Encode the final newline with converted text output

[x0Lazarus]

Converting with --encoding utf-16 currently succeeds but produces a file that cannot be decoded as UTF-16. The CLI encodes the result, then click.echo appends a raw one-byte newline. UTF-32 output has the same problem.

Include the final newline in the text before encoding it, and tell Click not to append another one. This keeps the existing separators and UTF-8 output while producing valid multibyte text, including a single initial BOM where the encoding requires one. The change covers string and JSON results sent to a file or stdout. Binary results and separate-file output are unchanged.

This also touches the dictionary output branch in #102, but leaves its separate -o fix to that PR.

Validation on Windows, Python 3.12:

• All 134 output tests pass with current pySigma and the minimum supported 1.3.0. On unchanged CLI code, 94 fail and 40 controls pass. Tests compare exact bytes for file/stdout, empty values, Unicode, JSON indentation and UTF-16/32 variants.
• The broader offline run has 234 passes and one existing skip. Twelve dataset-fetch tests fail with networking disabled on both the original and patched code; nine online plugin test functions (12 cases) were not run.
• Black 23.3 formats the new tests and changed output calls. The complete convert.py still has pre-existing formatting differences on both revisions; unrelated formatting is left out of this patch.