Skip to main content

Webapp for the cold start recommender

Project description

*******************************************
A web app to use the cold-start-recommender
*******************************************

This is a basic [tornado](http://www.tornadoweb.org/) application which provide a web interface to the [cold-start-recommender](https://github.com/elegans-io/cold-start-recommender)

# Dependencies

* [tornado](http://www.tornadoweb.org/)
* [cold-start-recommender](https://github.com/elegans-io/cold-start-recommender) version >= 4.0.0

# Installation

pip install csrec-webapp

# Running the application

csrec_webapp.py --port=8888

# APIs

## Item data Input

Items must have at least the following fields:

* A unique identifier. No matter if it is a string or an int, but must be unique.
* A type identifier. The recommender must be able to recognise which type of item is dealing with (program, series, news etc).
* Informative fields. Prior to use, the recommender must know which fields are the informative ones

## User behaviour data input

The application must tell which action is performed by which user on which item.

## Output

For each user, the recommender can recommend different types of items.

Note that:

* The recommender does not consider which programs/series are scheduled for a certain day.
* Items for a certain period must be filtered by the user after the query.
* The recommender can filter the type of items (eg if programs or series or news should be recommended to a certain user), but cannot recommend the informative fields (eg "this user likes running shoes").
* Whenever not enough information is available, the recommender recommends the most popular items.

## Functions

### Update items table

Insert or modify items.

#### methods

POST

#### URL Params

##### Required

* unique_id: the name of the field used as unique id for the item

#### Data Params

A list of items with type and category informations e.g.:

```
[{ "_id" : "item1", "type": "lady", "category" : "high_heels"},
{ "_id" : "item2", "type": "male", "category" : "mocassino"},
{ "_id" : "item3", "type": "unisex", "category" : "mocassino"},
{ "_id" : "item4", "type": "male", "category" : "mocassino"}]
```

#### Success Response

Code: 200

Content: {}

#### Error Response

Code: 404

Content: {}

#### Sample Call

```bash
curl -X POST -H "Content-Type: application/json" -d '[{ "_id" : "item1", "type": "lady", "category" : "high_heels"}, { "_id" : "item2", "type": "male", "category" : "mocassino"}, { "_id" : "item3", "type": "unisex", "category" : "mocassino"}, { "_id" : "item4", "type": "male", "category" : "mocassino"}]' 'http://elegans.it:8000/insertitems?unique_id=_id'
```

### Post user action on an item

#### methods

POST

#### URL Params

##### Required

* item: the item id (program, series etc)
* user: This could be the value of rating, or could correspond to different actions (e.g. hated: -1, loved: +1)
* code: the rating

##### Optional

* only_info: true or false, default is false

#### Description

Update/insert an action identified by code performed by {uid} on {item_id}. If users or item are not found they are created in the database.

#### Success Response

Code: 200

Content: {}

#### Error Response

Code: 404

Content: {}

#### Sample Call

```bash
curl -X POST -H "Content-Type: application/json" -d '{ "item_info" : ["type", "category"]}' 'http://elegans.it:8000/itemaction?item=item1&user=User1&code=1&only_info=false'
```

### Post a social action (users to users)

#### methods

POST

#### Description

Stores information about action performed by {uid} on {other_uid}, like "follow", "like" etc. If id's are not found they are created in the database.

#### URL Params

##### Required

* user: email or session_id of the user. NB Always use email if available.
* code: This could be the liking factor, or could correspond to different actions (e.g. downvote: -1, follow: +1)
* user_to: the other user's id

#### Success Response

Code: 200

Content: {}

#### Error Response

Code: 404

Content: {}

#### Sample Call

```bash
curl -X POST 'http://elegans.it:8000/socialaction?user=User1&user_to=User2&code=3'
```

### Get recommended items for a user

#### methods

GET

#### Description

Provide a list of recommended item_ids.

#### URL Params

##### Required

* user: email or session_id of the user. NB Always use email if available.

##### Optional

* fast: if set to any value, uses faster and less accurate algorithm
* limit: integer, number of items to return, default is 10
* type: type of items to be returned. NB: if a non-existent key is provided the recommender will return an empty list

#### Success Response

Code: 200

Content: a list of items e.g.: ```["item4", "item1", "item3", "item2", "User2"]```

#### Error Response

Code: 404

Content: {}

#### Sample Call

```bash
curl -X GET 'http://elegans.it:8000/recommend?user=User1&limit=10'
```

### Reconcile session_id with user ID

#### methods

POST

#### Description

Whenever a user logs in not in the first session, the app should tell which session_id s/he was using during the previous sessions. All action associated to user_old will be associated to user_new.

#### URL Params

##### Required

* user_old: old user id
* user_new: new user id

#### Success Response

Code: 200

Content: {}

#### Error Response

Code: 404

Content: {}

#### Sample Call

```bash
curl -X POST 'http://elegans.it:8000/reconcile?user_old=User1&user_new=User2
```

### Get action of a user

#### methods

GET

#### Description

Return a dictionary with two lists:

#### URL Params

* itemaction: actions performed on items
* socialaction: actions performed on other users

#### URL Params

##### Required

* user: email or session_id of the user. NB Always use email if available..

#### Success Response

Code: 200

Content: {}

#### Error Response

Code: 404

Content: {}

#### Sample Call

```bash
curl -X GET 'http://elegans.it:8000/info/user?user=User1'
```

### Get interactions on an item

#### methods

GET

#### Description

List of user who performed any action on the item, and which action

#### URL Params

##### Required

* item: id of the item

#### Success Response

Code: 200

Content: {}

#### Error Response

Code: 404

Content: {}

#### Sample Call

```bash
curl -X GET 'http://localhost:8000/info/item?item=item1'
```

### Serialize the data on files

#### methods

GET

#### Description

serialize data on file

#### URL Params

##### Optional

* filename: the path where to serialize the data, default is /tmp/dump.bin

#### Success Response

Code: 200

Content: {}

#### Error Response

Code: 404

Content: {}

#### Sample Call

```bash
curl -X GET 'http://localhost:8000/serialize?filename=/tmp/dump.bin'
```

### Restore data from file

#### methods

GET

#### Description

restore data from file

#### URL Params

##### Optional

* filename: the path of the file which contains the serialized data

#### Success Response

Code: 200

Content: {}

#### Error Response

Code: 404

Content: {}

#### Sample Call

```bash
curl -X GET 'http://localhost:8000/restore?filename=/tmp/dump.bin'
```

Project details


Download files

Download the file for your platform. If you're not sure which to choose, learn more about installing packages.

Source Distribution

cold-start-recommender-webapp-0.4.1.tar.gz (6.7 kB view details)

Uploaded Source

File details

Details for the file cold-start-recommender-webapp-0.4.1.tar.gz.

File metadata

File hashes

Hashes for cold-start-recommender-webapp-0.4.1.tar.gz
Algorithm Hash digest
SHA256 8000a28afef8748d988d6383f00f67ed0d228be8d4b0beb923a09bf545992b59
MD5 a61f207adcefb82911220b6e1c1b36ec
BLAKE2b-256 794c9d3c553107f3db9fee9af9d8d4b5233e14a1ec5251aea20c032e2b7566c2

See more details on using hashes here.

Supported by

AWS AWS Cloud computing and Security Sponsor Datadog Datadog Monitoring Fastly Fastly CDN Google Google Download Analytics Microsoft Microsoft PSF Sponsor Pingdom Pingdom Monitoring Sentry Sentry Error logging StatusPage StatusPage Status page