TG-31: Spam filtering

[sebasruii]

31) The system must prevent the principals from storing any data that can be considered spam. A piece of text is considered spam if the number of spam terms exceeds a predefined spam thresh-old. The default list of spam terms includes the following ones: “sex”, “viagra”, “cialis”, “one mil-lion”, “you’ve won”, “nigeria”, and their corresponding translations into the languages considered for internationalisation; the default spam threshold is 10%. Note that the previous default values can be changed at will by the administrators. Realise that a term must be considered spam irre-spective of its case and the blanks in between its words; for instance, “one␣million” is a spam term that matches “one␣million”, “ONE␣MILLION”, “OnE␣␣␣MiLLiOn”, or “One␣⏎␣Mil-lion”; it doesn’t match “One␣Millionaire”, “One␣or␣two␣millions”, or “One␣sex␣million”, though. The spam detector must be reusable across different projects; that is: it must be imple-mented as an independent project that must be packaged into a reusable dependency. Do not forget to deliver your spam detector project so that it can also be evaluated or, otherwise, this requirement shall not be considered fulfilled. This requirement must be fulfilled in this and every subsequent group or individual deliverable.

[sebasruii]

9

[Lucperrom]

7

[nxniya]

8