Struct TfidfFeatureExtractor 

pub struct TfidfFeatureExtractor { /* private fields */ }

TF-IDF feature extractor for commit messages

This extractor converts commit messages into TF-IDF feature vectors for ML classification. Implements Phase 2 of nlp-models-techniques-spec.md (Tier 2: TF-IDF + ML).

Examples

use organizational_intelligence_plugin::nlp::TfidfFeatureExtractor;

let messages: Vec<String> = vec![
    "fix: null pointer dereference".to_string(),
    "fix: race condition in mutex".to_string(),
    "feat: add new feature".to_string(),
];

let mut extractor = TfidfFeatureExtractor::new(1500);
let features = extractor.fit_transform(&messages).unwrap();

assert_eq!(features.n_rows(), 3); // 3 documents

Implementations

impl TfidfFeatureExtractor

pub fn new(max_features: usize) -> Self

Create a new TF-IDF feature extractor

Arguments

• max_features - Maximum number of features (vocabulary size)

Examples

use organizational_intelligence_plugin::nlp::TfidfFeatureExtractor;

let extractor = TfidfFeatureExtractor::new(1500);

pub fn fit_transform(&mut self, messages: &[String]) -> Result<Matrix<f64>>

Fit the vectorizer on training messages and transform them to TF-IDF features

Arguments

• messages - Training commit messages

Returns

• Ok(Matrix<f64>) - TF-IDF feature matrix (n_messages × vocabulary_size)
• Err - If vectorization fails

Examples

use organizational_intelligence_plugin::nlp::TfidfFeatureExtractor;

let messages: Vec<String> = vec![
    "fix: memory leak".to_string(),
    "fix: race condition".to_string(),
];

let mut extractor = TfidfFeatureExtractor::new(1000);
let features = extractor.fit_transform(&messages).unwrap();

assert_eq!(features.n_rows(), 2);

pub fn fit(&mut self, messages: &[String]) -> Result<()>

Fit the vectorizer on training messages

Arguments

• messages - Training commit messages

Examples

use organizational_intelligence_plugin::nlp::TfidfFeatureExtractor;

let messages = vec![
    "fix: memory leak".to_string(),
    "fix: race condition".to_string(),
];

let mut extractor = TfidfFeatureExtractor::new(1000);
extractor.fit(&messages).unwrap();

pub fn transform(&self, messages: &[String]) -> Result<Matrix<f64>>

Transform messages to TF-IDF features using fitted vocabulary

Arguments

• messages - Commit messages to transform

Returns

• Ok(Matrix<f64>) - TF-IDF feature matrix
• Err - If transformation fails

Examples

use organizational_intelligence_plugin::nlp::TfidfFeatureExtractor;

let train_messages = vec![
    "fix: memory leak".to_string(),
    "fix: race condition".to_string(),
];

let test_messages = vec!["fix: null pointer".to_string()];

let mut extractor = TfidfFeatureExtractor::new(1000);
extractor.fit(&train_messages).unwrap();

let features = extractor.transform(&test_messages).unwrap();
assert_eq!(features.n_rows(), 1);

pub fn vocabulary_size(&self) -> usize

Get the vocabulary size (number of features)

Returns

• usize - Number of features in vocabulary

Examples

use organizational_intelligence_plugin::nlp::TfidfFeatureExtractor;

let messages = vec![
    "fix: bug".to_string(),
    "feat: feature".to_string(),
];

let mut extractor = TfidfFeatureExtractor::new(1000);
extractor.fit(&messages).unwrap();

assert!(extractor.vocabulary_size() > 0);
assert!(extractor.vocabulary_size() <= 1000);

pub fn max_features(&self) -> usize

Get the maximum features configuration

Returns

• usize - Maximum number of features