Updated the normalize command to correct common errors in WhisperX-generated input transcripts

README.md

@@ -142,16 +142,16 @@ Global flags:
 - Top-level object with a `segments` array.
 - Bare top-level array of segment objects (for example, Audita-style output).
 
-`normalize` required segment fields:
-
-- `start`
-- `end`
-- `speaker`
-- `text`
-
 `normalize` behavior:
 
-- Validates `start >= 0`, `end >= start`, and non-empty `speaker`.
+- Repairs missing timing fields deterministically:
+  if one of `start`/`end` is present, sets both to that value;
+  if both are missing, uses midpoint of previous `end` and next `start`,
+  with edge fallback to available neighbor and `0.0` for single-segment inputs.
+- If `end < start`, swaps them.
+- Fills missing/empty `speaker` with `Unknown_Speaker`.
+- Drops segments with missing, empty, or whitespace-only `text`.
+- Validates repaired timing with `start >= 0`.
 - Accepts existing input `id` values as provenance only.
 - Reassigns output segment IDs sequentially from `1` to `N`.
 - Sorts deterministically by `(start, end, original_input_index, speaker)`.