Implement AI orchestration wedding demo

backend/tests/test_soniox_grouping.py

@@ -0,0 +1,17 @@
+from app.infrastructure.soniox import group_by_speaker
+
+
+def test_group_by_speaker_flushes_on_sentence_and_speaker_change():
+    segments = group_by_speaker(
+        [
+            {"speaker": "Bride", "text": "Create ", "start_ms": 0, "end_ms": 100},
+            {"speaker": "Bride", "text": "a task.", "start_ms": 100, "end_ms": 500},
+            {"speaker": "Planner", "text": "Add ", "start_ms": 600, "end_ms": 700},
+            {"speaker": "Planner", "text": "a note.", "start_ms": 700, "end_ms": 1000},
+        ]
+    )
+
+    assert len(segments) == 2
+    assert segments[0].speaker == "Bride"
+    assert segments[0].edited_text == "Create a task."
+    assert segments[1].speaker == "Planner"