Reimplement stitching algorithm

Singularity

@@ -188,11 +188,12 @@ From: centos:7
 %applabels denovo
     KIVE_INPUTS sample_info_csv fastq1 fastq2 bad_cycles_csv
     KIVE_OUTPUTS g2p_csv g2p_summary_csv remap_counts_csv \
-        remap_conseq_csv unmapped1_fastq unmapped2_fastq conseq_ins_csv \
+        remap_conseq_csv remap_unstitched_conseq_csv unmapped1_fastq unmapped2_fastq conseq_ins_csv \
         failed_csv cascade_csv nuc_csv amino_csv insertions_csv conseq_csv \
         conseq_all_csv concordance_csv concordance_seed_csv failed_align_csv \
         coverage_scores_csv coverage_maps_tar aligned_csv g2p_aligned_csv \
-        genome_coverage_csv genome_coverage_svg genome_concordance_svg contigs_csv \
+        genome_coverage_csv genome_coverage_svg genome_concordance_svg \
+        contigs_unstitched_csv contigs_csv \
         read_entropy_csv conseq_region_csv conseq_stitched_csv
     KIVE_THREADS 2
     KIVE_MEMORY 6000

docs/steps.md

@@ -44,8 +44,9 @@ Individual files are described after the list of steps.
   * in - fastq1
   * in - fastq2
   * in - merged_contigs.csv
-  * contigs.csv - the assembled contigs, plus any merged contigs, including
+  * contigs_unstitched.csv - the assembled contigs, plus any merged contigs, including
     the best blast results
+  * contigs.csv - stitched version of `contigs_unstitched`
   * blast.csv - multiple blast results for each contig
 * `remap`: iteratively use consensus from previous mapping as reference to try
     and map more reads. See [remap design] for more details. (The denovo version
@@ -58,6 +59,8 @@ Individual files are described after the list of steps.
     each stage.
   * remap_conseq.csv - downloaded - consensus sequence that reads were mapped to
     on the final iteration
+  * remap_unstitched_conseq.csv - downloaded - consensus sequence that reads were
+    mapped to the unstitched contigs.
   * unmapped1.fastq - FASTQ format (unstructured text) reads that didn't map to
     any of the final references.
   * unmapped2.fastq - FASTQ 
@@ -215,11 +218,15 @@ Individual files are described after the list of steps.
   * pos - 1-based position in the consensus sequence that this insertion follows
   * insert - the nucleotide sequence that was inserted
   * qual - the Phred quality scores for the inserted sequence
-* contigs.csv
-  * genotype - the reference name with the best BLAST result
+* contigs_unstitched.csv
+  * ref - the reference name with the best BLAST result
   * match - the fraction of the contig that matched in BLAST, negative for
     reverse-complemented matches
+  * group_ref - the reference name chosen to best match all of
+    the contigs in a sample
   * contig - the nucleotide sequence of the assembled contig
+* contigs.csv
+  Same as `contigs_unstitched.csv`, but contigs are stitched by `micall/core/contig_stitcher.py`.
 * coverage_scores.csv
   * project - the project this score is defined by
   * region - the region being displayed
@@ -343,6 +350,9 @@ Individual files are described after the list of steps.
 * remap_conseq.csv
   * region - the region mapped to
   * sequence - the consensus sequence used
+* remap_unstitched_conseq.csv
+  * region - the region mapped to
+  * sequence - the consensus sequence used
 * resistance.csv
   * region - the region code, like PR or RT
   * drug_class - the drug class code from the HIVdb rules, like NRTI