04-azure-cosmosdb.md

Module 4: Azure CosmosDB

The completion of this module will result in you connecting your Azure Function to a CosmosDB instance running MongoDB and reading the data from there.

Create an Azure Cosmos DB Account

1. In the portal, create a new Azure Cosmos DB Account. Select MongoDB or SQL API for your account. Create a new database named admin and a collection names Recipes
2. In VS Code, go to the extensions panel and install CosmosDB extension. Alternatively, go to the marketplace
3. Sign in and explore the account you have just created. Update the Recipes collection to include the two documents from https://tacofancy.blob.core.windows.net/data/data.json

Read data from Azure Functions

1. Install mongoose npm package

npm i mongoose

2. Create new folder called shared to place code that's going to be shared in the future:

   • mongo.js

     const mongoose = require('mongoose');
     /**
      * Set to Node.js native promises
      * Per http://mongoosejs.com/docs/promises.html
      */
     mongoose.Promise = global.Promise;
     
     const auth = {
       user: process.env.user,
       password: process.env.password
     };
     
     const mongoUri = `mongodb://${process.env.user}.documents.azure.com:${
       process.env.port
     }/?ssl=true`;
     
     let client = null;
     function connect() {
       mongoose.set('debug', true);
       return new Promise((resolve, reject) => {
         if (client == null) {
           mongoose
             .connect(
               mongoUri,
               { auth: auth }
             )
             .then(_client => {
               client = _client;
               resolve(_client);
             })
             .catch(err => reject(err.status));
         } else {
           resolve(client);
         }
       });
     }
     
     module.exports = {
       connect,
       mongoose
     };

   • recipe.model.js

     const mongoose = require('mongoose');
     const RecipeSchema = new mongoose.Schema(
       {
         category: {
           type: String,
           enum: [
             'SHELLS',
             'BASE_LAYERS',
             'MIXINS',
             'SEASONINGS',
             'CONDIMENTS',
             'FULL_TACOS'
           ]
         },
         title: String,
         description: String,
         ingredients: [String],
         steps: [String],
         tags: [String]
       },
       {
         collection: 'Recipes',
         read: 'nearest'
       }
     );
     const Recipe = mongoose.model('Recipe', RecipeSchema);
     module.exports = Recipe;

   • recipe.service.js

     const Recipe = require('./recipe.model');
     const ReadPreference = require('mongodb').ReadPreference;
     require('./mongo').connect();
     
     function getRecipes() {
       var countQuery = Recipe.count();
       const docQuery = Recipe.find({}).read(ReadPreference.NEAREST);
       return new Promise((resolve, reject) => {
         docQuery
           .exec()
           .then(recipes => {
             if (recipes.length === 0)
               resolve('There are no results matching your query.');
             else {
               countQuery.exec().then(count => {
                 resolve({ items: recipes, totalItems: count });
               });
             }
           })
           .catch(err => {
             reject(err.status);
           });
       });
     }
     
     module.exports = {
       getRecipes
     };

3. In index.js import recipe.service and make request to read data:

   const recipes = require('../shared/recipe.service');
   module.exports = async function(context, req) {
     context.log('JavaScript HTTP trigger function processed a request.');
     await recipes
       .getRecipes()
       .then(data => {
         context.res = {
           body: data
         };
       })
       .catch(err => {
         context.res = {
           status: 500,
           body: err
         };
       });
   };

4. Update local.setting.json to include your database config:

   {
     "IsEncrypted": false,
     "Values": {
       "AzureWebJobsStorage": "<function app storage connection string>",
       "FUNCTIONS_WORKER_RUNTIME": "node",
       "user": "<database user>",
       "password": "<database primary password>",
       "port": "10255"
     },
     "Host": {
       "CORS": "*"
     }
   }

5. Run and check that UI still works