cueseg: topic chapters from a transcript, and a worked pipeline in the readme

parity/cases/cueseg/sparse-text.ndjson

@@ -0,0 +1,4 @@
+{"start": 0.0, "end": 4.0, "text": "   "}
+{"start": 4.0, "end": 8.0, "text": "the mill stood"}
+{"start": 8.0, "end": 12.0, "text": ""}
+{"start": 3600.0, "end": 3604.0, "text": "parliament dissolved itself again"}