Skip to content
This repository has been archived by the owner on Jul 24, 2020. It is now read-only.

ponyriders/django-amazon-price-monitor

Repository files navigation

Build Status codecov.io Requirements Status Stories in Ready Code Health

Monitors prices of Amazon products via Product Advertising API.

Necessary settings changes for upcoming version 0.8 can be found in the history . UPDATE: 0.8 will never come, see the end of life announcement below.

!!! IMPORTANT !!!

Since January 23rd 2019 the efficiency guidelines for using the Product Advertising API have changed dramatically. Requests limits for the API are now calculated based on the revenue generated with the Amazon Associate account the Product Advertising API is connected with.

You can find the details here: https://docs.aws.amazon.com/AWSECommerceService/latest/DG/TroubleshootingApplications.html#efficiency-guidelines

For django-amazon-price-monitor this means the END OF LIFE from now on.

But why?

We started the project to solve the issue of saving money with Amazon while having control over your data by self hosting this project. To continue maintaining and extending the project we'd need to have enough revenue in our associate account, which we do not have. The result is that our API account used for developing this project is blocked and we cannot continue. Also our private little instance of pricemonitor, used by some of our friends and us, will be shut down, too.

The changes in the guidelines make it pretty hard for small projects that do not primarily focus on generating revenue with Amazon to continue. As the django-amazon-price-monitor did not get that much love in the last months and years from us as we focused on other things, the only logic consequence for us is to pull the plug.

Thanks to all involved for participating in this journey with us!

Best

Alex + Martin

This is a reusable Django app that can be plugged into your project. It consists basically of this parts:

  • Models
  • Frontend components
  • Angular Frontend API
  • Amazon API component
  • Product
    • representation of an Amazon product
  • Price
    • representation of a price of an Amazon product at a specific time
  • Subscription
    • subscribe to a product at a specific price with an email notification

The frontend displays all subscribed products with additional information and some graphs for price history.

The features are the following:

  • list products
  • show product details
  • show product price graphs
  • add subscriptions
  • adjust subscription price value
  • delete subscriptions

Simply the API consumed by AngularJS, based on Django REST Framework.

Fetches product information from Amazon Product Advertising API through several tasks powered by Celery and weaves the data into the models.

This software is licensed with the MIT license. So feel free to do with it whatever you like.

Python 3.4 3.5 3.6
Django 1.8, 1.9, 1.10, 1.11 1.9, 1.10, 1.11 1.11

For additional used packages see setup.py.

Add the following apps to INSTALLED_APPS:

INSTALLED_APPS = (
    ...
    'price_monitor',
    'price_monitor.product_advertising_api',
    'rest_framework',
)

Then migrate:

python manage.py migrate

Adjust the settings appropriately, see next chapter.

Include the url configuration.

Setup celery - you'll need the beat and a worker.

The values of the following displayed settings are their default values. If the value is '...' then there is no default value.

The following settings are absolutely necessary to the price monitor running, please set them:

You need to have a broker and a result backend set.

CELERY_BROKER_URL = ...
CELERY_RESULT_BACKEND = ...

# some additional settings
CELERY_ACCEPT_CONTENT = ['pickle', 'json']
CELERY_CHORD_PROPAGATES = True

We use Rest-Framework for Angular frontend:

REST_FRAMEWORK = {
    'PAGINATE_BY': 50,
    'PAGINATE_BY_PARAM': 'page_size',
    'MAX_PAGINATE_BY': 100,
}

Specify the base URL under which your site will be available. Defaults to: http://localhost:8000 Necessary for creating links to the site within the notification emails.

# base url to the site
PRICE_MONITOR_BASE_URL = 'https://....'
# your Amazon Web Services access key id
PRICE_MONITOR_AWS_ACCESS_KEY_ID = '...'

# your Amazon Web Services secret access key
PRICE_MONITOR_AWS_SECRET_ACCESS_KEY = '...'

# the region endpoint you want to use.
# Typically the country you'll run the price monitor in.
# possible values: CA, CN, DE, ES, FR, IT, JP, UK, US
PRICE_MONITOR_AMAZON_PRODUCT_API_REGION = '...'

# the assoc tag of the Amazon Product Advertising API
PRICE_MONITOR_AMAZON_PRODUCT_API_ASSOC_TAG = '...'

As the links to Amazon will be affiliate links with your Amazon associate tag (see above), you have to set your name for the disclaimer (see https://partnernet.amazon.de/gp/associates/agreement).

# name of you/your site
PRICE_MONITOR_AMAZON_ASSOCIATE_NAME = 'name/sitename'
# Amazon site being used, choose from on of the following
    'Amazon.co.uk'
    'Local.Amazon.co.uk'
    'Amazon.de'
    'de.BuyVIP.com'
    'Amazon.fr'
    'Amazon.it'
    'it.BuyVIP.com'
    'Amazon.es'
    'es.BuyVIP.com'
PRICE_MONITOR_AMAZON_ASSOCIATE_SITE = '<ONE FROM ABOVE>'
# if to use the HTTPS URLs for Amazon images.
# if you're running the monitor on SSL, set this to True
# INFO:
#  Product images are served directly from Amazon.
#  This is a restriction when using the Amazon Product Advertising API
PRICE_MONITOR_IMAGES_USE_SSL = True

# domain to use for image serving.
# typically analog to the api region following the URL pattern
#  https://images-<REGION>.ssl-images-amazon.com
PRICE_MONITOR_AMAZON_SSL_IMAGE_DOMAIN = 'https://images-eu.ssl-images-amazon.com'

The following settings can be adjusted but come with reasonable default values.

# time after which products shall be refreshed
# Amazon only allows caching up to 24 hours, so the maximum value is 1440!
PRICE_MONITOR_AMAZON_PRODUCT_REFRESH_THRESHOLD_MINUTES = 720  # 12 hours

To be able to send out the notification emails, set up a proper email backend (see Django documentation).

# time after which to notify the user again about a price limit hit (in minutes)
PRICE_MONITOR_SUBSCRIPTION_RENOTIFICATION_MINUTES = 10080  # 7 days

# sender address of the notification email
PRICE_MONITOR_EMAIL_SENDER = 'noreply@localhost'

# currency name to use on notifications
PRICE_MONITOR_DEFAULT_CURRENCY = 'EUR'

# subject and body of the notification emails
gettext = lambda x: x
PRICE_MONITOR_I18N_EMAIL_NOTIFICATION_SUBJECT = gettext(
    'Price limit for %(product)s reached'
)
PRICE_MONITOR_I18N_EMAIL_NOTIFICATION_BODY = gettext(
    'The price limit of %(price_limit)0.2f %(currency)s has been reached for the '
    'article "%(product_title)s" - the current price is %(price)0.2f %(currency)s.'
    '\n\nPlease support our platform by using this '
    'link for buying: %(link)s\n\n\nRegards,\nThe Team'
)

# name of the site in notifications
PRICE_MONITOR_SITENAME = 'Price Monitor'
# key of cache (according to project config) to use for graphs
# None disables caching.
PRICE_MONITOR_GRAPH_CACHE_NAME = None

# prefix for cache key used for graphs
PRICE_MONITOR_GRAPH_CACHE_KEY_PREFIX = 'graph_'

To be able to run the required Celery tasks, Celery itself has to be set up. Please see the Celery Documentation about how to setup the whole thing. You'll need a broker and a result backend configured.

The package comes with an easy to use Docker setup - you just need docker and docker-compose.

There are 5 containers:

db Postgres database
redis Celery broker
web a django project containing the django-amazon-price-monitor package
celery the celery for the django project
data container for mounted volumes

The web and celery containers are using a docker image being set up under docker/web.

Basic image with all necessary system packages and pre-installed lxml and psycopg2. The image can be found on Docker Hub.

It comes with a Django project with login/logout view, that can be found under docker/web/project. The image derives from pricemonitor/base from above.

The directory structure within the container is the following (base dir: /srv/):

root:/srv tree
├── logs                [log files]
├── media               [media files]
├── project             [the django project]
├── static              [static files]
└── pricemonitor        [the pricemonitor package]

Starts via the start script docker/web/web_run.sh that does migrations and the starts the runserver.

Basically the same as web, but starts the Celery worker with beat.

If you want to develop anything involving tasks, see the Usage section below.

The containers mount several paths:

Folder in container Folder on host Information
/var/lib/postgresql/data <PROJECTROOT>/docker/postgres
  • Postgres data directory
  • Keeps the DB data even if container is removed
/srv/logs <PROJECTROOT>/docker/logs Django logs (see project settings)
/srv/media <PROJECTROOT>/docker/media Django media files
/srv/project <PROJECTROOT>/docker/web/project
  • the Django project
  • is copied on Dockerfile to get it up and running
  • then mounted over (the copy is overwritten)
/srv/pricemonitor <PROJECTROOT>
  • the django-amazon-price-monitor lib
  • is copied on Dockerfile to get it up and running
  • then mounted over (the copy is overwritten)

To override some settings as well as to set up the required AWS settings you can create a docker-compose.override.yml and fill with the specific values (also see docker-compose documentation).

Please see or adjust the docker\web\project\glue\settings.py for all settings that are read from the environment. They can be overwritten.

A sample docker-compose.override.yml file could look like this:

version: '3'
services:
  celery:
    command: /bin/true
    environment:
      PRICE_MONITOR_AWS_ACCESS_KEY_ID: XXX
      PRICE_MONITOR_AWS_SECRET_ACCESS_KEY: XXX
      PRICE_MONITOR_AMAZON_PRODUCT_API_REGION: DE
      PRICE_MONITOR_AMAZON_PRODUCT_API_ASSOC_TAG: XXX
      PRICE_MONITOR_AMAZON_PRODUCT_REFRESH_THRESHOLD_MINUTES: 5
      PRICE_MONITOR_SUBSCRIPTION_RENOTIFICATION_MINUTES: 60

It will avoid the automatic startup of celery (command: /bin/true) and set the required settings for AWS (in fact they are only needed in the celery container). You can then manually start the container and execute celery which is quite useful if you develop anything that includes changes in the tasks and thus requires the celery to be restarted (execute from the docker folder!):

alex@tyrion:~/projects/github/django-amazon-price-monitor/docker$ docker-compose run celery bash
Starting docker_data_1


# check environment variables

root@9d64bbd23e98:/srv/project# env
HOSTNAME=9d64bbd23e98
EMAIL_BACKEND=django.core.mail.backends.filebased.EmailBackend
POSTGRES_DB=pm_db
TERM=xterm
PYTHONUNBUFFERED=1
PRICE_MONITOR_SUBSCRIPTION_RENOTIFICATION_MINUTES=60
POSTGRES_PASSWORD=6i2vmzq5C6BuSf5k33A6tmMSHwKKv0Pu
PATH=/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin
SECRET_KEY=Vceev7yWMtEQzHaTZX52
PWD=/srv/project
CELERY_BROKER_URL=redis://redis/1
C_FORCE_ROOT='True'
PRICE_MONITOR_AWS_SECRET_ACCESS_KEY=XXX
POSTGRES_USER=pm_user
SHLVL=1
HOME=/root
PRICE_MONITOR_AMAZON_PRODUCT_REFRESH_THRESHOLD_MINUTES=5
PRICE_MONITOR_AMAZON_PRODUCT_API_REGION=DE
PRICE_MONITOR_AMAZON_PRODUCT_API_ASSOC_TAG=XXX
DEBUG='True'
PRICE_MONITOR_AWS_ACCESS_KEY_ID=XXX
_=/usr/bin/env


# start celery (worker and beat) (can also execute /srv/celery_run.sh)

root@9d64bbd23e98:/srv/project# celery --beat -A glue worker

 -------------- celery@9d64bbd23e98 v3.1.23 (Cipater)
---- **** -----
--- * ***  * -- Linux-3.16.0-4-amd64-x86_64-with-debian-8.0
-- * - **** ---
- ** ---------- [config]
- ** ---------- .> app:         glue:0x7fc6b5269e10
- ** ---------- .> transport:   redis://redis:6379/1
- ** ---------- .> results:     disabled://
- *** --- * --- .> concurrency: 8 (prefork)
-- ******* ----
--- ***** ----- [queues]
 -------------- .> celery           exchange=celery(direct) key=celery

[2016-03-20 10:02:26,776: WARNING/MainProcess] celery@9d64bbd23e98 ready.
cd docker

# start
docker-compose up -d

# stop
docker-compose stop

# inspect
docker-compose logs -f

A fixture with a Django user admin and the password password is loaded automatically.

To build the images, use the Makefile from the root directory.

As the fronted is done by Angular, there is only a single template with very limited possibilities to adjust, price_monitor/angular_index_view.html. You can extends the template and adjust the following blocks.

Is rendered on the very bottom of the page. You have to use Bootstrap compatible markup, e.g.:

{% block footer %}
        <div class="row">
                <div class="col-md-12">Additonal footer</div>
        </div>
{% endblock %}

A management command to batch create a number of products by providing their ASIN:

python manage.py price_monitor_batch_create_products <ASIN1> <ASIN2> <ASIN3>

Recreates a product with the given asin. If product already exists, it is deleted. Only use in development!

python manage.py price_monitor_recreate_product <ASIN>

Searches for products at Amazon (not within the database) with the given ASINs and prints out their details.

python manage.py price_monitor_search <ASIN1> <ASIN2> ...

The app uses the logger "price_monitor" to log all error and info messages that are not included within a dedicated other logger. Please see the Django logging documentation for how to setup loggers.

Logger for everything related to the ProductAdvertisingAPI wrapper class that accesses the Amazon Product Advertising API through bottlenose.

Logger for the utils module.

codecov-graph

Model Graph
Image of Product advertising api synchronization workflow

Image of Product advertising api synchronization workflow