Fix off-by-one error in parse_range_str page range handling

[yijinlee]

parse_range_str returns 1-based page numbers, but load_pdf_images iterates with a 0-based index. This causes --page-range 1 to return page 2, --page-range 2 to return page 3, etc.

Fix in chandra/input.py, parse_range_str.

Steps to reproduce error:

1. Take a PDF where page 1 and page 2 have distinct content
2. Run chandra input.pdf ./output --page-range 1
3. Output contains page 2's content instead of page 1's