This library provides PHP implementation of Keboola writer to DataBreakers DataAPI.

Clone the git repository and use Composer to install dependencies:

$ git clone github/... .
$ composer install

Run the script php src/run.php --data=/data with path to data parameter. In Keboola enviroment the data path is /data.

Script needs a standard Keboola configuration file placed in /data/configuration.json or /data/configuration.yml with following custom parameters.

{
    "parameters": {
        "databreakers_account": DATABREAKERS_ACCOUNT_ID,
        "databreakers_secret_key": DATABREAKERS_API_SECRET,
        "delete_old_data": true|false
    },
}

You need to get DATABREAKERS_ACCOUNT_ID and DATABREAKERS_API_SECRET first to use the writer. Please contact keboola@databreakers.com if you don't have your account yet. delete_old_data signifies if all actual data in your DataBreakers account will be deleted or you want to append new data to account actual dataset.

Writer operates with three entites: items, users, interactions. Item entities are... User entities are... Interaction entites are...

To upload a set of entites into DataBreakers API, you have to create csv exports named accordingly to entity type: users.csv, items.csv, interactions.csv.

items.csv has to have PK defined on one column. Then it can have as many attributes (columns) as you need. Name of every column (except PK) will correspod to item's attributes of items in DataBreakers dataset. This file is not required, you can use it to add explicit attributes to items.

users.csv has to have PK defined on one column. Then it can have as many attributes (columns) as you need. Name of every column (except PK) will correspod to user's attributes of users in DataBreakers dataset. This file is not required, you can use it to add explicit attributes to users.

interactions.csv has to have exactly 4 columns named as user_id, item_id, timestamp, interaction_id and PK on all these four columns. Default interaction_id set is defined as [Detail view, Purchase, Recommendation, Dislike, Like, Bookmark]. You can have your own extended set if you need, just contact support. This file is required. It will create empty user or item entity if user_id or item_id is not presented in your dataset.

DataBreakers – we are your data sense