Fraudpointer is a service and an application that can help online shops protect from fraudsters. More about this on Fraudpointer.
Special Features of the Application
- Multiple Accounts
- Multiple Users per Account
- Multiple Roles per User
- User configurable Rules to identify fraud cases
- Fraud Case Management
- Cases are Queued to Accept/Reject/Review and to various Profiles
- Dynamic Black lists/White Lists and any other kind of Lists - user configurable
- Real-time graphs of Fraud Case Management
- Any data can be fed into Fraudpointer database and be used for fraud evaluation
- Attributes of various types are supported (String, Numbers, Boolean, Dates, Countries, Emails)
- Calculated attributes
- Velocity Attributes
- Reputation Database
- Similarity Matching of Cases
- Google Apps Integration (docs, email, contacts)
- Matlab Integration
- Linux distribution (Debian 6)
- Amazon RDS - MySQL
- Ruby on Rails
- Hosted on Amazon Cloud
Ruby on Rails Special Considerations
- Full internationalization
- acts_as_list: Sortable list of objects
- kaminari: Used for pagination
- redis: Used for background asynchronously long running tasks, processed out of the web request cycle
- memcached: Used for memory caching.
- state_machine: Used to implement a state machine
- jQuery and jQuery UI
- Online payments. Integration with payment gateway using activemerchant.
- Google Apps Market Integration. With the help of various gems:
- wicked-pdf: PDf Generation
- paper_trail. Tracks changes on data.
- ar-octopus: MySQL Master/Slave Replication
- Test Unit
- validator_attachment. Used to check whether specific validators are attached to a model.
- capybara. For testing the UI.]
- [cucumber](http://github.com/cucumber/cucumber/tree/master] for Acceptance Tests
- new relic. Used for real-time process monitoring.
- sunspot_rails. Integration with Solr
- About 10 months
- About 65 models
- About 60 tables
- About 40 controllers
- About 2900 tests
- About 6200 lines of code