package jalview.ws.datamodel.alphafold;
+import java.awt.Color;
+import java.io.BufferedInputStream;
+import java.io.File;
+import java.io.FileInputStream;
+import java.io.IOException;
+import java.util.ArrayList;
+import java.util.BitSet;
+import java.util.HashMap;
import java.util.Iterator;
import java.util.List;
import java.util.Map;
+import java.util.Map.Entry;
+import org.json.simple.JSONObject;
+
+import jalview.analysis.AverageDistanceEngine;
+import jalview.bin.Console;
+import jalview.datamodel.Annotation;
+import jalview.datamodel.BinaryNode;
import jalview.datamodel.ContactListI;
import jalview.datamodel.ContactListImpl;
import jalview.datamodel.ContactListProviderI;
import jalview.datamodel.ContactMatrixI;
+import jalview.datamodel.GroupSet;
+import jalview.datamodel.GroupSetI;
+import jalview.datamodel.Mapping;
+import jalview.datamodel.SequenceDummy;
import jalview.datamodel.SequenceI;
+import jalview.io.DataSourceType;
+import jalview.io.FileFormatException;
+import jalview.io.FileParse;
+import jalview.util.MapList;
+import jalview.util.MapUtils;
+import jalview.ws.dbsources.EBIAlfaFold;
-public class PAEContactMatrix implements ContactMatrixI
+public class PAEContactMatrix extends MappableContactMatrix<PAEContactMatrix> implements ContactMatrixI
{
-
- SequenceI refSeq = null;
-
int maxrow = 0, maxcol = 0;
- int[] indices1, indices2;
-
float[][] elements;
float maxscore;
+
@SuppressWarnings("unchecked")
- public PAEContactMatrix(SequenceI _refSeq, Map<String, Object> pae_obj)
- throws Exception
+ public PAEContactMatrix(SequenceI _refSeq, Map<String, Object> pae_obj) throws FileFormatException
{
- refSeq = _refSeq;
- while (refSeq.getDatasetSequence() != null)
+ setRefSeq(_refSeq);
+ // convert the lists to primitive arrays and store
+
+ if (!MapUtils.containsAKey(pae_obj, "predicted_aligned_error", "pae"))
{
- refSeq = refSeq.getDatasetSequence();
+ parse_version_1_pAE(pae_obj);
+ return;
}
- // convert the lists to primitive arrays and store
- int length = _refSeq.getEnd() - _refSeq.getStart() + 1;
+ else
+ {
+ parse_version_2_pAE(pae_obj);
+ }
+ }
+
+ /**
+ * construct a sequence associated PAE matrix directly from a float array
+ *
+ * @param _refSeq
+ * @param matrix
+ */
+ public PAEContactMatrix(SequenceI _refSeq, float[][] matrix)
+ {
+ setRefSeq(_refSeq);
+ maxcol = 0;
+ for (float[] row : matrix)
+ {
+ if (row.length > maxcol)
+ {
+ maxcol = row.length;
+ }
+ maxscore = row[0];
+ for (float f : row)
+ {
+ if (maxscore < f)
+ {
+ maxscore = f;
+ }
+ }
+ }
+ maxrow = matrix.length;
+ elements = matrix;
+
+ }
+ /**
+ * new matrix with specific mapping to a reference sequence
+ * @param newRefSeq
+ * @param newFromMapList
+ * @param elements2
+ * @param grps2
+ */
+ public PAEContactMatrix(SequenceI newRefSeq,
+ MapList newFromMapList, float[][] elements2, GroupSet grps2)
+ {
+ this(newRefSeq,elements2);
+ toSeq = newFromMapList;
+ grps= grps2;
+ }
+
+ /**
+ * parse a sane JSON representation of the pAE
+ *
+ * @param pae_obj
+ */
+ @SuppressWarnings("unchecked")
+ private void parse_version_2_pAE(Map<String, Object> pae_obj)
+ {
+ maxscore = -1;
+ // look for a maxscore element - if there is one...
+ try
+ {
+ // this is never going to be reached by the integer rounding.. or is it ?
+ maxscore = ((Double) MapUtils.getFirst(pae_obj,
+ "max_predicted_aligned_error", "max_pae")).floatValue();
+ } catch (Throwable t)
+ {
+ // ignore if a key is not found.
+ }
+ List<List<Long>> scoreRows = ((List<List<Long>>) MapUtils
+ .getFirst(pae_obj, "predicted_aligned_error", "pae"));
+ elements = new float[scoreRows.size()][scoreRows.size()];
+ int row = 0, col = 0;
+ for (List<Long> scoreRow : scoreRows)
+ {
+ Iterator<Long> scores = scoreRow.iterator();
+ while (scores.hasNext())
+ {
+ Object d = scores.next();
+
+ if (d instanceof Double)
+ {
+ elements[row][col++] = ((Double) d).longValue();
+ }
+ else
+ {
+ elements[row][col++] = (float) ((Long) d).longValue();
+ }
+
+ if (maxscore < elements[row][col - 1])
+ {
+ maxscore = elements[row][col - 1];
+ }
+ }
+ row++;
+ col = 0;
+ }
+ maxcol = length;
+ maxrow = length;
+ }
+
+ /**
+ * v1 format got ditched 28th July 2022 see
+ * https://alphafold.ebi.ac.uk/faq#:~:text=We%20updated%20the%20PAE%20JSON%20file%20format%20on%2028th%20July%202022
+ *
+ * @param pae_obj
+ */
+ @SuppressWarnings("unchecked")
+ private void parse_version_1_pAE(Map<String, Object> pae_obj)
+ {
// assume indices are with respect to range defined by _refSeq on the
// dataset refSeq
Iterator<Long> rows = ((List<Long>) pae_obj.get("residue1")).iterator();
Iterator<Long> cols = ((List<Long>) pae_obj.get("residue2")).iterator();
+ // two pass - to allocate the elements array
+ while (rows.hasNext())
+ {
+ int row = rows.next().intValue();
+ int col = cols.next().intValue();
+ if (maxrow < row)
+ {
+ maxrow = row;
+ }
+ if (maxcol < col)
+ {
+ maxcol = col;
+ }
+
+ }
+ rows = ((List<Long>) pae_obj.get("residue1")).iterator();
+ cols = ((List<Long>) pae_obj.get("residue2")).iterator();
Iterator<Double> scores = ((List<Double>) pae_obj.get("distance"))
.iterator();
-
- elements = new float[length][length];
+ elements = new float[maxrow][maxcol];
while (scores.hasNext())
{
float escore = scores.next().floatValue();
elements[row - 1][col - 1] = escore;
}
- maxscore = ((Double) pae_obj.get("max_predicted_aligned_error"))
- .floatValue();
+ maxscore = ((Double) MapUtils.getFirst(pae_obj,
+ "max_predicted_aligned_error", "max_pae")).floatValue();
}
@Override
- public ContactListI getContactList(final int _column)
+ public ContactListI getContactList(final int column)
{
- if (_column < 0 || _column >= elements.length)
+// final int _column;
+// if (toSeq != null)
+// {
+// int[] word = toSeq.locateInTo(column, column);
+// if (word == null)
+// {
+// return null;
+// }
+// _column = word[0];
+// }
+// else
+// {
+// _column = column;
+// }
+ if (column < 0 || column >= elements.length)
{
return null;
}
return new ContactListImpl(new ContactListProviderI()
{
@Override
+ public int getPosition()
+ {
+ return column;
+ }
+
+ @Override
public int getContactHeight()
{
return maxcol - 1;
}
@Override
- public double getContactAt(int column)
+ public double getContactAt(int mcolumn)
{
- if (column < 0 || column >= elements[_column].length)
+ if (mcolumn < 0 || mcolumn >= elements[column].length)
{
return -1;
}
- // TODO Auto-generated method stub
- return elements[_column][column];
+ return elements[column][mcolumn];
}
});
}
@Override
+ protected double getElementAt(int _column, int i)
+ {
+ return elements[_column][i];
+ }
+ @Override
public float getMin()
{
return 0;
}
@Override
- public boolean hasReferenceSeq()
+ public String getAnnotDescr()
{
- return (refSeq != null);
+ return "Predicted Alignment Error"+((refSeq==null) ? "" : (" for " + refSeq.getName()));
}
@Override
- public SequenceI getReferenceSeq()
+ public String getAnnotLabel()
{
- return refSeq;
+ StringBuilder label = new StringBuilder("PAE Matrix");
+ //if (this.getReferenceSeq() != null)
+ //{
+ // label.append(":").append(this.getReferenceSeq().getDisplayId(false));
+ //}
+ return label.toString();
}
+ public static final String PAEMATRIX = "PAE_MATRIX";
+
+ @Override
+ public String getType()
+ {
+ return PAEMATRIX;
+ }
+
+ @Override
+ public int getWidth()
+ {
+ return length;
+ }
+
+ @Override
+ public int getHeight()
+ {
+ return length;
+ }
+
+ public static void validateContactMatrixFile(String fileName) throws FileFormatException,IOException
+ {
+ FileInputStream infile=null;
+ try {
+ infile = new FileInputStream(new File(fileName));
+ } catch (Throwable t)
+ {
+ new IOException("Couldn't open "+fileName,t);
+ }
+
+
+ JSONObject paeDict=null;
+ try {
+ paeDict = EBIAlfaFold.parseJSONtoPAEContactMatrix(infile);
+ } catch (Throwable t)
+ {
+ new FileFormatException("Couldn't parse "+fileName+" as a JSON dict or array containing a dict");
+ }
+
+ PAEContactMatrix matrix = new PAEContactMatrix(new SequenceDummy("Predicted"), (Map<String,Object>)paeDict);
+ if (matrix.getWidth()<=0)
+ {
+ throw new FileFormatException("No data in PAE matrix read from '"+fileName+"'");
+ }
+ }
+
+ @Override
+ protected PAEContactMatrix newMappableContactMatrix(
+ SequenceI newRefSeq, MapList newFromMapList)
+ {
+ PAEContactMatrix pae=new PAEContactMatrix(newRefSeq, newFromMapList,
+ elements, new GroupSet(grps));
+ return pae;
+ }
}