Struct KeywordClassifier 

pub struct KeywordClassifier<T> { /* private fields */ }

Aho-Corasick based keyword classifier.

Given a set of rules (keywords, classification) ordered by priority, classify scans the input in a single O(n) pass and returns the classification of the first (highest-priority) keyword that matches as a substring. All matching is case-insensitive – keywords are lowercased at insert time and input bytes are lowercased inline during the scan (no allocation).

Example

use spider_browser::retry::keyword_classifier::KeywordClassifier;

let classifier: KeywordClassifier<&str> = KeywordClassifier::new(&[
    (&["blocked", "captcha", "403"], "blocked"),
    (&["timeout", "err_connection_reset"], "transient"),
]);

assert_eq!(classifier.classify("Error 403 Forbidden"), Some(&"blocked"));
assert_eq!(classifier.classify("ERR_CONNECTION_RESET"), Some(&"transient"));
assert_eq!(classifier.classify("all good"), None);

Implementations

impl<T: Clone> KeywordClassifier<T>

pub fn new(rules: &[(&[&str], T)]) -> Self

Build a new classifier from priority-ordered rules.

Each rule is (keywords, classification). Rules are checked in order; if two rules contain overlapping keywords, the first rule’s classification wins.

All keywords are stored lowercased internally.

pub fn classify(&self, text: &str) -> Option<&T>

Classify a string by scanning it once for all keywords.

Returns the classification of the highest-priority matching keyword, or None if no keyword matches.

Runs in O(n) where n = text.len() with inline ASCII lowercasing (no heap allocation).