pad the target when aligning

[ekg]

Changes so far:

1. A new parameter target_padding was added to the Parameters struct in align_parameters.hpp. This parameter adds additional padding around target sequences.

2. In computeAlignments.hpp, the main changes are:

   • Modified parseMashmapRow function to accept a new target_padding parameter
   • Added logic to apply padding to reference sequence coordinates (rStartPos and rEndPos) while ensuring they stay within valid bounds (not below 0 or above reference length)
   • Updated function calls to pass through the new target_padding parameter

3. In parse_args.hpp, added command-line support for the target padding feature:

   • Added new command-line option -E or --target-padding to specify padding around target sequence
   • Default value is set to 0
   • Added validation to ensure the padding value is non-negative
   • Integrated the parameter into the alignment parameters structure

Objectives:

• Integrate https://github.com/ekg/indelswizzle/blob/main/cigar_swap.cpp, ideally within the biWFA alignment function in wflign. This will mean left and right-aligning leading/trailing indels, so that we have gaps at the start and end of the alignment. We should then trim back the coordinates of the target.
• Test!
• Only run this when we are on mappings which have been split (via -P 50k for instance) and not at the starts and ends of chains. This would mean integrating a new bit of information into the mapping records to say where we are in the chain in addition to which chain id we have. We could stash that in a single variable or put multiple ones on the row (three are needed—chain id, chain length, position in chain).