This crate provides a pure-Rust implementation of Jonathan Snook's spam detection algorithm for blog comments.
As described in the afore-linked post, it works on a points system. Points are awarded and deducted based on a variety of rules. If a comments final score is greater than or equal to 1, the comment is considered valid. If the comments final score is 0 then it's considered to be worth of moderating. If the comments final score is below 0 then it's considered to be spam. Each comment starts with a score of 0.
If you're using Cargo, just add Snooker to your Cargo.toml
:
toml
[dependencies]
snooker = "0.1.0"
Snooker gives the example comment below a score of -10 based off of the following patterns:
body
has less that 2 links in it: +2 pointsbody
is more that 20 characters long but contains 1 link: +1 pointbody
contains one keyword considered spammy ("free"): -1
pointbody
contains one phrase considered spammy ("limited time only"): -1
pointbody
starts with a word considered spammy when it's the first word of
the comment ("nice"): -10 pointsauthor
field doesn't contain http://
or https://
: +0 points
(unchanged)url
field contains a keyword considered spammy ("free"): -1 point```rust use snooker::{Comment, Snooker, Status};
let comment = Comment { author: Some("Johnny B. Goode".tostring()), url: Some("http://my-free-ebook.com".tostring()), body: String::from("
Nice post! Check out our free (for a limited time only) eBook here that's totally relevant
"), previouslyacceptedforemail: None, previouslyrejectedforemail: None, previouscommentbodies: None, };let snookerresult = Snooker::new(comment); asserteq!(snookerresult.score, -10); asserteq!(snooker_result.status, Status::Spam); ```
Snooker is released under the MIT LICENSE
.
This crate was written by Elliot Jackson.