
}