Update CRAM conversion and round trip tests.

[cmnbroad]

The existing CRAM compliance tests do only partial record validation after conversion, rather than testing for full record equality. This PR separates out the tests that fail if full record validation is used (tickets filed) and:
-changes the remaining tests to require full SAMRecord equality.
-Adds a BAM->CRAM->BAM round trip validation test based on an NA12878 subset.
-Removes a couple of unreferenced or duplicate alignment and reference files.

Checklist

• Code compiles correctly
• New tests covering changes and new functionality
• All tests passing
• Extended the README / documentation, if necessary
• Is not backward compatible (breaks binary or source compatibility)

[coveralls]

Coverage increased (+0.1%) to 69.151% when pulling 78cc3a2 on cmnbroad:cn_cram_roundtrip into 88b6719 on samtools:master.

[coveralls]

Coverage increased (+0.1%) to 69.151% when pulling e5304a6 on cmnbroad:cn_cram_roundtrip into 88b6719 on samtools:master.

[vadimzalunin]

except for 2 minor changes looks very good in general, binary BAM comparison is my only concern as this depends on system properties.

[vadimzalunin]

not sure about binary equality of BAM files, for example changing compression level will break this.

[vadimzalunin]

remove semicolumn

[vadimzalunin]

remove 'throws IOException'

[vadimzalunin]

summoning @droazen @yfarjoun @cmnbroad into this thread
It appears I'm blocked on this PR... Given the changes I've requested are not important, can we just merge this?

[cmnbroad]

Its fine with me - this PR only has test changes.

[codecov-io]

Codecov Report

Merging #720 into master will increase coverage by 0.114%.

@@               Coverage Diff               @@
##              master      #720       +/-   ##
===============================================
+ Coverage     64.623%   64.737%   +0.114%     
- Complexity      7113      7137       +24     
===============================================
  Files            525       525               
  Lines          31679     31679               
  Branches        5414      5414               
===============================================
+ Hits           20472     20508       +36     
+ Misses          9061      9029       -32     
+ Partials        2146      2142        -4

Impacted Files	Coverage Δ	Complexity Δ
...va/htsjdk/samtools/util/AsyncBufferedIterator.java	68.478% <ø> (-1.087%)	17% <ø> (ø)
src/main/java/htsjdk/samtools/SAMRecord.java	65.931% <ø> (+0.123%)	285% <ø> (+2%)	✅
...ava/htsjdk/samtools/cram/build/CramNormalizer.java	81.356% <ø> (+0.565%)	57% <ø> (+1%)	✅
...main/java/htsjdk/samtools/SAMRecordSetBuilder.java	89.113% <ø> (+0.806%)	64% <ø> (+2%)	✅
...samtools/cram/structure/CramCompressionRecord.java	69.748% <ø> (+0.84%)	77% <ø> (+1%)	✅
...dk/samtools/seekablestream/SeekableHTTPStream.java	56.061% <ø> (+1.515%)	10% <ø> (+1%)	✅
src/main/java/htsjdk/samtools/BinaryTagCodec.java	76.243% <ø> (+17.127%)	68% <ø> (+17%)	✅

---

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 b05d1f2...b54150e. Read the comment docs.

[cmnbroad]

I'm not sure why the codecov report shows a change in src/main/java/htsjdk/samtools/util/AsyncBufferedIterator.java that looks unrelated to this PR, but is looks to be timing related. The tests are passing, so I"m merging.