Revert "Use consensus for breaking conflicts in stitching"

This reverts commit 2d27096.

micall/core/aln2counts.py

@@ -173,74 +173,23 @@ def get_insertion_info(left, report_aminos, report_nucleotides):
     return insert_behind, insertion_coverage
 
 
-
-def build_consensus_table(nuc_dict, region_nucleotides, region_start):
-    """Builds a consensus table with pairs of existing and new nucleotides that share a consensus index."""
-
-    consensus_nucleotides = defaultdict(lambda: {'left': None, 'right': None})
-
-    for nuc_index, nuclotide in enumerate(region_nucleotides):
-        position = region_start + nuc_index - 1
-        existing_nucleotide = nuc_dict.get(position, None)
-        new_nucleotide = nuclotide.seed_nucleotide
-
-        if existing_nucleotide and existing_nucleotide.consensus_index is not None:
-            consensus_nucleotides[existing_nucleotide.consensus_index]['left'] = existing_nucleotide
-        if new_nucleotide.consensus_index is not None:
-            consensus_nucleotides[new_nucleotide.consensus_index]['right'] = new_nucleotide
-
-    return consensus_nucleotides
-
-
-def resolve_nucleotide_conflict(left, right, nuc_index, overlap_midpoint, is_amino):
-    """Resolves a conflict between two nucleotides based on the regions overlap midpoint."""
-
-    if left and right:
-        # we prefer the next region after the middle of the overlap with the previous translated region for an amino nucleotide
-        if nuc_index >= overlap_midpoint and is_amino:
-            return right
-        else:
-            return left
-    elif left:
-        return left
-    elif right:
-        return right
-
-
 def combine_region_nucleotides(nuc_dict, region_nucleotides, region_start, prev_region_end, is_amino, region_name):
     assert region_start is not None
     mismatch = False
     overlap_midpoint = int((prev_region_end - region_start + 1) / 2)
     if overlap_midpoint < 0:
         overlap_midpoint = 0
-
-    consensus_nucleotides = build_consensus_table(nuc_dict, region_nucleotides, region_start)
-    previous_consensus_index = None
-
     for nuc_index, nucleotide in enumerate(region_nucleotides):
         position = region_start + nuc_index - 1
-        existing_nucleotide = nuc_dict.get(position, None)
-
-        def consensus_nucleotide():
-            if previous_consensus_index is not None:
-                left = consensus_nucleotides[previous_consensus_index + 1]['left']
-                right = consensus_nucleotides[previous_consensus_index + 1]['right']
-                return resolve_nucleotide_conflict(left, right,
-                                                   nuc_index, overlap_midpoint,
-                                                   is_amino)
-
-        def region_nucleotide():
-            return resolve_nucleotide_conflict(existing_nucleotide,
-                                               nucleotide.seed_nucleotide,
-                                               nuc_index, overlap_midpoint,
-                                               is_amino)
-
-        nuc_dict[position] = consensus_nucleotide() or region_nucleotide()
-        if existing_nucleotide and existing_nucleotide.counts != nuc_dict[position].counts:
-            mismatch = True
-
-        previous_consensus_index = nuc_dict[position].consensus_index
-
+        if position not in nuc_dict:
+            # if we have not seen this position before, add it
+            nuc_dict[position] = nucleotide.seed_nucleotide
+        else:
+            if is_amino and nuc_index >= overlap_midpoint:
+                # after the middle of the overlap with the previous translated region, we use this region
+                nuc_dict[position] = nucleotide.seed_nucleotide
+            if nuc_dict[position].counts != nucleotide.seed_nucleotide.counts:
+                mismatch = True
     if mismatch:
         logger.debug(f"Disagreement in counts for region {region_name}")