README.md

Using the Search Backend REST API

Starting ElasticSearch and the API

First install docker and docker-compose

To persist the elasticsearch database, create a folder [your/directory/on/your/local/]. Use chmod 777 [your/directory/on/your/local/] to set the correct permissions for elastic search.

Also, select a port number PORT_ES (default 9200) which is used to connect to elasticsearch and a port number PORT_API (default 9880) to connect to the search app API.

Create a docker-compose-local.yml file containing the following:

version: "2"
services:
  search_app:
    ports:
      - "PORT_API:8880"
  es:
    volumes:
      - "[your/directory/on/your/local/]:/usr/share/elasticsearch/data"
    ports:
      - "PORT_ES:9200"

The first time, you need to build the images using:

$ docker-compose -f docker-compose.yml -f docker-compose-local.yml  build

By typing -f docker-compose.yml and then -f docker-compose-local.yml the data from the local file overwrites the default.

Then you can start the services using:

$ docker-compose -f docker-compose.yml -f docker-compose-local.yml  up

ElasticSearch will be accessible on your local machine at 127.0.0.1:PORT_ES.

Add docs to the index

Create a data folder in the root directory. Add AMiner_sample.jsonl,wikipedia_taxonomy.json,wikipedia_taxonomy.xml, acm_ccs.xml files to the data/ folder. You can find these files on gdrive (link inside Trello)

$ python scripts/add_docs.py

You can parametrise the script by passing other jsonl file with the --data_path param.

Connect to the API

The API accepts the following requests:

/search

Method: POST

Input:

{
    "query": <your-query>
}

Output:

{
    "results": {
        "hits": {
            "hits": [list-of-es-document-dicts]
        }
    }
}