Reduce memory and time needed for sample_concordance_plink.

[rajwanir]

Description:

This PR reduces time and memory requirement for rule sample_concordance_plink. It does so by:

1. Implementing a faster _sort_ids sort IDs in read_genome. (789d401).
2. Processes the .genome file in chunks to avoid high memory usage. (a6e5c73).

Testing:

Per current default branch, the sample_concordance_plink has following performance metrics for a sample of 123,244:

attempt	cpus	requested_max_gb	requested_time_hr	execution_time_hr	used_max_gb	status
1	1	57.3	10	10.0	203.3	failed
2	1	86.0	48	48.0	517.6	failed
3	1	114.7	96	54.9	611.2	completed

With this PR the sample_concordance_plink uses 0.9Gb memory and completes in 6.2 minutes.

Related issues:
#358, #241, #312

[jaamarks]

Nice job! This looks like it's going to be a massive improvement in computational efficiency. I've tested this on an N=500 sample with success.

sample_concordance_plink time comparison

previously

start: [Mon Dec  2 13:46:22 2024]
end:   [Mon Dec  2 13:52:52 2024]

Current

start: [Tue Dec  3 15:52:06 2024]
end:   [Tue Dec  3 15:52:07 2024]

[jaamarks]

LGTM