Add all scaffolding, design, and decision documentation AKA first commit

src-tauri/src/diarization/mod.rs

@@ -0,0 +1,40 @@
+//! Speaker diarization (Phase 4, FR-SPK-*). sherpa-onnx: segmentation +
+//! embedding + clustering, fully offline (ADR-0005).
+//!
+//! Runs as a post-stop pass over the recorded WAV (audio = source of truth),
+//! then aligns speaker spans to transcript segments by timestamp overlap.
+//! Speaker IDs are internal ("S1"…); names are applied at render time and never
+//! destructively rewritten onto segments (FR-SPK-5).
+
+use crate::models::{SpeakerSpan, TranscriptSegment};
+use std::path::Path;
+
+#[derive(Debug, thiserror::Error)]
+pub enum DiarError {
+    #[error("model load failed: {0}")]
+    Load(String),
+    #[error("diarization failed: {0}")]
+    Run(String),
+}
+
+pub trait Diarizer: Send + Sync {
+    /// Partition audio into speaker spans.
+    fn diarize(&self, wav: &Path) -> Result<Vec<SpeakerSpan>, DiarError>;
+    /// Assign speaker labels to transcript segments by overlap with spans.
+    fn assign(&self, segments: &mut [TranscriptSegment], spans: &[SpeakerSpan]);
+}
+
+#[cfg(feature = "diarization")]
+pub struct SherpaDiarizer;
+
+#[cfg(feature = "diarization")]
+impl Diarizer for SherpaDiarizer {
+    fn diarize(&self, _wav: &Path) -> Result<Vec<SpeakerSpan>, DiarError> {
+        // T4.1: sherpa-onnx segmentation + embedding + clustering via FFI.
+        todo!("Phase 4 — diarize")
+    }
+    fn assign(&self, _segments: &mut [TranscriptSegment], _spans: &[SpeakerSpan]) {
+        // T4.2: timestamp-overlap alignment.
+        todo!("Phase 4 — assign speakers to segments")
+    }
+}