Skip to content

main.md

Latest commit

 

History

History
69 lines (48 loc) · 5.73 KB

main.md

File metadata and controls

69 lines (48 loc) · 5.73 KB

Databases and Docker intro

Today we're going to continue talking about databases and how to integrate them into your projects. Databases offer centralized persistent storage with some fantastic garuntees about data consistency, availability, and/or partition tolerance (generally you only get two out of the three, unless you're Google Spanner).

MongoDB

MongoDB is a non-relational "NoSQL" database that is essentially an object store -- it lets you build structured objects and put them into storage with easy queries for retrieval later. Generally MongoDB is a schemaless database (unlike SQL databases), which means you generally do not have to specify a data structure schema for objects before inserting them into collections. This isn't great for programming because we'd like validation on our data models we're putting into and getting out of the database, so we're going to use pymodm to help us interface with MongoDB in that manner.

MongoDB is just a program (like your flask server) that runs on a machine and receives requests from folks looking to insert entities into a database or query for existing entities in a database. This MongoDB program can be running on your machine, or it could be running on a completley different machine. To know how to reach a MongoDB instance, we need a URL (just like you do for flask). If you're running MongoDB on the same machine as your flask server, the URL your flask server will use will look something like this mongodb://127.0.0.1:27017/bme590. If it's running on another machine, say on mlab, the url would look something like this: mongodb://<dbuser>:<dbpassword>@ds157223.mlab.com:57223/bme590. Notice on mlab the databases have a username and password that you must provide in the URL itself.

Let's look through a simple example of using a database from a python program:

👀 To see a detailed example walkthrough of using pymodm see the example jupyter notebook 👀

from pymodm import connect
from pymodm import MongoModel, fields

connect("mongodb://localhost:27017/bme590") # connect to database

class User(MongoModel):
    email = fields.EmailField(primary_key=True)
    first_name = fields.CharField()
    last_name = fields.CharField()
    password = fields.CharField()

u = User('user1@email.com', last_name='Ross', first_name='Bob')
u2 = User('user2@email.com', last_name='Ross', first_name='Rob')

u.save()
u2.save()

for user in User.objects.raw({"first_name":"Rob"}):
        print(user)
	print(user.first_name)
	print(user.last_name)

How to create a MongoDB instance

You have many options to setup your own mongodb instance:

  • You can set it up using a free service like mlab. See instructions here.
  • You can simply install and run MongoDB on your machine (see community edition here )
  • You can install MongoDB easily using Docker, if you want to learn Docker. Instructions are in the Docker section below.

MySQL/postgres

An example of a postgres project is here. For SQL databases, we store data across tables and specify relationships across those tables (as discussed in the previous database lecture). If you use MySQL or Postgres in this class, you will not have to use an ORM or schema manager since SQL databases impose their internal schemas.

Docker

To install Docker, follow these instructions for mac and for ubuntu and finally if you really have to, for windows.

Docker allows us to run programs (with all of their dependencies) in a consistent environment called a container. You can think of this container as a "virtual machine" that runs on your computer, like a mini linux computer running inside your own (although it's way more efficient than that). You can push and pull containers that have all of the programs and dependencies already setup to a central place called DockerHub.

Some common docker commands:

  • docker pull mongo this pulls down the full mongodb database container from DockerHub
  • docker run mongo this runs the docker container and the program (mongodb) that comes with it.
  • docker run -it mongo bash this starts up the mongodb container, and opens up bash shell inside the container (kind of like ssh-ing into the container.
  • docker exec -it <container_id> bash -- "ssh" into a currently running container.

When we run containers, we can bind container ports to our local machine ports and even bind parts of our filesystem into the container.

In the future, we'll explore Docker further, and chat about how to create your own containers for your software.

Run MongoDB using Docker

Below is the minimal example of a program that connects to a MongoDB database called bme590 running on localhost:27017. To get a database running using Docker (which we will introduce in class) simply run

docker run -v $PWD/db:/data/db -p 27017:27017 mongo

which will run the mongo container, binding the container's 27017 port to your computer's 27017 and binding the current directory's folder db to be the place where data is stored inside the container.