Ignore metadata comments in experimental data parsing and improve header dialog sizing

[KWWyatt]

Motivation

• Experimental data files may include comment lines with metadata or prefixed inline comments that should not be interpreted as column labels, and the header-selection dialog needs to behave well on varied screen sizes and long file paths.

Description

• Updated experimental data parsing to detect and ignore comment/metadata lines and to strip comment prefixes before tokenizing by adding _strip_comment_prefix, _is_comment_metadata_line, and _tokens_look_like_column_labels, and by updating _guess_experimental_header_rows, _read_experimental_column_names, and _first_data_tokens to use these helpers.
• Improved robustness of column-name detection by checking whether tokens look like numeric data or plausible column labels before treating a line as a header.
• Enhanced the ExperimentalDataHeaderDialog UI to size itself relative to the primary screen, set sensible minimums and maximums, enable word-wrapping for long file paths, change preview wrapping and minimum height, and configure column QComboBox sizing via a helper _configure_column_combo.
• Updated imports and tests to cover the new behavior and UI changes.

Testing

• Added test_experimental_data_metadata_comments_are_not_used_as_columns, test_experimental_data_ignores_metadata_prefix_before_inline_header, and test_experimental_data_header_dialog_geometry_is_screen_bounded to tests/test_saxs_ui.py and ran them with pytest; they passed.
• Ran the updated tests/test_saxs_ui.py file via pytest tests/test_saxs_ui.py and the suite completed successfully.

---

Codex Task