I'd like to find movies to watch from Amazon Prime Instant Video, but I hate the Amazon UI. It's cluttered with videos that aren't Prime eligible, shows only a few covers at a time without any useful metadata, and doesn't provide useful ways to filter results. I wrote some scripts to scrape the data from Amazon's website, put it into an Elasticsearch server, and make it browse-able with a simple web front end.
Requirements:
- a web server
- Recline library (https://github.com/okfn/recline)
- Elasticsearch backend (https://github.com/okfn/elasticsearch.js/)
- Elasticsearch (http://www.elasticsearch.org/)
- Python 2.7.x with BeautifulSoap, pyelasticsearch, and selenium modules
- create the
primeindex - run
setup.pyto create thevideomapping - run
prime_video.pyto scrape data from Amazon's site into Elasticsearch index; re-run weekly to get new data - run
add_metadata.pyto update video documents with metadata; repeat until no records have a missing metadata field
- copy
webappdirectory to web server - copy recline to
webapp_dir/lib/ - copy elasticsearch.js to
webapp_dir/lib/ - point browser to
webserver/index.html
- scroll left nav with content
- color queue button on click in Chrome
- remove from queue filtered list on de-select
- hide regular version if HD version available
- add not found genre if no metadata
- better metadata?
- get data from recently added (7 days)
- verify that videos are still available in Prime