Flask uploading file to s3 with Boto

Skip to Step 5 if you have an existing app from class and you want to add file upload support.

Our requirements.txt

Flask==0.10
Flask-mongoengine==0.7.0
boto==2.17.0

Changes to App.py & Models.py

We're only requiring Flask-mongoengine

Since we started with Databases and Form validation we have been including both mongoengine and flask-mongoengine. This was bad form on my part.

Mongoengine and Flask-mongoengine do the same thing but Flask-mongoengine is a little better for our needs.

app.py updates

Top of the file currently, we imported mongoengine

# import all of mongoengine
from mongoengine import *

Change this to...

# import all of mongoengine
from flask.ext.mongoengine import mongoengine

We connected to the database a few lines from the top with

# --------- Database Connection ---------
# MongoDB connection to MongoLab's database
connect('mydata', host=os.environ.get('MONGOLAB_URI'))

Change this to...

mongoengine.connect('mydata', host=os.environ.get('MONGOLAB_URI'))

models.py updates

The import and from/import statements at the top of models.py is currently

from mongoengine import *
from flask.ext.mongoengine.wtf import model_form
from datetime import datetime

Change to this...

from flask.ext.mongoengine.wtf import model_form
from wtforms.fields import * # for our custom signup form
from flask.ext.mongoengine.wtf.orm import validators
from flask.ext.mongoengine import *
from datetime import datetime

We are updating the imports to use more field validation techniques.

Our models change a bit too. Previously we defined Models as such,

class Blogpost(Document):
	title = StringField(max_length=120, required=True, verbose_name="First name")
	description = StringField(required=True)
	...

We now define models like so,

class Blogpost(mongoengine.Document):
	title = mongoengine.StringField(max_length=120, required=True)
	description = mongoengine.StringField()
	...

We create forms like this

photo_form = model_form(Image)

But we can also use WTForms directly like this,

# Create a WTForm form for the photo upload.
# This form will inhirit the Photo model above
# It will have all the fields of the Photo model
# We are adding in a separate field for the file upload called 'fileupload'

class photo_upload_form(photo_form):
	fileupload = FileField('Upload an image file', validators=[])

photo_upload_form can be used like any other form we use before. Using WTForm directly gives us a few more options than we previously had. We will use this again in the User management app demo.

Getting Started

Step 1 : Download code, setup Git, heroku create

1. Download the sample code from Github

2. Navigate to code directory in Terminal. Create Git repo

    git init
    git add .
    git commit -m "initial commit"
   

   Create virtual environment and install requirements

    virtualenv venv
    . runpip
   

3. Create a new Heroku app

    heroku create
   

Step 2 : Adding MongoLabs to your Heroku App

Heroku offers a lot of different Add-ons for your apps. Many different types of databases, image tools, cache utilities are available from 3rd party companies. Many offer a trial plan to test and develop with before you commit to a paid plan.

MongoLabs offers a 496 MB MongoDB instance for free (see here) : ) How wonderful.

To install the MongoLabs

• Navigate to the code folder of your app

• In Terminal, add the MongoLab sandbox plan

    heroku addons:add mongolab:sandbox
  

This has added MongoLab to your app.

Step 3: Configure your local environment

When adding Add-ons, Heroku will add the required configuration variables for the services including username, password, urls, etc.

We must create a local configuration file to allow our local development server to connect to the MongoLabs MongoDB instance. We can grab a copy of our Heroku configuration variables and put them inside a .env file, our environment variable file.

Run the following command inside your code folder.

heroku config --shell | grep MONGOLAB_URI >> .env

This will create a new file, .env and it will contain a single line starting with MONGOLAB_URI and followed by a long connection url. This is the username and password for your MongoLabs account.

.env

MONGOLAB_URI=mongodb://heroku_app8083291:sadlfkwewe........

Step 4: Add .env to .gitignore file

We want GIT to ignore the .env file, VERY IMPORTANT This keeps our environment variables safe and they won't get included inside our GIT repository (or worse, get pushed to Github).

Open your .gitignore file and add '.env' on a new line. Save the file.

.gitignore

.env
venv
*.pyc

Step 5: Register with Amazon Web Services

Create an account on Amazon Web Services, you can use your Amazon account, http://aws.amazon.com/console/ Click on the Sign Up button.

Step 6: Log in and create new S3 bucket

When you're registered and logged into the AWS site, visit the console, https://console.aws.amazon.com/console/home?#

In the Storage and Content Delivery section click on S3, scalable storage in the cloud, https://console.aws.amazon.com/s3/home.

Now we will create a bucket (like a directory). The bucket will be the container for your uploaded files. On the left panel of the S3 console, click 'Create Bucket'. Provide a bucket name and leave the Region to US Standard. Then click Create.

Step 7: Add environment variables to .env and Heroku

Inside the AWS Console, on the top menu bar click on your name, then click SECURITY CREDENTIALS.

On the SECURITY CREDENTIALS page, you will have access to

• ACCESS KEY ID
• SECRET ACCESS KEY

Open your .env file and add 3 new variables for Amazon AWS

.env

AWS_BUCKET=YOURBUCKNAME
AWS_ACCESS_KEY_ID=XXXXXXXXXXXX
AWS_SECRET_ACCESS_KEY=XXXXXXXXXXX
SECRET_KEY=SOMETHINGSECRETFORFLASK

Save your .env file.

Now let's push the new AWS variable to Heroku config, run the commands in Terminal

heroku config:add SECRET_KEY=SOMETHINGSECRETFORFLASK
heroku config:add AWS_BUCKET=YOURBUCKNAME
heroku config:add AWS_ACCESS_KEY_ID=XXXXXXXXXXXX
heroku config:add AWS_SECRET_ACCESS_KEY=XXXXXXXXXXXXX

You can confirm the AWS variables are on heroku by running the command,

heroku config

Start your servers

. start

or

foreman start