feat(cli): separate-branches rename, short aliases, clean dry-run

- --normalize-to-one renamed to --separate-branches (no alias; the
  name now describes the output choice: total spectrum with
  intensity-normalized branches vs separate branches, each
  normalized to unit area with intensities left to the data user).
  SpectrumConfig.normalize_to_one -> separate_branches, CSV header
  word to_one -> separate.
- every long option gains a single-letter alias: -N -i -s -m -o -p
  -L -d -c -V -S -b (only -v/-q existed before)
- dry-run drops the top-level Transition: line — the per-branch
  list below already reports each branch's transition type
- version bumped to 0.4.1 (pyproject, __version__, CSV semantics
  docstring, TODO roadmap entry)

tests/quality/test_logging_utils.py

@@ -154,7 +154,7 @@ class TestCSVMetadataHeader:
             lines = f.readlines()
 
         assert any("# beta-spectrum v" in line for line in lines)
-        assert any("0.4.0" in line for line in lines)
+        assert any("0.4.1" in line for line in lines)
         os.unlink(filename)
 
     def test_header_contains_timestamp(self):