Merge branch 'develop' into features/r2_11_2_alphafold/JAL-2349_JAL-3855
authorJim Procter <j.procter@dundee.ac.uk>
Wed, 16 Feb 2022 13:27:25 +0000 (13:27 +0000)
committerJim Procter <j.procter@dundee.ac.uk>
Wed, 16 Feb 2022 13:27:25 +0000 (13:27 +0000)
combined new logic for 2.11.2 3d-beacons with original hack to resolve PAE and CIF
 Conflicts:
src/jalview/ws/dbsources/EBIAlfaFold.java

1  2 
src/jalview/datamodel/AlignmentAnnotation.java
src/jalview/io/AppletFormatAdapter.java
src/jalview/renderer/AnnotationRenderer.java
src/jalview/ws/SequenceFetcher.java
src/jalview/ws/dbsources/EBIAlfaFold.java
test/jalview/ws/seqfetcher/DbRefFetcherTest.java

Simple merge
  package jalview.ws.dbsources;
  
  import jalview.api.FeatureSettingsModelI;
 +import jalview.bin.Cache;
++import jalview.bin.Console;
  import jalview.datamodel.AlignmentAnnotation;
  import jalview.datamodel.AlignmentI;
 +import jalview.datamodel.ContactMatrix;
 +import jalview.datamodel.ContactMatrixI;
  import jalview.datamodel.DBRefEntry;
  import jalview.datamodel.DBRefSource;
  import jalview.datamodel.PDBEntry;
@@@ -161,8 -156,10 +174,11 @@@ public class EBIAlfaFold extends EbiFil
  
      try
      {
-       File tmpFile = File.createTempFile(id, "cif");
+       File tmpFile = File.createTempFile(id, ".cif");
++      Console.debug("Retrieving structure file for "+id+" from "+alphaFoldCif);
        UrlDownloadClient.download(alphaFoldCif, tmpFile);
+       
+       // may not need this check ?
        file = tmpFile.getAbsolutePath();
        if (file == null)
        {
                  { id, ((chain == null) ? "' '" : chain) }));
        }
  
 +      // import PAE as contact matrix - assume this will work if there was a
 +      // model
 +      File pae = File.createTempFile(id, "pae_json");
 +      String paeURL = getAlphaFoldPaeDownloadUrl(id);
-       UrlDownloadClient.download(paeURL, pae);
++      
++      if (retrievalUrl!=null) {
++        // manufacture the PAE url from a url like ...-model-vN.cif
++        paeURL = retrievalUrl.replace("model","predicted_aligned_error").replace(".cif",".json");
++      }
++      Console.debug("Downloading pae from " + paeURL
++              + " to " + pae.toString() + "");
++
++      try {
++        UrlDownloadClient.download(paeURL, pae);        
 +      if (!importPaeJSONAsContactMatrix(pdbAlignment, pae))
 +      {
-         Cache.log.debug("Couln't import contact matrix from " + paeURL
++        Console.warn("Couln't import contact matrix from " + paeURL
 +                + " (stored in " + pae.toString() + ")");
 +      }
++      } catch (Exception pae_ex) {
++        Console.debug("Couldn't download PAE",pae_ex);
++      }
 +
      } catch (Exception ex) // Problem parsing PDB file
      {
        stopQuery();
      return pdbAlignment;
    }
  
 +  private boolean importPaeJSONAsContactMatrix(AlignmentI pdbAlignment,
 +          File pae) throws Exception
 +  {
 +    FileInputStream pae_input = new FileInputStream(pae);
 +
 +    List<Object> pae_obj = (List<Object>) Platform
 +            .parseJSON(pae_input);
 +    if (pae_obj == null)
 +    {
 +      return false;
 +    }
 +    ContactMatrixI matrix = new PAEContactMatrix(
 +            pdbAlignment.getSequenceAt(0), (Map<String, Object>)pae_obj.get(0));
 +
 +    pdbAlignment.getSequenceAt(0).addAlignmentAnnotation(pdbAlignment.addContactList(matrix));
 +    return true;
 +  }
 +
+   /**
+    * general purpose structure importer - designed to yield alignment useful for transfer of annotation to associated sequences
+    * @param alphaFoldCif
+    * @param tmpFile
+    * @param id
+    * @param chain
+    * @param dbSource
+    * @param dbVersion
+    * @return
+    * @throws Exception
+    */
+   public static AlignmentI importDownloadedStructureFromUrl(String alphaFoldCif,
+           File tmpFile, String id, String chain, String dbSource, String dbVersion) throws Exception
+   {
+     String file = tmpFile.getAbsolutePath();
+     // todo get rid of Type and use FileFormatI instead?
+     FileFormatI fileFormat = FileFormat.MMCif;
+     AlignmentI pdbAlignment = new FormatAdapter().readFile(tmpFile,
+             DataSourceType.FILE, fileFormat);
+     if (pdbAlignment != null)
+     {
+       List<SequenceI> toremove = new ArrayList<SequenceI>();
+       for (SequenceI pdbcs : pdbAlignment.getSequences())
+       {
+         String chid = null;
+         // Mapping map=null;
+         for (PDBEntry pid : pdbcs.getAllPDBEntries())
+         {
+           if (pid.getFile() == file)
+           {
+             chid = pid.getChainCode();
+           }
+         }
+         if (chain == null || (chid != null && (chid.equals(chain)
+                 || chid.trim().equals(chain.trim())
+                 || (chain.trim().length() == 0 && chid.equals("_")))))
+         {
+           // FIXME seems to result in 'PDB|1QIP|1qip|A' - 1QIP is redundant.
+           // TODO: suggest simplify naming to 1qip|A as default name defined
+           pdbcs.setName(id + SEPARATOR + pdbcs.getName());
+           // Might need to add more metadata to the PDBEntry object
+           // like below
+           /*
+            * PDBEntry entry = new PDBEntry(); // Construct the PDBEntry
+            * entry.setId(id); if (entry.getProperty() == null)
+            * entry.setProperty(new Hashtable());
+            * entry.getProperty().put("chains", pdbchain.id + "=" +
+            * sq.getStart() + "-" + sq.getEnd());
+            * sq.getDatasetSequence().addPDBId(entry);
+            */
+           // Add PDB DB Refs
+           // We make a DBRefEtntry because we have obtained the PDB file from
+           // a
+           // verifiable source
+           // JBPNote - PDB DBRefEntry should also carry the chain and mapping
+           // information
+           if (dbSource != null)
+           {
+             DBRefEntry dbentry = new DBRefEntry(dbSource,
+                     dbVersion, (chid == null ? id : id + chid));
+             // dbentry.setMap()
+             pdbcs.addDBRef(dbentry);
+             // update any feature groups
+             List<SequenceFeature> allsf = pdbcs.getFeatures().getAllFeatures();
+             List<SequenceFeature> newsf = new ArrayList<SequenceFeature>();
+             if (allsf!=null && allsf.size()>0)
+             {
+               for (SequenceFeature f:allsf)
+               {
+                 if (file.equals(f.getFeatureGroup()))
+                 {
+                   f = new SequenceFeature(f, f.type, f.begin, f.end, id, f.score);
+                 }
+                 newsf.add(f);
+               }
+               pdbcs.setSequenceFeatures(newsf);
+             }
+           }
+         }
+         else
+         {
+           // mark this sequence to be removed from the alignment
+           // - since it's not from the right chain
+           toremove.add(pdbcs);
+         }
+       }
+       // now remove marked sequences
+       for (SequenceI pdbcs : toremove)
+       {
+         pdbAlignment.deleteSequence(pdbcs);
+         if (pdbcs.getAnnotation() != null)
+         {
+           for (AlignmentAnnotation aa : pdbcs.getAnnotation())
+           {
+             pdbAlignment.deleteAnnotation(aa);
+           }
+         }
+       }
+     }
+     return pdbAlignment;
+   }
    /*
     * (non-Javadoc)
     * 
@@@ -36,8 -35,8 +37,9 @@@ import jalview.datamodel.SequenceFeatur
  import jalview.datamodel.SequenceI;
  import jalview.gui.JvOptionPane;
  import jalview.util.DBRefUtils;
+ import jalview.ws.DBRefFetcher;
  import jalview.ws.SequenceFetcher;
 +import jalview.ws.dbsources.EBIAlfaFold;
  import jalview.ws.dbsources.Pdb;
  import jalview.ws.dbsources.Uniprot;