grant/ten31-transcripts
1
0
Fork 0
Code Issues Pull Requests Actions Packages Projects Releases Wiki Activity

Revert adjacent same-speaker segment collapse

Browse Source 
User found the merged transcript lines harder to read — too many sentences
joined into one statement. Remove SpeakerReconciler.mergeAdjacent, its wiring in
finishBackend (restore the no-LLM early return), and its tests. Back to one
segment per diarized utterance.
This commit is contained in:
Grant Gilliam
2026-06-08 15:52:27 -05:00
parent d9e8c6ceab
commit 5c80e827a1
3 changed files with 8 additions and 70 deletions
Download Patch File Download Diff File Expand all files Collapse all files
Ten31TranscriptsTests/SpeakerReconcilerTests.swift
-32
View File
@@ -73,38 +73,6 @@ final class SpeakerReconcilerTests: XCTestCase {
XCTAssertTrue(out.speakers.contains { $0.name == "Me" }) // self never dissolved
}
func testMergeAdjacentCollapsesSameSpeakerAndJoinsText() {
let f = file([sp("A", "content"), sp("B", "content")], [
SpeakersFile.Segment(start: 0, end: 1, speaker: "A", text: "I"),
SpeakersFile.Segment(start: 1.5, end: 4, speaker: "A", text: "need to switch it back"),
SpeakersFile.Segment(start: 4.2, end: 6, speaker: "B", text: "Sure"),
])
let out = SpeakerReconciler.mergeAdjacent(f, maxGap: 2.0)
XCTAssertEqual(out.segments.count, 2) // two A's collapsed
XCTAssertEqual(out.segments[0].speaker, "A")
XCTAssertEqual(out.segments[0].start, 0, accuracy: 0.001)
XCTAssertEqual(out.segments[0].end, 4, accuracy: 0.001)
XCTAssertEqual(out.segments[0].text, "I need to switch it back")
XCTAssertEqual(out.segments[1].speaker, "B") // different speaker untouched
}
func testMergeAdjacentRespectsMaxGapAndSpeakerBoundaries() {
let f = file([sp("A", "content")], [
SpeakersFile.Segment(start: 0, end: 1, speaker: "A", text: "one"),
SpeakersFile.Segment(start: 5, end: 6, speaker: "A", text: "two"), // gap 4s > maxGap
])
let out = SpeakerReconciler.mergeAdjacent(f, maxGap: 2.0)
XCTAssertEqual(out.segments.count, 2) // large gap → not merged
// A B A must stay three segments (intervening speaker breaks the run).
let g = file([sp("A", "content"), sp("B", "content")], [
SpeakersFile.Segment(start: 0, end: 1, speaker: "A", text: "a1"),
SpeakersFile.Segment(start: 1.2, end: 2, speaker: "B", text: "b"),
SpeakersFile.Segment(start: 2.2, end: 3, speaker: "A", text: "a2"),
])
XCTAssertEqual(SpeakerReconciler.mergeAdjacent(g, maxGap: 2.0).segments.count, 3)
}
func testParseNamingDropsNullAndKeepsConfidence() {
let json = #"{"speakers":[{"current":"MH","name":"Jonathan Kirkwood","confidence":"high"},{"current":"Unknown_0","name":null,"confidence":"low"}]}"#
let m = SpeakerReconciler.parseNaming(json)