JAL-3858 report errors when trying to parse JSON to extract PAE and raise a warning...

[jalview.git] / src / jalview / ws / datamodel / alphafold / PAEContactMatrix.java

diff --git a/src/jalview/ws/datamodel/alphafold/PAEContactMatrix.java b/src/jalview/ws/datamodel/alphafold/PAEContactMatrix.java
index 41e677a..22dd7fb 100644 (file)
--- a/src/jalview/ws/datamodel/alphafold/PAEContactMatrix.java
+++ b/src/jalview/ws/datamodel/alphafold/PAEContactMatrix.java
@@ -1,6 +1,10 @@
 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;
@@ -8,6 +12,8 @@ import java.util.Iterator;
 import java.util.List;
 import java.util.Map;
 
+import org.json.simple.JSONObject;
+
 import jalview.analysis.AverageDistanceEngine;
 import jalview.bin.Console;
 import jalview.datamodel.BinaryNode;
@@ -15,8 +21,13 @@ import jalview.datamodel.ContactListI;
 import jalview.datamodel.ContactListImpl;
 import jalview.datamodel.ContactListProviderI;
 import jalview.datamodel.ContactMatrixI;
+import jalview.datamodel.SequenceDummy;
 import jalview.datamodel.SequenceI;
+import jalview.io.DataSourceType;
+import jalview.io.FileFormatException;
+import jalview.io.FileParse;
 import jalview.util.MapUtils;
+import jalview.ws.dbsources.EBIAlfaFold;
 
 public class PAEContactMatrix implements ContactMatrixI
 {
@@ -47,7 +58,7 @@ public class PAEContactMatrix implements ContactMatrixI
   }
 
   @SuppressWarnings("unchecked")
-  public PAEContactMatrix(SequenceI _refSeq, Map<String, Object> pae_obj)
+  public PAEContactMatrix(SequenceI _refSeq, Map<String, Object> pae_obj) throws FileFormatException
   {
     setRefSeq(_refSeq);
     // convert the lists to primitive arrays and store
@@ -101,9 +112,17 @@ public class PAEContactMatrix implements ContactMatrixI
   @SuppressWarnings("unchecked")
   private void parse_version_2_pAE(Map<String, Object> pae_obj)
   {
-    // 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();
+    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()];
@@ -114,10 +133,20 @@ public class PAEContactMatrix implements ContactMatrixI
       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;
@@ -139,10 +168,26 @@ public class PAEContactMatrix implements ContactMatrixI
     // 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();
-    // assume square matrix
-    elements = new float[length][length];
+    elements = new float[maxrow][maxcol];
     while (scores.hasNext())
     {
       float escore = scores.next().floatValue();
@@ -373,4 +418,30 @@ public class PAEContactMatrix implements ContactMatrixI
   {
     return treeType;
   }
+
+  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+"'");
+    }
+  } 
 }