Feed experiment

feed.py

import datetime

from elasticsearch import Elasticsearch


INDEX = 'feed'
today = datetime.date.today()


es = Elasticsearch()


if es.indices.exists(INDEX):
    es.indices.delete(INDEX)


# Create index.
es.indices.create(INDEX, {
    'settings': {
        'index': {
            'number_of_replicas': 0
        }
    }
})


# Populate some test data.
#
# Let's make up 3 content types. Let's assume a blog type app that combines
# blog posts, flickr photos, and twitter posts.

# Blog posts
for i in range(1, 31, 4):  # 30 days back, a blog post ea 4 days.
    body = {
        'created': today - datetime.timedelta(days=i),
        'title': 'Blog %s' % i,
        'content': 'foo',
    }
    es.index(INDEX, 'blog', body=body)

# Photos
for i in range(1, 31, 2):  # 30 days back, a photo ea 2 days.
    body = {
        'created': today - datetime.timedelta(days=i),
        'img': '/some/url/to/image/%s' % i,
        'caption': 'foo',
    }
    es.index(INDEX, 'photo', body=body)

# Tweets
for i in range(1, 31):  # 30 days back, a tweet erryday.
    body = {
        'created': today - datetime.timedelta(days=i),
        'tweet': 'tweet %s' % i,
    }
    es.index(INDEX, 'tweet', body=body)


# Refresh index.
es.indices.refresh(INDEX)


# Get the feed, 25 items.
res = es.search(INDEX, body={
    'query' : {
        'match_all': {}
    },
    'sort': {
        'created': {'order': 'desc'}
    },
    # If things were given regions to appear in, we could do a region filter
    # in here too.
    'size': 25
})


print 'Feed Items:'
for obj in res['hits']['hits']:
    print "Doctype: %s" % obj['_type']
    print "Source: %s" % obj['_source']
    print '-' * 72