X-Git-Url: http://source.jalview.org/gitweb/?a=blobdiff_plain;f=src%2Fjalview%2Fio%2FIdentifyFile.java;h=ff959b0d79b281f5bb078e500bc768e270931c93;hb=f4766a7bbcfae845fc95923b01fa14ff83d589ff;hp=e90e0ef613f6b75b24c1c449e2e3ca9a6b118bb9;hpb=d7c00f48b00e3ede57c46ae4daf6a9203b9b6059;p=jalview.git

diff --git a/src/jalview/io/IdentifyFile.java b/src/jalview/io/IdentifyFile.java
index e90e0ef..ff959b0 100755
--- a/src/jalview/io/IdentifyFile.java
+++ b/src/jalview/io/IdentifyFile.java
@@ -1,28 +1,30 @@
 /*
- * Jalview - A Sequence Alignment Editor and Viewer (Version 2.7)
- * Copyright (C) 2011 J Procter, AM Waterhouse, J Engelhardt, LM Lui, G Barton, M Clamp, S Searle
- * 
+ * Jalview - A Sequence Alignment Editor and Viewer ($$Version-Rel$$)
+ * Copyright (C) $$Year-Rel$$ The Jalview Authors
+ *
  * This file is part of Jalview.
- * 
+ *
  * Jalview is free software: you can redistribute it and/or
- * modify it under the terms of the GNU General Public License 
- * as published by the Free Software Foundation, either version 3 of the License, or (at your option) any later version.
- * 
- * Jalview is distributed in the hope that it will be useful, but 
- * WITHOUT ANY WARRANTY; without even the implied warranty 
- * of MERCHANTABILITY or FITNESS FOR A PARTICULAR 
+ * modify it under the terms of the GNU General Public License
+ * as published by the Free Software Foundation, either version 3
+ * of the License, or (at your option) any later version.
+ *
+ * Jalview is distributed in the hope that it will be useful, but
+ * WITHOUT ANY WARRANTY; without even the implied warranty
+ * of MERCHANTABILITY or FITNESS FOR A PARTICULAR
  * PURPOSE.  See the GNU General Public License for more details.
- * 
- * You should have received a copy of the GNU General Public License along with Jalview.  If not, see <http://www.gnu.org/licenses/>.
+ *
+ * You should have received a copy of the GNU General Public License
+ * along with Jalview.  If not, see <http://www.gnu.org/licenses/>.
+ * The Jalview Authors are detailed in the 'AUTHORS' file.
  */
 package jalview.io;
 
-import java.io.*;
-import java.net.*;
+import java.io.IOException;
 
 /**
  * DOCUMENT ME!
- * 
+ *
  * @author $author$
  * @version $Revision$
  */
@@ -30,26 +32,26 @@ public class IdentifyFile
 {
   /**
    * Identify a datasource's file content.
-   * 
+   *
    * @note Do not use this method for stream sources - create a FileParse object
    *       instead.
-   * 
+   *
    * @param file
-   *          DOCUMENT ME!
-   * @param protocol
-   *          DOCUMENT ME!
-   * @return ID String
+   * @param sourceType
+   * @return
+   * @throws FileFormatException
    */
-  public String Identify(String file, String protocol)
+  public FileFormatI identify(String file, DataSourceType sourceType)
+          throws FileFormatException
   {
     String emessage = "UNIDENTIFIED FILE PARSING ERROR";
     FileParse parser = null;
     try
     {
-      parser = new FileParse(file, protocol);
+      parser = new FileParse(file, sourceType);
       if (parser.isValid())
       {
-        return Identify(parser);
+        return identify(parser);
       }
     } catch (Exception e)
     {
@@ -58,41 +60,58 @@ public class IdentifyFile
       emessage = e.getMessage();
     }
     if (parser != null)
-      return parser.errormessage;
-    return emessage;
+    {
+      throw new FileFormatException(parser.errormessage);
+    }
+    throw new FileFormatException(emessage);
+  }
+
+  public FileFormatI identify(FileParse source) throws FileFormatException
+  {
+    return identify(source, true);
+    // preserves original behaviour prior to version 2.3
   }
 
-  public String Identify(FileParse source)
+  public FileFormatI identify(AlignmentFileReaderI file,
+          boolean closeSource) throws IOException
   {
-    return Identify(source, true); // preserves original behaviour prior to
-    // version 2.3
+    FileParse fp = new FileParse(file.getInFile(),
+            file.getDataSourceType());
+    return identify(fp, closeSource);
   }
 
   /**
    * Identify contents of source, closing it or resetting source to start
    * afterwards.
-   * 
+   *
    * @param source
    * @param closeSource
-   * @return filetype string
+   * @return (best guess at) file format
+   * @throws FileFormatException
    */
-  public String Identify(FileParse source, boolean closeSource)
+  public FileFormatI identify(FileParse source, boolean closeSource)
+          throws FileFormatException
   {
-    String reply = "PFAM";
+    FileFormatI reply = FileFormat.Pfam;
     String data;
-    int length = 0;
+    int bytesRead = 0;
+    int trimmedLength = 0;
     boolean lineswereskipped = false;
     boolean isBinary = false; // true if length is non-zero and non-printable
     // characters are encountered
+
     try
     {
       if (!closeSource)
       {
         source.mark();
       }
+      boolean aaIndexHeaderRead = false;
+
       while ((data = source.nextLine()) != null)
       {
-        length += data.length();
+        bytesRead += data.length();
+        trimmedLength += data.trim().length();
         if (!lineswereskipped)
         {
           for (int i = 0; !isBinary && i < data.length(); i++)
@@ -115,101 +134,105 @@ public class IdentifyFile
             if (fileStr.lastIndexOf(".jar") > -1
                     || fileStr.lastIndexOf(".zip") > -1)
             {
-              reply = "Jalview";
+              reply = FileFormat.Jalview;
             }
           }
           if (!lineswereskipped && data.startsWith("PK"))
           {
-            reply = "Jalview"; // archive.
+            reply = FileFormat.Jalview; // archive.
             break;
           }
         }
         data = data.toUpperCase();
 
-        if ((data.indexOf("# STOCKHOLM") > -1))
+        if (data.startsWith(ScoreMatrixFile.SCOREMATRIX))
         {
-          reply = "STH";
-
+          reply = FileFormat.ScoreMatrix;
           break;
         }
-        
-        if ((data.indexOf("<") > -1))
+        if (data.startsWith("H ") && !aaIndexHeaderRead)
+        {
+          aaIndexHeaderRead = true;
+        }
+        if (data.startsWith("D ") && aaIndexHeaderRead)
         {
-          reply = "RNAML";
-          
+          reply = FileFormat.ScoreMatrix;
           break;
         }
-
-        if ((data.length() < 1) || (data.indexOf("#") == 0))
+        if (data.startsWith("##GFF-VERSION"))
         {
-          lineswereskipped = true;
-          continue;
+          // GFF - possibly embedded in a Jalview features file!
+          reply = FileFormat.Features;
+          break;
         }
-
-        if (data.indexOf("PILEUP") > -1)
+        if (looksLikeFeatureData(data))
         {
-          reply = "PileUp";
-
+          reply = FileFormat.Features;
           break;
         }
-
-        if ((data.indexOf("//") == 0)
-                || ((data.indexOf("!!") > -1) && (data.indexOf("!!") < data
-                        .indexOf("_MULTIPLE_ALIGNMENT "))))
+        if (data.indexOf("# STOCKHOLM") > -1)
         {
-          reply = "MSF";
-
+          reply = FileFormat.Stockholm;
           break;
         }
-        else if (data.indexOf("CLUSTAL") > -1)
+        if (data.indexOf("_ENTRY.ID") > -1
+                || data.indexOf("_AUDIT_AUTHOR.NAME") > -1
+                || data.indexOf("_ATOM_SITE.") > -1)
         {
-          reply = "CLUSTAL";
-
+          reply = FileFormat.MMCif;
           break;
         }
-        
-        
-        else if (data.indexOf(">") > -1)
+        // if (data.indexOf(">") > -1)
+        if (data.startsWith(">"))
         {
           // FASTA, PIR file or BLC file
-          boolean checkPIR = false,starterm=false;
+          boolean checkPIR = false, starterm = false;
           if ((data.indexOf(">P1;") > -1) || (data.indexOf(">DL;") > -1))
           {
             // watch for PIR file attributes
             checkPIR = true;
-            reply = "PIR";
+            reply = FileFormat.PIR;
           }
           // could also be BLC file, read next line to confirm
           data = source.nextLine();
 
           if (data.indexOf(">") > -1)
           {
-            reply = "BLC";
+            reply = FileFormat.BLC;
           }
           else
           {
             // Is this a single line BLC file?
             String data1 = source.nextLine();
             String data2 = source.nextLine();
+            int c1;
             if (checkPIR)
             {
-              starterm = (data1!=null && data1.indexOf("*")>-1) || (data2!=null && data2.indexOf("*")>-1);
+              starterm = (data1 != null && data1.indexOf("*") > -1)
+                      || (data2 != null && data2.indexOf("*") > -1);
             }
-            if (data2 != null && data.indexOf("*") > -1)
+            if (data2 != null && (c1 = data.indexOf("*")) > -1)
             {
-              if (data.indexOf("*") == data2.indexOf("*"))
+              if (c1 == 0 && c1 == data2.indexOf("*"))
               {
-                reply = "BLC";
+                reply = FileFormat.BLC;
+              }
+              else
+              {
+                reply = FileFormat.Fasta; // possibly a bad choice - may be
+                                          // recognised as
+                // PIR
               }
               // otherwise can still possibly be a PIR file
             }
             else
             {
-              reply = "FASTA";
+              reply = FileFormat.Fasta;
               // TODO : AMSA File is indicated if there is annotation in the
               // FASTA file - but FASTA will automatically generate this at the
               // mo.
-              if (!checkPIR) {
+              if (!checkPIR)
+              {
                 break;
               }
             }
@@ -217,56 +240,108 @@ public class IdentifyFile
           // final check for PIR content. require
           // >P1;title\n<blah>\nterminated sequence to occur at least once.
 
-          // TODO the PIR/fasta ambiguity may be the use case that is needed to have
+          // TODO the PIR/fasta ambiguity may be the use case that is needed to
+          // have
           // a 'Parse as type XXX' parameter for the applet/application.
           if (checkPIR)
           {
-            String dta=null;
+            String dta = null;
             if (!starterm)
             {
-              do {
-                try {
+              do
+              {
+                try
+                {
                   dta = source.nextLine();
-                } catch(IOException ex) {};
-                if (dta!=null && dta.indexOf("*")>-1)
+                } catch (IOException ex)
+                {
+                }
+                if (dta != null && dta.indexOf("*") > -1)
                 {
                   starterm = true;
                 }
-              } while (dta!=null && !starterm);
+              } while (dta != null && !starterm);
             }
             if (starterm)
             {
-              reply="PIR";
+              reply = FileFormat.PIR;
               break;
-            } else {
-              reply="FASTA"; // probably a bad choice!
+            }
+            else
+            {
+              reply = FileFormat.Fasta; // probably a bad choice!
             }
           }
           // read as a FASTA (probably)
           break;
         }
+        if (data.indexOf("{\"") > -1)
+        {
+          reply = FileFormat.Json;
+          break;
+        }
+        int lessThan = data.indexOf("<");
+        if ((lessThan > -1)) // possible Markup Language data i.e HTML,
+                             // RNAML, XML
+        {
+          String upper = data.toUpperCase();
+          if (upper.substring(lessThan).startsWith("<HTML"))
+          {
+            reply = FileFormat.Html;
+            break;
+          }
+          if (upper.substring(lessThan).startsWith("<RNAML"))
+          {
+            reply = FileFormat.Rnaml;
+            break;
+          }
+        }
+
+        if ((data.length() < 1) || (data.indexOf("#") == 0))
+        {
+          lineswereskipped = true;
+          continue;
+        }
+
+        if (data.indexOf("PILEUP") > -1)
+        {
+          reply = FileFormat.Pileup;
+
+          break;
+        }
+
+        if ((data.indexOf("//") == 0) || ((data.indexOf("!!") > -1) && (data
+                .indexOf("!!") < data.indexOf("_MULTIPLE_ALIGNMENT "))))
+        {
+          reply = FileFormat.MSF;
+
+          break;
+        }
+        else if (data.indexOf("CLUSTAL") > -1)
+        {
+          reply = FileFormat.Clustal;
+
+          break;
+        }
+
         else if (data.indexOf("HEADER") == 0 || data.indexOf("ATOM") == 0)
         {
-          reply = "PDB";
+          reply = FileFormat.PDB;
           break;
         }
-        /*
-         * // TODO comment out SimpleBLAST identification for Jalview 2.4.1 else
-         * if (!lineswereskipped && data.indexOf("BLAST")<4) { reply =
-         * "SimpleBLAST"; break;
-         * 
-         * } // end comments for Jalview 2.4.1
-         */
-        else if (!lineswereskipped && data.charAt(0) != '*'
-                && data.charAt(0) != ' '
-                && data.indexOf(":") < data.indexOf(",")) // &&
-        // data.indexOf(",")<data.indexOf(",",
-        // data.indexOf(",")))
+        else if (data.matches("\\s*\\d+\\s+\\d+\\s*"))
         {
-          // file looks like a concise JNet file
-          reply = "JnetFile";
+          reply = FileFormat.Phylip;
           break;
         }
+        else
+        {
+          if (!lineswereskipped && looksLikeJnetData(data))
+          {
+            reply = FileFormat.Jnet;
+            break;
+          }
+        }
 
         lineswereskipped = true; // this means there was some junk before any
         // key file signature
@@ -277,29 +352,86 @@ public class IdentifyFile
       }
       else
       {
-        source.reset(); // so the file can be parsed from the beginning again.
+        source.reset(bytesRead); // so the file can be parsed from the mark
       }
     } catch (Exception ex)
     {
       System.err.println("File Identification failed!\n" + ex);
-      return source.errormessage;
+      throw new FileFormatException(source.errormessage);
     }
-    if (length == 0)
+    if (trimmedLength == 0)
     {
-      System.err
-              .println("File Identification failed! - Empty file was read.");
-      return "EMPTY DATA FILE";
+      System.err.println(
+              "File Identification failed! - Empty file was read.");
+      throw new FileFormatException("EMPTY DATA FILE");
     }
+    System.out.println("File format identified as " + reply.toString());
     return reply;
   }
 
+  /**
+   * Returns true if the data appears to be Jnet concise annotation format
+   * 
+   * @param data
+   * @return
+   */
+  protected boolean looksLikeJnetData(String data)
+  {
+    char firstChar = data.charAt(0);
+    int colonPos = data.indexOf(":");
+    int commaPos = data.indexOf(",");
+    boolean isJnet = firstChar != '*' && firstChar != ' ' && colonPos > -1
+            && commaPos > -1 && colonPos < commaPos;
+    // && data.indexOf(",")<data.indexOf(",", data.indexOf(","))) / ??
+    return isJnet;
+  }
+
+  /**
+   * Returns true if the data has at least 6 tab-delimited fields _and_ fields 4
+   * and 5 are integer (start/end)
+   * 
+   * @param data
+   * @return
+   */
+  protected boolean looksLikeFeatureData(String data)
+  {
+    if (data == null)
+    {
+      return false;
+    }
+    String[] columns = data.split("\t");
+    if (columns.length < 6)
+    {
+      return false;
+    }
+    for (int col = 3; col < 5; col++)
+    {
+      try
+      {
+        Integer.parseInt(columns[col]);
+      } catch (NumberFormatException e)
+      {
+        return false;
+      }
+    }
+    return true;
+  }
+
   public static void main(String[] args)
   {
-	 
     for (int i = 0; args != null && i < args.length; i++)
     {
       IdentifyFile ider = new IdentifyFile();
-      String type = ider.Identify(args[i], AppletFormatAdapter.FILE);
+      FileFormatI type = null;
+      try
+      {
+        type = ider.identify(args[i], DataSourceType.FILE);
+      } catch (FileFormatException e)
+      {
+        System.err.println(
+                String.format("Error '%s' identifying file type for %s",
+                        args[i], e.getMessage()));
+      }
       System.out.println("Type of " + args[i] + " is " + type);
     }
     if (args == null || args.length == 0)