package jalview.ws2.actions.annotation;

import java.util.ArrayList;
import java.util.BitSet;
import java.util.HashMap;
import java.util.List;
import java.util.Map;

import jalview.analysis.AlignSeq;
import jalview.analysis.SeqsetUtils;
import jalview.datamodel.Sequence;
import jalview.datamodel.SequenceCollectionI;
import jalview.datamodel.SequenceI;
import jalview.schemes.ResidueProperties;
import jalview.util.Comparison;
import jalview.ws2.actions.BaseJob;

public class AnnotationJob extends BaseJob
{
  final boolean[] gapMap;

  final Map<String, SequenceI> seqNames;

  final int regionStart, regionEnd;

  final int minSize;

  public AnnotationJob(List<SequenceI> inputSeqs, boolean[] gapMap,
          Map<String, SequenceI> seqNames, int start, int end, int minSize)
  {
    super(inputSeqs);
    this.gapMap = gapMap;
    this.seqNames = seqNames;
    this.regionStart = start;
    this.regionEnd = end;
    this.minSize = minSize;
  }

  @Override
  public boolean isInputValid()
  {
    int nvalid = 0;
    for (SequenceI sq : getInputSequences())
      if (sq.getStart() <= sq.getEnd())
        nvalid++;
    return nvalid >= minSize;
  }

  public static AnnotationJob create(SequenceCollectionI inputSeqs,
          boolean bySequence, boolean submitGaps, boolean requireAligned,
          boolean filterNonStandardResidues, int minSize)
  {
    List<SequenceI> seqences = new ArrayList<>();
    int minlen = 10;
    int width = 0;
    Map<String, SequenceI> namesMap = bySequence ? new HashMap<>() : null;
    BitSet residueMap = new BitSet();
    int start = inputSeqs.getStartRes();
    int end = inputSeqs.getEndRes();
    // TODO: URGENT! unify with JPred / MSA code to handle hidden regions
    // correctly
    // TODO: push attributes into WsJob instance (so they can be safely
    // persisted/restored
    for (SequenceI sq : inputSeqs.getSequences())
    {
      int sqLen = (bySequence)
              ? sq.findPosition(end + 1) - sq.findPosition(start + 1)
              : sq.getEnd() - sq.getStart();
      if (sqLen < minlen)
        continue;
      String newName = SeqsetUtils.unique_name(seqences.size() + 1);
      if (namesMap != null)
        namesMap.put(newName, sq);
      Sequence seq;
      if (submitGaps)
      {
        seq = new Sequence(newName, sq.getSequenceAsString());
        updateResidueMap(residueMap, seq, filterNonStandardResidues);
      }
      else
      {
        // TODO: add ability to exclude hidden regions
        seq = new Sequence(newName,
                AlignSeq.extractGaps(Comparison.GapChars,
                        sq.getSequenceAsString(start, end + 1)));
        // for annotation need to also record map to sequence start/end
        // position in range
        // then transfer back to original sequence on return.
      }
      seqences.add(seq);
      width = Math.max(width, seq.getLength());
    }

    if (requireAligned && submitGaps)
    {
      for (int i = 0; i < seqences.size(); i++)
      {
        SequenceI sq = seqences.get(i);
        char[] padded = fitSequenceToResidueMap(sq.getSequence(),
                residueMap);
        seqences.set(i, new Sequence(sq.getName(), padded));
      }
    }
    boolean[] gapMapArray = null;
    if (submitGaps)
    {
      gapMapArray = new boolean[width];
      for (int i = 0; i < width; i++)
        gapMapArray[i] = residueMap.get(i);
    }
    return new AnnotationJob(seqences, gapMapArray, namesMap, start, end,
            minSize);
  }

  private static void updateResidueMap(BitSet residueMap, SequenceI seq,
          boolean filterNonStandardResidues)
  {
    for (int pos : seq.gapMap())
    {
      char sqchr = seq.getCharAt(pos);
      boolean include = !filterNonStandardResidues;
      include |= seq.isProtein() ? ResidueProperties.aaIndex[sqchr] < 20
              : ResidueProperties.nucleotideIndex[sqchr] < 5;
      if (include)
        residueMap.set(pos);
    }
  }

  /**
   * Fits the sequence to the residue map removing empty columns where residue
   * map is unset and padding the sequence with gaps at the end if needed.
   */
  private static char[] fitSequenceToResidueMap(char[] sequence,
          BitSet residueMap)
  {
    int width = residueMap.cardinality();
    char[] padded = new char[width];
    for (int op = 0, pp = 0; pp < width; op++)
    {
      if (residueMap.get(op))
      {
        if (sequence.length > op)
          padded[pp++] = sequence[op];
        else
          padded[pp++] = '-';
      }
    }
    return padded;
  }
}