DEPRECATED

Go to the main Replica repo folder instead.

ReplicaFinder

This is a Flask web service that allows indexing of images in order to make them searchable.

Principles :

• All images are indexed by the web url at which they are accessible, usually encoded plainly as image_url in the json or directly into the API url.
• Arbitrary metadata can be linked to an image.
• All operations require data as json for the request and give a json answer as well.

State of features in progress can be seen here

API

It is divided into two parts database which are operations to modify the database and search for the searching operations.

<web-server-url>/database

• POST

Body : json with fields :

• image_url
• (optional) webpage_url : webpage describing the image if there is any
• (optional) metadata : see below for precisions
• (optional) origin : if you hesitate, leave it blank. Will default to web-app

Returns : json with fields :

• id: unique id of the inserted element

<web-server-url>/database/url/<image-url>

• GET

Body : none

Returns : json describing the image

• id

• image_url

• metadata

• webpage_url if available

• origin

• features features computed for this image

• indexes search indexes it is registered in

• PUT

Body : json with only metadata field to be replaced

• DELETE

Body : none

<web-server-url>/database/id/<id>

See above

<web-server-url>/search/urls

• POST

Body : json with fields :

• positive_image_urls : list of positive examples
• (optional) negative_image_urls : list of negative examples
• (optional) nb_results : number of results to be returned. Default : 30

<web-server-url>/search/ids

• POST

Body : json with fields :

• positive_ids : list of positive examples
• (optional) negative_ids : list of negative examples
• (optional) nb_results : number of results to be returned. Default : 30

<web-server-url>/search/img

• POST

Body : json with fields :

• image : a base64 encoding of the image file that needs to be matched, preferably jpeg
• (optional) nb_results : number of results to be returned. Default : 30

Important : all the search requests might change in the future by returning a result_id that could be queried later to check the results. This if searches appear to take a long time and/or more ressources than expected.

Metadata structure

The metadata is a completely free JSON object stored. However, we recommend some guidelines :

• author : FAMILY NAME, First Name format preferably.
• title : title of the image.
• date : string representation of the date. Can be a year (1563, 1897), or a timeframe (1567-1679, 1550-1650).

author and title will be used for a text search on the metadata. date might be used later.

Any other field is possible as long as it is JSON valid, use explicit and self-sufficient naming though.

Examples

• Adding an image from the Web Gallery to the database

POST <web-server-url>/database

{
  "image_url": "http://www.wga.hu/art/l/leonardo/04/0monalis.jpg",
  "webpage_url": "http://www.wga.hu/frames-e.html?/html/l/leonardo/04/1monali.html",
  "metadata" : {
      "author": "LEONARDO da Vinci",
      "title": "Mona Lisa (La Gioconda)",
      "date": "1503-5",
      "WGA_id": 153647
  }
}

• Searching similar paintings

POST /search/urls

{
  "positive_image_urls": ["http://www.wga.hu/art/l/leonardo/04/0monalis.jpg"]
}