Implement long targets pipeline for consistent nhmmer results

[althonos]

Based on tests made after issue #12 I realised the implementation of nhmmer was inconsistent with the binary provided with HMMER3. This PR does the following:

• Add the LongTargetsPipeline to use when searching long nucleotide sequences, using a sliding window over the target instead of loading everything.
• Change nhmmer to use LongTargetsPipeline instead of Pipeline
• Make Pipeline.search_hmm also accept a query Profile or OptimizedProfile
• Refactor Pipeline.search_seq and Pipeline.search_msa to use Pipeline.search_hmm instead of duplicating code.

[codecov]

Codecov Report

Merging #13 (3d79ab4) into master (eb6fe7c) will decrease coverage by 0.25%.
The diff coverage is 72.17%.

@@            Coverage Diff             @@
##           master      #13      +/-   ##
==========================================
- Coverage   80.54%   80.28%   -0.26%     
==========================================
  Files           3        3              
  Lines        4122     4296     +174     
==========================================
+ Hits         3320     3449     +129     
- Misses        802      847      +45     

Flag	Coverage Δ
CPython	80.28% <72.17%> (-0.26%)	⬇️
Linux	80.28% <72.17%> (-0.26%)	⬇️
OSX	80.28% <72.17%> (-0.26%)	⬇️
v3.6	80.23% <72.17%> (-0.26%)	⬇️
v3.7	80.23% <72.17%> (-0.26%)	⬇️
v3.8	80.28% <72.17%> (-0.26%)	⬇️
v3.9	80.28% <72.17%> (-0.26%)	⬇️

Flags with carried forward coverage won't be shown. Click here to find out more.

Impacted Files	Coverage Δ
pyhmmer/plan7.pyx	73.68% <70.63%> (+0.07%)	⬆️
pyhmmer/easel.pyx	85.64% <100.00%> (-0.10%)	⬇️

---

Continue to review full report at Codecov.

Legend - Click here to learn more
Δ = absolute <relative> (impact), ø = not affected, ? = missing data
Powered by Codecov. Last update eb6fe7c...3d79ab4. Read the comment docs.