Deploy R in online mode - complete example notebook

r-online-deployment-demo.ipynb

{
 "cells": [
  {
   "cell_type": "code",
   "execution_count": 6,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "'SUCCESS'"
      ]
     },
     "execution_count": 6,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "import requests\n",
    "import ibm_watson_machine_learning\n",
    "\n",
    "SPACE_ID = \"4a9d3d67-8208-4564-a78f-bebdebf864f8\"\n",
    "API_KEY = \"xxx\"\n",
    "CPD_URL = \"https://cpd-cpd.itzroks-550003aw18-72fhj7-6ccd7f378ae819553d37d5f2ee142bd6-0000.au-syd.containers.appdomain.cloud\"\n",
    "wml_credentials = {\n",
    "    \"username\": \"admin\",\n",
    "    \"apikey\": API_KEY,\n",
    "    \"url\": CPD_URL,\n",
    "    \"instance_id\": \"openshift\",\n",
    "    \"version\": \"4.0\",\n",
    "}\n",
    "wml_client = ibm_watson_machine_learning.APIClient(wml_credentials)\n",
    "wml_client.set.default_space(SPACE_ID)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 26,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Creating package extensions\n",
      "SUCCESS\n",
      "SUCCESS\n"
     ]
    },
    {
     "data": {
      "text/plain": [
       "'SUCCESS'"
      ]
     },
     "execution_count": 26,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "config_yaml = \"\"\"\n",
    "name: custom\n",
    "channels:\n",
    "  - conda-forge\n",
    "dependencies:\n",
    "  - r\n",
    "  - libiconv\n",
    "\"\"\"\n",
    "with open(\"config.yaml\", \"w\", encoding=\"utf-8\") as f:\n",
    "    f.write(config_yaml)\n",
    "\n",
    "# create a custom software specification\n",
    "# note, you can also create a custom software specificaton using the user interface from Watson Machine Learning.\n",
    "\n",
    "software_spec_uid = wml_client.software_specifications.get_uid_by_name(\"custom-r\")\n",
    "if software_spec_uid != \"Not Found\":\n",
    "    software_spec_details = wml_client.software_specifications.get_details(software_spec_uid)\n",
    "    package_ext_uid = software_spec_details[\"entity\"][\"software_specification\"][\"package_extensions\"][0][\"metadata\"][\"asset_id\"]\n",
    "    wml_client.package_extensions.delete(package_ext_uid)\n",
    "    wml_client.software_specifications.delete(software_spec_uid)\n",
    "\n",
    "meta_props = {wml_client.package_extensions.ConfigurationMetaNames.NAME: \"custom-r\", wml_client.package_extensions.ConfigurationMetaNames.TYPE: \"conda_yml\"}\n",
    "package_ext_details = wml_client.package_extensions.store(meta_props=meta_props, file_path=\"config.yaml\")\n",
    "package_ext_uid = wml_client.package_extensions.get_uid(package_ext_details)\n",
    "meta_props = {\n",
    "    wml_client.software_specifications.ConfigurationMetaNames.NAME: \"custom-r\",\n",
    "    wml_client.software_specifications.ConfigurationMetaNames.BASE_SOFTWARE_SPECIFICATION: {\"guid\": wml_client.software_specifications.get_uid_by_name(\"runtime-22.1-py3.9\")},\n",
    "}\n",
    "software_spec_details = wml_client.software_specifications.store(meta_props=meta_props)\n",
    "software_spec_uid = wml_client.software_specifications.get_uid(software_spec_details)\n",
    "wml_client.software_specifications.add_package_extension(software_spec_uid, package_ext_uid)\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 29,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\n",
      "\n",
      "#######################################################################################\n",
      "\n",
      "Synchronous deployment creation for uid: '7981d407-f0a9-42c4-9d42-d5c71534ee6b' started\n",
      "\n",
      "#######################################################################################\n",
      "\n",
      "\n",
      "initializing\n",
      "Note: online_url is deprecated and will be removed in a future release. Use serving_urls instead.\n",
      "...........................................................................................................................\n",
      "ready\n",
      "\n",
      "\n",
      "------------------------------------------------------------------------------------------------\n",
      "Successfully finished deployment creation, deployment_uid='d5e98822-5461-4397-b067-8f033f0d8e10'\n",
      "------------------------------------------------------------------------------------------------\n",
      "\n",
      "\n"
     ]
    },
    {
     "data": {
      "text/plain": [
       "{'entity': {'asset': {'id': '7981d407-f0a9-42c4-9d42-d5c71534ee6b'},\n",
       "  'custom': {},\n",
       "  'deployed_asset_type': 'function',\n",
       "  'hardware_spec': {'id': 'b128f957-581d-46d0-95b6-8af5cd5be580',\n",
       "   'name': 'XXS',\n",
       "   'num_nodes': 1},\n",
       "  'name': 'iris xgb deployment',\n",
       "  'online': {},\n",
       "  'space_id': '4a9d3d67-8208-4564-a78f-bebdebf864f8',\n",
       "  'status': {'online_url': {'url': 'https://cpd-cpd.itzroks-550003aw18-72fhj7-6ccd7f378ae819553d37d5f2ee142bd6-0000.au-syd.containers.appdomain.cloud/ml/v4/deployments/d5e98822-5461-4397-b067-8f033f0d8e10/predictions'},\n",
       "   'serving_urls': ['https://cpd-cpd.itzroks-550003aw18-72fhj7-6ccd7f378ae819553d37d5f2ee142bd6-0000.au-syd.containers.appdomain.cloud/ml/v4/deployments/d5e98822-5461-4397-b067-8f033f0d8e10/predictions'],\n",
       "   'state': 'ready'}},\n",
       " 'metadata': {'created_at': '2022-06-07T05:02:45.598Z',\n",
       "  'id': 'd5e98822-5461-4397-b067-8f033f0d8e10',\n",
       "  'modified_at': '2022-06-07T05:02:45.598Z',\n",
       "  'name': 'iris xgb deployment',\n",
       "  'owner': '1000330999',\n",
       "  'space_id': '4a9d3d67-8208-4564-a78f-bebdebf864f8'},\n",
       " 'system': {'warnings': [{'id': 'Deprecated',\n",
       "    'message': 'online_url is deprecated and will be removed in a future release. Use serving_urls instead.'}]}}"
      ]
     },
     "execution_count": 29,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "params = {\n",
    "    \"wml_credentials\": wml_credentials, \n",
    "    \"space_id\": SPACE_ID,\n",
    "    \"script\": \"predict.R\",\n",
    "    \"required_assets\": [\"iris_xgb.model\"],\n",
    "    \"r_packages\": [\"jsonlite\", \"xgboost\"],\n",
    "}\n",
    "\n",
    "def r_function(params=params):\n",
    "    \"\"\"\n",
    "    Deployable Python function which downloads the required assets from a WML deployment space and runs the specified R Script.\n",
    "    \"\"\"\n",
    "    import json\n",
    "    import subprocess\n",
    "    from ibm_watson_machine_learning import APIClient\n",
    "\n",
    "    wml_client = APIClient(params[\"wml_credentials\"])\n",
    "    wml_client.set.default_space(params[\"space_id\"])\n",
    "    script = params[\"script\"]\n",
    "    assets = {x[\"metadata\"][\"name\"]: x[\"metadata\"][\"asset_id\"] for x in wml_client.data_assets.get_details()[\"resources\"]}\n",
    "\n",
    "    # download required assets\n",
    "    wml_client.data_assets.download(assets[script], script)\n",
    "    for x in params[\"required_assets\"]:\n",
    "        wml_client.data_assets.download(assets[x], x)\n",
    "    \n",
    "    # install R and other libraries\n",
    "    subprocess.run([\"Rscript\", \"-e\", 'install.packages(c(\"jsonlite\", \"data.table\", \"stringi\", \"stringr\"), repos=\"https://cloud.r-project.org\")'])\n",
    "    if \"xgboost\" in params[\"r_packages\"]:\n",
    "        params[\"r_packages\"].remove(\"xgboost\")\n",
    "        subprocess.run([\"Rscript\", \"-e\", 'install.packages(\"https://cloud.r-project.org/src/contrib/Archive/xgboost/xgboost_1.5.2.1.tar.gz\", repos=NULL, type=\"source\")'])\n",
    "    subprocess.run([\"Rscript\", \"-e\", 'install.packages(c(' + \",\".join([f'\"{x}\"' for x in params[\"r_packages\"]]) + '), repos=\"https://cloud.r-project.org\")'])\n",
    "\n",
    "    def score(payload):\n",
    "        d = payload[\"input_data\"][0][\"values\"][0]\n",
    "        inputs = d[\"inputs\"]\n",
    "        inputs = json.dumps(inputs)\n",
    "        result = subprocess.run([\"Rscript\", \"--vanilla\", script, inputs], stdout=subprocess.PIPE, stderr=subprocess.PIPE)\n",
    "        out = result.stdout.decode(\"utf-8\")\n",
    "        try:\n",
    "            out = json.loads(out)\n",
    "        except:\n",
    "            out = result.stdout.decode(\"utf-8\")\n",
    "        err = result.stderr.decode(\"utf-8\") \n",
    "        return {\"predictions\": [{\"out\": out, \"err\": err}]}\n",
    "\n",
    "    return score\n",
    "\n",
    "meta_props = {\n",
    "    wml_client.repository.FunctionMetaNames.NAME: \"iris xgb\",\n",
    "    wml_client.repository.FunctionMetaNames.SOFTWARE_SPEC_ID: wml_client.software_specifications.get_uid_by_name(\"custom-r\"),\n",
    "}\n",
    "function_details = wml_client.repository.store_function(function=r_function, meta_props=meta_props)\n",
    "function_uid = wml_client.repository.get_function_id(function_details)\n",
    "meta_props = {\n",
    "    wml_client.deployments.ConfigurationMetaNames.NAME: \"iris xgb deployment\",\n",
    "    wml_client.deployments.ConfigurationMetaNames.ONLINE: {},\n",
    "}\n",
    "wml_client.deployments.create(function_uid, meta_props=meta_props)\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 33,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "[{'setosa': 0.3561,\n",
       "  'versicolor': 0.3218,\n",
       "  'virginica': 0.3221,\n",
       "  'probability': 0.3561,\n",
       "  'prediction': 'setosa'},\n",
       " {'setosa': 0.3561,\n",
       "  'versicolor': 0.3218,\n",
       "  'virginica': 0.3221,\n",
       "  'probability': 0.3561,\n",
       "  'prediction': 'setosa'},\n",
       " {'setosa': 0.3561,\n",
       "  'versicolor': 0.3218,\n",
       "  'virginica': 0.3221,\n",
       "  'probability': 0.3561,\n",
       "  'prediction': 'setosa'}]"
      ]
     },
     "execution_count": 33,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "endpoint = \"https://cpd-cpd.itzroks-550003aw18-72fhj7-6ccd7f378ae819553d37d5f2ee142bd6-0000.au-syd.containers.appdomain.cloud/ml/v4/deployments/d5e98822-5461-4397-b067-8f033f0d8e10/predictions?version=2022-06-07\"\n",
    "\n",
    "inputs = [\n",
    "    {\"Sepal.Length\": 5.1, \"Sepal.Width\": 3.5, \"Petal.Length\": 1.4, \"Petal.Width\": 0.2},\n",
    "    {\"Sepal.Length\": 4.9, \"Sepal.Width\": 3.1, \"Petal.Length\": 1.4, \"Petal.Width\": 0.2},\n",
    "    {\"Sepal.Length\": 5.3, \"Sepal.Width\": 6.1, \"Petal.Length\": 1.4, \"Petal.Width\": 0.2},\n",
    "]\n",
    "\n",
    "payload = {\"input_data\": [{\"values\": [{\"inputs\": inputs}]}]}\n",
    "headers = {\"Content-Type\": \"application/json\", \"Accept\": \"application/json\", \"Authorization\": wml_client._get_headers()[\"Authorization\"]}\n",
    "\n",
    "response = requests.post(url=endpoint, headers=headers,json=payload, verify=False).json()\n",
    "response[\"predictions\"][0][\"out\"]\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": []
  }
 ],
 "metadata": {
  "interpreter": {
   "hash": "2fd8ffa6155755da6e46d79320aeff097111fde3e362ae0d899ff1703ff23477"
  },
  "kernelspec": {
   "display_name": "Python 3.10.4 ('ml')",
   "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": "3.10.4"
  },
  "orig_nbformat": 4
 },
 "nbformat": 4,
 "nbformat_minor": 2
}