Support for delta encoding of AP data series for sorted CRAM files

[athos]

This PR adds support for delta encoding of the AP data series for position-sorted CRAM files.

The CRAM specification says that if the AP flag is set in the container's compression header, delta encoding for the AP data series will be enabled. When enabled, the AP data series encodes the difference in the :pos field between each two consecutive records. For the first record in a slice, the difference is calculated relative to the alignment start in the slice header. See the following sections in the CRAM specification for more details:

• 8.4 Compression header block
• 10.2 CRAM positional data

In this implementation, AP delta encoding is enabled if (and only if) the CRAM file is declared as SO:coordinate in the CRAM header. With this change, alignment stats, which were previously calculated during record encoding, must be calculated during preprocessing. The calculated alignment stats are now passed to record encoding via container context/slice contexts.

To cover more comprehensive testing scenarios for AP delta encoding, the existing test resource medium.cram has been re-encoded with AP delta encoding enabled. The re-encoding was done using the following script, which is a modified version of the script previously used:

  (require '[clojure.java.io :as io])
- (import '[htsjdk.samtools CRAMFileWriter SamReaderFactory SAMRecord]
+ (import '[htsjdk.samtools CRAMFileWriter SamReaderFactory SAMRecord SAMFileHeader$SortOrder]
          '[htsjdk.samtools.cram.ref ReferenceSource]
          '[htsjdk.samtools.cram.structure CRAMEncodingStrategy]))

  (let [rdr-factory (SamReaderFactory/makeDefault)
        strategy (doto (CRAMEncodingStrategy.)
                   (.setMinimumSingleReferenceSliceSize 300)
                   (.setReadsPerSlice 500)
                   (.setSlicesPerContainer 10))]
    (with-open [r (.open rdr-factory (io/file "test-resources/bam/medium.bam"))
                os (io/output-stream "test-resources/cram/medium.cram")
                w (CRAMFileWriter. strategy
                                   os
                                   nil
-                                  false
+                                  true
                                   (ReferenceSource. (io/file "hg19.fa"))
-                                  (.getFileHeader r)
+                                  (doto (.clone (.getFileHeader r))
+                                    (.setSortOrder SAMFileHeader$SortOrder/coordinate))
                                   "test-resources/cram/medium.cram")]
      (doseq [^SAMRecord aln (iterator-seq (.iterator r))]
        (.addAlignment w aln)))))

[codecov]

Codecov Report

All modified and coverable lines are covered by tests ✅

Project coverage is 89.80%. Comparing base (b70109a) to head (9209c52).

Additional details and impacted files

@@            Coverage Diff             @@
##           master     #323      +/-   ##
==========================================
+ Coverage   89.78%   89.80%   +0.01%     
==========================================
  Files         102      102              
  Lines        9186     9197      +11     
  Branches      480      480              
==========================================
+ Hits         8248     8259      +11     
  Misses        458      458              
  Partials      480      480              

☔ View full report in Codecov by Sentry.
📢 Have feedback on the report? Share it here.

[niyarin]

Sorry for the late reviewing.
LGTM.

[athos]

Thank you both for reviewing!