Struct TimestampChunker 

pub struct TimestampChunker { /* private fields */ }

Timestamp-aware chunker for subtitle/transcript content.

Groups subtitle cues into chunks based on time duration rather than character count. Each chunk carries start_secs and end_secs in its metadata for timestamp-aware retrieval and citation.

Falls back to RecursiveChunker for documents without subtitle cue metadata.

Example

use aprender_rag::chunk::{TimestampChunker, Chunker};
use aprender_rag::Document;
use aprender_rag::media::SubtitleCue;

let cues = vec![
    SubtitleCue { index: 0, start_secs: 0.0, end_secs: 30.0, text: "First segment.".into() },
    SubtitleCue { index: 1, start_secs: 30.0, end_secs: 65.0, text: "Second segment.".into() },
    SubtitleCue { index: 2, start_secs: 65.0, end_secs: 90.0, text: "Third segment.".into() },
];

let mut doc = Document::new("First segment. Second segment. Third segment.");
doc.metadata.insert(
    "subtitle_cues".into(),
    serde_json::to_value(&cues).unwrap(),
);
doc.metadata.insert("duration_secs".into(), serde_json::json!(90.0));

let chunker = TimestampChunker::new(60.0);
let chunks = chunker.chunk(&doc).unwrap();
assert!(chunks.len() >= 2);
assert!(chunks[0].metadata.custom.contains_key("start_secs"));

Implementations

impl TimestampChunker

pub fn new(target_duration_secs: f64) -> Self

Create a timestamp chunker with the given target duration.

pub fn with_min_duration(self, secs: f64) -> Self

Set minimum chunk duration.

pub fn with_max_duration(self, secs: f64) -> Self

Set maximum chunk duration.

pub fn with_overlap(self, secs: f64) -> Self

Set overlap duration.

Trait Implementations

impl Chunker for TimestampChunker

fn chunk(&self, document: &Document) -> Result<Vec<Chunk>>

Split document into chunks

fn estimate_chunks(&self, document: &Document) -> usize

Estimate chunk count without materializing

impl Clone for TimestampChunker

fn clone(&self) -> TimestampChunker

Returns a duplicate of the value.

fn clone_from(&mut self, source: &Self)

Performs copy-assignment from source.

impl Debug for TimestampChunker

fn fmt(&self, f: &mut Formatter<'_>) -> Result

Formats the value using the given formatter.

impl Default for TimestampChunker

fn default() -> Self

Returns the “default value” for a type.