git-svn-id: http://www.xerial.org/svn/project/XerialJ/trunk/xerial-core@3403 ae02f08e...

diff --git a/src/main/java/org/xerial/silk/schema/impl/SilkSchema.g b/src/main/java/org/xerial/silk/schema/impl/SilkSchema.g
index 2fe43b3..dda0a9a 100644 (file)
--- a/src/main/java/org/xerial/silk/schema/impl/SilkSchema.g
+++ b/src/main/java/org/xerial/silk/schema/impl/SilkSchema.g
@@ -95,13 +95,21 @@ package org.xerial.silk.schema.impl;
 }\r
  \r
 \r
-// lexer rules \r
+// lexer rules\r
 \r
+\r
+// Line Comment\r
+fragment LineBreakChar: '\n' | '\r'; // r: <CR> n : <LF>\r
+LineComment: '#' ~(LineBreakChar)* { $channel=HIDDEN; };\r
+\r
+\r
+LineBreak: ('\r' '\n' | '\r' | '\n' ); \r
 WhiteSpace: (' ' | '\r' | '\t' | '\u000C' | '\n') { $channel=HIDDEN; };\r
 \r
 fragment Digit: '0' .. '9';\r
 fragment Letter: 'A' .. 'F' | 'a' .. 'f';\r
 fragment HexDigit: Digit | Letter;\r
+fragment NonWhiteUnicodeChar: ~('"' | '\\' | \r
 fragment UnicodeChar: ~('"'| '\\');\r
 fragment EscapeSequence\r
   : '\\' ('\"' | '\\' | '/' | 'b' | 'f' | 'n' | 'r' | 't' | 'u' HexDigit HexDigit HexDigit HexDigit)\r
@@ -113,7 +121,6 @@ fragment StringChar_s: StringChar*;
 String: '"' s=StringChar_s '"' { setText($s.text); };\r
 \r
 \r
-Colon: ':';\r
 Comma: ',';\r
 Integer: '-'? ('0' | '1'..'9' Digit*);\r
 fragment Frac: '.' Digit+;\r
@@ -125,18 +132,55 @@ RBrace: '}' ;
 LBracket: '[' ;\r
 RBracket: ']' ;\r
 \r
+Lt: '<';\r
+Eq: '=';\r
+\r
 LParen: '(';\r
 RParen: ')';\r
 \r
+Star: '*';\r
+\r
+fragment\r
+UnsafeUnicodeChar: '(' | ')' | [' | ']' | '{' | '}' | ',' | ':' | '#' | '<' | '>' | '|' | '*' | '\'' | '"' | '@' | '%' | '\\'; \r
+\r
+fragment:\r
+NonWhiteSpaceChar: ~(UnsafeUncodeChar | WhiteSpace);\r
+\r
+\r
+Symbol: { ':' NonWhiteSpaceChar} =>  ':' NonWhiteSpaceChar+ ;\r
+\r
 Class: 'class';\r
+Includes: 'includes';\r
+End: 'end';\r
+\r
+fragment SafeFirstLetter: 'A' .. 'Z' | 'a' .. 'z';\r
+fragment SafeLetter: SafeFirstLetter | '0' .. '9' | '-' | '_';\r
+\r
+ModuleDef: 'module' WhiteSpace* SafeFistLetter SafeLetter* ('.' SafeFirstLetter SafeLetter)*; \r
+\r
+QName: ~(UnsafeUnicodeChar | WhiteSpace);\r
+\r
 \r
 // parser rules \r
 \r
 schema\r
-  : (classDefinition)*\r
+  : (classDefinition | moduleDefinition )*\r
   ;  \r
   \r
+moduleDefinition:\r
+       ModuleDef LineBreak \r
+       ;  \r
+  \r
 classDefinition \r
-  : Class LBrace RBrace\r
-  ;\r
+  :  Class LineBreak\r
+      ((includeStatment | attributes) LineBreak)* \r
+     End LineBreak\r
+  ; \r
+  \r
+includeStatement: Includes QName (',' QName);   \r
+       \r
+\r
+\r
+\r
+\r
 \r

diff --git a/src/test/java/org/xerial/silk/RelationExtractTest.java b/src/test/java/org/xerial/silk/RelationExtractTest.java
new file mode 100644 (file)
index 0000000..a217cd6
--- /dev/null
+++ b/src/test/java/org/xerial/silk/RelationExtractTest.java
@@ -0,0 +1,97 @@
+/*--------------------------------------------------------------------------\r
+ *  Copyright 2009 Taro L. Saito\r
+ *\r
+ *  Licensed under the Apache License, Version 2.0 (the "License");\r
+ *  you may not use this file except in compliance with the License.\r
+ *  You may obtain a copy of the License at\r
+ *\r
+ *     http://www.apache.org/licenses/LICENSE-2.0\r
+ *\r
+ *  Unless required by applicable law or agreed to in writing, software\r
+ *  distributed under the License is distributed on an "AS IS" BASIS,\r
+ *  WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.\r
+ *  See the License for the specific language governing permissions and\r
+ *  limitations under the License.\r
+ *--------------------------------------------------------------------------*/\r
+//--------------------------------------\r
+// XerialJ\r
+//\r
+// RelationExtractTest.java\r
+// Since: Jul 2, 2009 2:38:03 PM\r
+//\r
+// $URL$\r
+// $Author$\r
+//--------------------------------------\r
+package org.xerial.silk;\r
+\r
+import java.util.ArrayList;\r
+import java.util.List;\r
+\r
+import org.junit.After;\r
+import org.junit.Before;\r
+import org.junit.Test;\r
+import org.xerial.lens.Lens;\r
+import org.xerial.lens.ObjectLens;\r
+import org.xerial.util.FileResource;\r
+import org.xerial.util.log.Logger;\r
+\r
+public class RelationExtractTest\r
+{\r
+    private static Logger _logger = Logger.getLogger(RelationExtractTest.class);\r
+\r
+    @Before\r
+    public void setUp() throws Exception\r
+    {}\r
+\r
+    @After\r
+    public void tearDown() throws Exception\r
+    {}\r
+\r
+    public static class Coordinate\r
+    {\r
+        public String name;\r
+        public String species;\r
+\r
+    }\r
+\r
+    public static class GeneList\r
+    {\r
+        private List<Gene> geneList = new ArrayList<Gene>();\r
+\r
+        public void add(Coordinate coordinate, Gene gene)\r
+        {\r
+            _logger.info(String\r
+                    .format("corrdinate %s, gene %s", ObjectLens.toJSON(coordinate), ObjectLens.toJSON(gene)));\r
+        }\r
+\r
+        public void add(Gene g)\r
+        {\r
+            geneList.add(g);\r
+        }\r
+\r
+        public List<Gene> getGeneList()\r
+        {\r
+            return geneList;\r
+        }\r
+\r
+    }\r
+\r
+    public static class Gene\r
+    {\r
+        public String name;\r
+        public long start;\r
+        public String strand;\r
+        public String sequence;\r
+\r
+    }\r
+\r
+    @Test\r
+    public void test() throws Exception\r
+    {\r
+        GeneList geneList = Lens\r
+                .loadSilk(GeneList.class, FileResource.open(RelationExtractTest.class, "sequence.silk"));\r
+\r
+        _logger.info(ObjectLens.toJSON(geneList));\r
+    }\r
+\r
+}\r

diff --git a/src/test/java/org/xerial/silk/schema/schema.silk b/src/test/java/org/xerial/silk/schema/schema.silk
index 2d4dc14..0d83997 100644 (file)
--- a/src/test/java/org/xerial/silk/schema/schema.silk
+++ b/src/test/java/org/xerial/silk/schema/schema.silk
@@ -1,34 +1,45 @@
-module org.utgenome.\r
+%silk(version:1.0, type:schema)\r
 \r
+module org.utgenome\r
+\r
+# coordinate system informatio\r
 class Coordinate\r
-  group\r
-  species\r
-  revision\r
-  name\r
+  string :group = "utgb"\r
+  :species \r
+  :revision\r
+  :name\r
 end\r
 \r
+# region on a genome sequence\r
 class Locus\r
-  start: integer\r
-  end: integer \r
+  integer :start \r
+  integer :end \r
 end\r
 \r
+# gene data with sequence data\r
 class Gene < Locus\r
-  id: integer \r
-  name\r
+  integer :id \r
+  sequence :sequence \r
 end\r
 \r
-\r
-class Reference\r
-  mix in Locus  \r
-  sequence: string\r
-  read*: Read\r
+# reference sequence\r
+class Reference < Locus\r
+  Coordinate :coodinate\r
+  sequence :sequence \r
 end\r
 \r
+# short-read data\r
 class Read \r
-  include Locus\r
-  sequence: string\r
-  qv*: integer\r
+  includes Locus\r
+  string :sequence \r
+  integer* :qv \r
 end \r
 \r
+# Alignment of reads to a reference genome\r
+class Alignment\r
+  Reference :reference\r
+  Read* :read \r
+end\r
 \r
 \r
+end \r

diff --git a/src/test/java/org/xerial/silk/sequence.silk b/src/test/java/org/xerial/silk/sequence.silk
index 336bc29..5e8e03e 100644 (file)
--- a/src/test/java/org/xerial/silk/sequence.silk
+++ b/src/test/java/org/xerial/silk/sequence.silk
@@ -5,6 +5,9 @@
 TAG1   100     +       ACCCGGTTTTGGCGCTTTCCTTTC[----]TTGGCCTTGGGGCCCATCB       {"link":"http://somewhere.org/", "description":"reference sequence"}\r
 TAG2   100     -       ACCCGG---GGCGCTTTCCCTTTC --AC TTGGCCTTGGGGCCCATCG       \r
 \r
+\r
+\r
+\r
 -(group:utgb, type:chromosome, species:human, revision:hg18)\r
  -coordinate(name:chr2)\r
   -gene(name, start, strand, sequence, param[json])|\r