FinServer

Builds a REST interface for data stored in a MongoDB collection.

Uses environment variables in dev.env to learn where the database is. Defaults to one on the cloud.

Hosting

Locally:

This starts the REST server and uses cloud db: NODE_ENV=development node index.js or ./startRESTServer.sh

This starts the REST server and uses a local db: MONGODB_URI_JAF=mongodb://127.0.0.1:27017/FinKittyData NODE_ENV=development node index.js or ./startLocalRESTServer.sh

On Heroku

The Procfile defines the actions of the corresponding Heroku app. Heroku uses 'dynos' which run this Node app, list for REST calls and can access MongoDB.

Set appropriate environment variables (for MongoDB access and encryption) in the Heroku app at https://dashboard.heroku.com/. Useful Heroku commands:

heroku login
git push heroku master
heroku ps
heroku ps:restart
heroku logs --tail

The REST interface

The API

The interface is defined according to routes defined in FinServer/src/routes/finkitty.route.js Roughly speaking, the routes are as follows:

A GET request, given a userID, obtains a list of the modelNames for that user.
A GET request, given a userID and a modelName, obtains the model data.
A PUT request, given a userID, a modelName and model data, writes as an update to an existing document.
A POST request, given a userID, a modelName and model data, creates a document with that data.
A DELETE request, given a userID and a modelName, deletes any matching document.

The actions

The work undertaken by this app is defined in FinServer/src/controllers/finkitty.controller.js Roughly speaking, the actions are:

For a model list, obtain the userID as a query parameter, get the model names from MongoDB, make a list and return.
For a model, obtain the userID and modelName as query parameters, get the model from MongoDB, decrypt and return.
For update, get the userID, modelName and model from the request body. Encrypt the model and save.
For create, get the userID, modelName and model from the request body. Encrypt the model and save. Expect failure for duplicates.
For delete, get the userID and modelName from the request body. Delete any matching document.

The MongoDB data

Data structure

Each document in MongoDB is stored according to a schema defined in FinServer/src/models/FinKittyModel.model.js Roughly speaking, the schema consists of

a user ID
a model name
the model data itself

These are strings.

Data location

The app uses an environment variable (with mongo username, password and cluster completed) to know where to go for data. The cloud location is structured like this: MONGODB_URI_JAF=mongodb+srv://<username>:<password>@<cluster>.mongodb.net/

To run mongo locally: On macos, need a data folder (default one no nlonger works since Catalina). Give permissions: sudo chown -R `id -un` Documents/git/FinKitty/mongo/ Note a new database (a new data folder) needs a uniqueness index added; read on and run the createIndex command in the mongo console.

Run the mongodb server mongod --dbpath Documents/git/FinKitty/mongo/

Explore the mongo data in a different terminal mongo there are many mongo commands; try

use FinKittyData $ switches to the right database
show collections $ lists these, expect to see finkittymodels
db.stats() $ expect 1 collection and some count of objects
db.finkittymodels.stats() $ more detailed stats
db.finkittymodels.deleteMany({FinKittyModelName:"Simple"}) $ if the uniqueness breaks you'll have to delete duplicates before adding uniqueness key
db.finkittymodels.createIndex({ "FinKittyUserID": 1,   "FinKittyModelName": -1 }, { unique: true }) $ app requires this otherwise switching to an existing model simply creates a new one with same name

A new mongo database will have a database called test. This app uses a database called FinKittyData (note capitalisation). And inside that, we use a collection called finkittymodels (note capitalisation) and its keys are FinKittyUserID, FinKittyModelName, FinKittyModel. Each model is a document in that collection.

Data encryption

The app encrypts any model information before write to MongoDB, and decrypts it after reading from MongoDB.
An encryption secret needs to be stored as an environment variable

MONGO_CRYPT_SECRET=<my_secret>

Data integrity

Setting up a MongoDB collection:

Network access: MongoDB clusters can have restrictions on access to certain IP addresses, so failure to connect can be due to attempting to connect from a not-included IP address
Unique data keys: MongoDB will maintain uniqueness if it's given keys.
This app requires the following key to be applied:

{
  "FinKittyUserID": 1,
  "FinKittyModelName": -1,
}
{
  unique: true
}

jeanflower / FinServer