Move seeding-related code into a new "seeding" module

[marcelm]

• Rename IndexParameters to SeedingParameters: This reflects better that the parameters are used both when processing the reference and the query sequences.
• Create a new seeding module and move this code into it:
  • Syncmer generation (including hash computation from hash.rs)
  • Randstrobe generation
  • SeedingParameters (and the seeding parameter profiles, i.e. things related to canonical read length). This makes index.rs a bit shorter and with the rename from IndexParameters to SeedingParameters, it makes it even clearer than before that this code shouldn’t be in that file.
  • randstrobes_query. This reduces the size of mapper.rs a little bit.

I am usually in favor of flat hierarchies, but just having everything be a top module is a bit unwieldy. I think this "seeding" subcategory is a nice, well-contained part the project.

There is now only one project-internal use statement within the seeding module that refers to something outside the module, which is REF_RANDSTROBE_HASH_MASK. I think this is fine because it shows that there is a dependency between the index and how we choose to compute the hash for the randstrobes.