README.md

Run

See https://github.com/akdh/cst-tools for JSON schemas.

To run install the required dependencies and launch the server.

pip install flask jsonschema
python main.py

Suggestion requests can now be made to this server, see https://gist.github.com/akdh/564cbb5f9c6c92c99d8a.

Note that you must have docs.db in the same directory as main.py.

You can create this file with by running:

cat lod_docs.sql | sqlite3 docs.db

(collection_2015.csv must be in the current directory when this is run).

load_docs.sql

.mode csv
CREATE TABLE docs (id TEXT PRIMARY KEY, location INT, url TEXT, title TEXT);
.import collection_2015.csv docs

main.py

from flask import Flask, jsonify, request, abort
import random, sqlite3, jsonschema, json

app = Flask(__name__)

def get_ids_for_location(location, db_filename = 'docs.db'):
    conn = sqlite3.connect(db_filename)
    cur = conn.cursor()
    cur.execute('select id from docs where location = ?', (location, ))
    rows = cur.fetchall()
    conn.close()
    return [row[0] for row in rows]

@app.route('/suggestions', methods=['POST'])
def suggestions():
    data = request.get_json()
    with open("request_schema.json") as f:
        schema = json.loads(f.read())
        print(schema)
    jsonschema.validate(schema, data)
    location = data['location']
    person = data['person']
    # Get a set of random IDs for the location
    # TODO: Replace this with a more intelligent recommendation strategy
    ids = get_ids_for_location(location['id'])
    suggestions = random.sample(ids, 10)
    return jsonify({'suggestions': suggestions})

if __name__ == '__main__':
    app.run(host='0.0.0.0', port=5002)