ASIS (Advanced Social Image Search) indexes Flickr and Instagram images and provides a search API across both indexes.
Clone or download
MothOnMars SRCH-272 fix MRSS photo importer (#39)
- use file fixture helper in specs
- update some gems
Latest commit ba853a3 Jan 3, 2019

ASIS Server

CircleCI Code Climate Test Coverage

ASIS (Advanced Social Image Search) indexes Flickr and Instagram images and provides a search API across both indexes.

Current version

You are reading documentation for ASIS API v1.

Contribute to the code

The server code that runs the image search component of is here on Github. Fork this repo to add features like additional datasets, or to fix bugs.


This code is currently tested against Ruby 2.3.


  1. Copy instagram.yml.example to instagram.yml and update the fields with your Instagram credentials.
  2. Copy flickr.yml.example to flickr.yml and update the fields with your Flickr credentials.


We use bundler to manage gems. You can install bundler and other required gems like this:

gem install bundler
bundle install


We're using Elasticsearch (>= 5.6.5) for fulltext search. On a Mac, it's easy to install with Homebrew.

$ brew install elasticsearch@5.6

Otherwise, follow the instructions to download and run it.

You can generally leave the defaults in elasticsearch.yml as they are, with the exception of:


Scripting is disabled by default in Elasticsearch 5.6. Set both script.inline and script.stored to true:

  inline: true
  stored: true

Elasticsearch with Docker

Install Docker if you haven't done so yet. Follow the instruction here Once you have Docker installed on your machine, run the following command from the project root

$ .docker/bin/elasticsearch-docker

This will download an docker image containing elasticsearch=5.6 from, run it, and expose port 9200 & 9300 to your machine. You can verify your setup with the following command.

$ curl localhost:9200
  "name" : ...random string...,
  "cluster_name" : "elasticsearch",
  "cluster_uuid" : ...uuid...
  "version" : {
    "number" : "5.6.5",
    "build_hash" : "6a37571",
    "build_timestamp" : "2017-12-04T07:50:10.466Z",
    "build_snapshot" : false,
    "lucene_version" : "6.6.1"
  "tagline" : "You Know, for Search"


Sidekiq (see below) uses Redis, so make sure you have that installed and running.

Seed some image data

You can bootstrap the system with some government Flickr/Instagram profiles and MRSS feeds to see the system working. Sample lists are in config/instagram_profiles.csv and config/flickr_profiles.csv and config/mrss_profiles.csv.

bundle exec rake oasis:seed_profiles

You can keep the indexes up to date by periodically refreshing the last day's images. To do this manually via the Rails console:


The Capistrano deploy script has whenever hooks so this refresh happens automatically via cron in your production environment.

Running it

Fire up a server and try it all out.

bundle exec rails s

Here are the profiles you have just bootstrapped. Note: Chrome does a nice job of pretty-printing the JSON response.




You can add a new profile manually via the REST API:

	curl -XPOST "http://localhost:3000/api/v1/instagram_profiles.json?username=deptofdefense&id=542835249"
	curl -XPOST "http://localhost:3000/api/v1/flickr_profiles.json?name=commercegov&id=61913304@N07&profile_type=user"
	curl -XPOST "http://localhost:3000/api/v1/mrss_profiles.json?"

MRSS profiles work a little differently than Flickr and Instagram profiles. When you create the MRSS profile, Oasis assigns a short name to it that you will use when performing searches. The JSON result from the POST request will look something like this:

    "created_at": "2014-10-26T18:25:21.167+00:00",
    "updated_at": "2014-10-26T18:25:21.173+00:00",
    "name": "72",
    "id": "http:\/\/\/gallery\/rss.php?cat=all"

Asynchronous job processing

We use Sidekiq for job processing. You can see all your Flickr and Instagram jobs queued up here:


Kick off the indexing process:

bundle exec sidekiq

In the Rails console, you can query each index manually using the Elasticsearch Query DSL:

bin/rails c


These parameters are accepted for the blended search API:

  1. query
  2. flickr_groups (comma separated)
  3. flickr_users (comma separated)
  4. instagram_profiles (comma separated)
  5. mrss_names (comma separated list of Oasis-assigned names)
  6. size
  7. from


The top level JSON contains these fields:

  • total
  • offset
  • suggestion: The overridden spelling suggestion
  • results

Each result contains these fields:

  • type: FlickrPhoto | InstagramPhoto
  • title
  • url
  • thumbnail_url
  • taken_at

API versioning

We support API versioning with the JSON format. The current version is v1. You can specify a specific JSON API version like this:

curl "http://localhost:3000/api/v1/image.json?flickr_groups=1058319@N21&flickr_users=35067687@n04,24662369@n07&instagram_profiles=nasa&mrss_names=72,73&query=earth"


These require an Elasticsearch server and Redis server to be running.

bundle exec rspec

Code coverage

We track test coverage of the codebase over time, to help identify areas where we could write better tests and to see when poorly tested code got introduced.

After running your tests, view the report by opening coverage/index.html.

Click around on the files that have < 100% coverage to see what lines weren't exercised.

Code samples

We "eat our own dog food" and use this ASIS API to display image results on the government websites that use

See the results for a search for moon on


You can send feedback via Github Issues.

Loren Siebert and contributors.