hapiger
Hapi.js HTTP wrapper for Good Enough Recommendation engine (GER)
Last updated 5 years ago by grahamjenson .
MIT · Repository · Bugs · Original npm · Tarball · package.json
$ cnpm install hapiger 
SYNC missed versions from official npm registry.

HapiGER

HapiGER logo

Providing good recommendations can create greater user engagement and directly provide value by recommending items the customer might additionally like. However, many applications don't provide recommendations to users because of the difficulty in implementing a custom engine or the pain of using an off-the-shelf engine.

HapiGER is a recommendations service that uses the Good Enough Recommendations (GER), a scalable, simple recommendations engine, and the Hapi.js framework. It has been developed to be easy to integrate, easy to use and very scalable.

Project Site

Quick Start Guide


*** #### Install HapiGER

Install with npm

npm install -g hapiger

***

Start HapiGER

By default it will start with an in-memory event store (events are not persisted)

hapiger

There are also PostgreSQL and RethinkDB event stores for persistence and scaling


***

Create a Namespace

A Namespace is a bucket where all the events are put:

curl -X POST 'http://localhost:3456/namespaces' -d'{
    "namespace": "movies"
  }'

***

Create some Events

An event occurs when a person actions something, e.g. Alice views Harry Potter:

curl -X POST 'http://localhost:3456/events' -d '{
    "events": [
    {
      "namespace": "movies",
      "person":    "Alice",
      "action":    "view",
      "thing":     "Harry Potter"
    }
  ]
}'

Then, Bob also views Harry Potter (now Bob has similar viewing habits to Alice)

curl -X POST 'http://localhost:3456/events' -d '{
    "events": [
    {
      "namespace": "movies",
      "person":    "Bob",
      "action":    "view",
      "thing":     "Harry Potter"
    }
  ]
}'

When a person actions and thing, it serves two purposes in HapiGER:

  1. It is used to measure a persons similarity to other people
  2. It can be a recommendation of that thing

For example, when Bob buys LOTR

curl -X POST 'http://localhost:3456/events' -d '{
    "events": [
    {
      "namespace":  "movies",
      "person":     "Bob",
      "action":     "buy",
      "thing":      "LOTR",
      "expires_at": "2016-10-12"
    }
  ]
}'

This is an action that can be used to find similar people AND it can be seen as Bob recommending LOTR. For an event to be a recommendation as well it must have an expiration date set with expires_at, which is how long the recommendation will be available for.


***

Recommendations

What books should Alice buy?

curl -X POST 'http://localhost:3456/recommendations' -d '{
    "namespace": "movies",
    "person": "Alice",
    "configuration": {
      "actions" : {"view": 5, "buy": 10}
    }
}'
{
  "recommendations": [
    {
      "thing": "LOTR",
      "weight": 0.44721359549996,
      "last_actioned_at": "2015-10-12T17:04:14+01:00",
      "last_expires_at": "2016-10-12T01:00:00+01:00",
      "people": [
        "Bob"
      ]
    }
  ],
  "neighbourhood": {
    "Bob": 0.44721359549996,
    "Alice": 1
  },
  "confidence": 0.00036398962692384
}

Alice should buy The Hobbit as it was recommended by Bob with a weight of about 0.2.

The configuration defines many variables that can be used to customise the search for recommendations. The object is directly passed to GER and the available variables are listed in the GER Documentation.

The confidence of these recommendations is pretty low because there are not many events in the system


***

How HapiGER Works (the Quick Version)

The HapiGER API calculates recommendations for Alice to buy by:

  1. Finding people (neighbors) that are like Alice by looking at her past events
  2. Calculating the similarities between Alice and her neighbors
  3. Looking at the recent things that those similar people recommended
  4. Weight those recommendations using the similarity of the people

***

Event Stores

The "in-memory" memory event store is the default, this will not scale well or persist event so is not recommended for production.

The recommended event store is PostgreSQL, which can be used with:

hapiger --es pg --esoptions '{
    "connection":"postgres://localhost/hapiger"
  }'

Options are passed to knex.

HapiGER also supports a RethinkDB event store:

hapiger --es rethinkdb --esoptions '{
    "host":"127.0.0.1",
    "port": 28015,
    "db":"hapiger"
  }'

Options passed to rethinkdbdash.


***

Compacting the Event Store

The event store needs to be regularly maintained by removing old, outdated, or superfluous events; this is called compacting

curl -X POST 'http://localhost:3456/compact' -d '{
  "namespace": "movies"
}'

***

Namespaces

In addition to creating namespaces, you can also list and destroy them:

curl -X GET 'http://localhost:3456/namespaces'

To delete a namespace (and all its events!):

curl -X DELETE 'http://localhost:3456/namespaces/movies'

***

Changelog

12/10/15 -- Updated README, new version of GER 8/02/15 -- Updated readme and bumped version

Current Tags

  • 0.0.12                                ...           latest (5 years ago)

4 Versions

  • 0.0.12                                ...           5 years ago
  • 0.0.11                                ...           5 years ago
  • 0.0.2                                ...           5 years ago
  • 0.0.1                                ...           5 years ago
Maintainers (1)
Downloads
Today 0
This Week 0
This Month 0
Last Day 0
Last Week 0
Last Month 0
Dependencies (13)
Dev Dependencies (3)
Dependents (0)
None

Copyright 2014 - 2016 © taobao.org |