Struct DocumentSummarizer 

pub struct DocumentSummarizer { /* private fields */ }

Production-quality document summarizer supporting five summarization strategies.

Implementations

impl DocumentSummarizer

pub fn new(config: SummarizerConfig) -> Self

Create a new summarizer with the supplied configuration.

pub fn with_defaults() -> Self

Create a summarizer with a default extractive (3-sentence) configuration.

pub fn stats(&self) -> &SummarizerStats

Return a reference to the accumulated statistics.

pub fn summarize( &mut self, text: &str, embeddings: Option<Vec<Vec<f64>>>, ) -> Result<SummaryResult, SummarizerError>

Summarize text using the configured strategy.

embeddings, when provided, must contain one vector per sentence in document order and must all share the same dimension.

pub fn score_sentence( &self, sentence: &str, index: usize, total: usize, corpus: &[Vec<String>], ) -> SentenceScore

Score a single sentence.

corpus holds the tokenized form of every sentence in the document (used for IDF).

pub fn extract_keyphrases(&self, text: &str, n: usize) -> Vec<String>

Extract the top-n n-gram keyphrases from text.

pub fn chunk_document( &self, text: &str, chunk_size: usize, ) -> Vec<DocumentChunk>

Split text into overlapping chunks of approximately chunk_size characters.

Chunks overlap by 10% of chunk_size to preserve context across boundaries.

pub fn quality_score(&self, original: &str, summary: &str) -> f64

Compute a quality score in [0, 1] as the fraction of original’s top keyphrases that appear (as substrings) in summary.