Merge pull request #33178 from vespa-engine/bratseth/to_uri

Add to_uri as prophesied in the documentation
vespa-engine · Jan 25, 2025 · cbaef48 · cbaef48
2 parents 99826c0 + 495a7ef
commit cbaef48
Show file tree

Hide file tree

Showing 5 changed files with 129 additions and 38 deletions.
diff --git a/...glanguage/src/main/java/com/yahoo/vespa/indexinglanguage/expressions/ToUriExpression.java b/...glanguage/src/main/java/com/yahoo/vespa/indexinglanguage/expressions/ToUriExpression.java
@@ -0,0 +1,53 @@
+// Copyright Vespa.ai. Licensed under the terms of the Apache 2.0 license. See LICENSE in the project root.
+package com.yahoo.vespa.indexinglanguage.expressions;
+
+import com.yahoo.document.DataType;
+import com.yahoo.document.datatypes.StringFieldValue;
+import com.yahoo.document.datatypes.UriFieldValue;
+
+/**
+ * @author bratseth
+ */
+public final class ToUriExpression extends Expression {
+
+    @Override
+    public DataType setInputType(DataType input, VerificationContext context) {
+        super.setInputType(input, context);
+        return DataType.URI;
+    }
+
+    @Override
+    public DataType setOutputType(DataType output, VerificationContext context) {
+        super.setOutputType(DataType.URI, output, null, context);
+        return getInputType(context);
+    }
+
+    @Override
+    protected void doVerify(VerificationContext context) {
+        if (context.getCurrentType() == null)
+            throw new VerificationException(this, "Expected input, but no input is provided");
+        context.setCurrentType(createdOutputType());
+    }
+
+    @Override
+    protected void doExecute(ExecutionContext context) {
+        context.setCurrentValue(new UriFieldValue(String.valueOf(context.getCurrentValue())));
+    }
+
+    @Override
+    public DataType createdOutputType() { return DataType.URI; }
+
+    @Override
+    public String toString() { return "to_uri"; }
+
+    @Override
+    public boolean equals(Object obj) {
+        return obj instanceof ToUriExpression;
+    }
+
+    @Override
+    public int hashCode() {
+        return getClass().hashCode();
+    }
+
+}
diff --git a/indexinglanguage/src/main/javacc/IndexingParser.jj b/indexinglanguage/src/main/javacc/IndexingParser.jj
@@ -204,16 +204,17 @@ TOKEN :
     <THIS: "this"> |
     <TOKENIZE: "tokenize"> |
     <TO_ARRAY: "to_array"> |
-    <TO_BYTE: "to_byte"> |
+    <TO_BOOL: "to_bool"> |
+     <TO_BYTE: "to_byte"> |
     <TO_DOUBLE: "to_double"> |
     <TO_FLOAT: "to_float"> |
     <TO_INT: "to_int"> |
     <TO_LONG: "to_long"> |
     <TO_POS: "to_pos"> |
     <TO_EPOCH_SECOND: "to_epoch_second"> |
     <TO_STRING: "to_string"> |
+    <TO_URI: "to_uri"> |
     <TO_WSET: "to_wset"> |
-    <TO_BOOL: "to_bool"> |
     <TRIM: "trim"> |
     <ZCURVE: "zcurve"> |
     <TRUE: "true" > |
@@ -352,16 +353,17 @@ Expression value() :
       val = thisExp()               |
       val = tokenizeExp()           |
       val = toArrayExp()            |
+      val = toBoolExp()             |
       val = toByteExp()             |
       val = toDoubleExp()           |
+      val = toEpochSecondExp()      |
       val = toFloatExp()            |
       val = toIntExp()              |
       val = toLongExp()             |
       val = toPosExp()              |
-      val = toEpochSecondExp()      |
       val = toStringExp()           |
+      val = toUriExp()              |
       val = toWsetExp()             |
-      val = toBoolExp()             |
       val = trimExp()               |
       val = zcurveExp()             |
       ( <LPAREN> val = statement() <RPAREN> { val = new ParenthesisExpression(val); } ) )
@@ -741,6 +743,12 @@ Expression toArrayExp() : { }
     { return new ToArrayExpression(); }
 }
 
+Expression toBoolExp() : { }
+{
+    ( <TO_BOOL> )
+    { return new ToBoolExpression(); }
+}
+
 Expression toByteExp() : { }
 {
     ( <TO_BYTE> )
@@ -753,6 +761,12 @@ Expression toDoubleExp() : { }
     { return new ToDoubleExpression(); }
 }
 
+Expression toEpochSecondExp() : { }
+{
+    ( <TO_EPOCH_SECOND> )
+    { return new ToEpochSecondExpression(); }
+}
+
 Expression toFloatExp() : { }
 {
     ( <TO_FLOAT> )
@@ -777,18 +791,18 @@ Expression toPosExp() : { }
     { return new ToPositionExpression(); }
 }
 
-Expression toEpochSecondExp() : { }
-{
-    ( <TO_EPOCH_SECOND> )
-    { return new ToEpochSecondExpression(); }
-}
-
 Expression toStringExp() : { }
 {
     ( <TO_STRING> )
     { return new ToStringExpression(); }
 }
 
+Expression toUriExp() : { }
+{
+    ( <TO_URI> )
+    { return new ToUriExpression(); }
+}
+
 Expression toWsetExp() :
 {
     boolean createIfNonExistent = false;
@@ -800,12 +814,6 @@ Expression toWsetExp() :
     { return new ToWsetExpression(createIfNonExistent, removeIfZero); }
 }
 
-Expression toBoolExp() : { }
-{
-    ( <TO_BOOL> )
-    { return new ToBoolExpression(); }
-}
-
 Expression trimExp() : { }
 {
     ( <TRIM> )
@@ -894,14 +902,17 @@ String identifier() :
         <SWITCH>                 |
         <THIS>                   |
         <TO_ARRAY>               |
+        <TO_BOOL>                |
+        <TO_BYTE>                |
         <TO_DOUBLE>              |
+        <TO_EPOCH_SECOND>        |
         <TO_FLOAT>               |
         <TO_INT>                 |
         <TO_LONG>                |
         <TO_POS>                 |
         <TO_STRING>              |
+        <TO_URI>                 |
         <TO_WSET>                |
-        <TO_BOOL>                |
         <TOKENIZE>               |
         <TRIM>                   |
         <TRUE>                   |

diff --git a/indexinglanguage/src/test/java/com/yahoo/vespa/indexinglanguage/ScriptTestCase.java b/indexinglanguage/src/test/java/com/yahoo/vespa/indexinglanguage/ScriptTestCase.java
@@ -12,6 +12,7 @@
 import com.yahoo.document.datatypes.IntegerFieldValue;
 import com.yahoo.document.datatypes.LongFieldValue;
 import com.yahoo.document.datatypes.StringFieldValue;
+import com.yahoo.document.datatypes.UriFieldValue;
 import com.yahoo.vespa.indexinglanguage.expressions.AttributeExpression;
 import com.yahoo.vespa.indexinglanguage.expressions.ExecutionContext;
 import com.yahoo.vespa.indexinglanguage.expressions.Expression;
@@ -345,4 +346,21 @@ public void testMultiStatementInput() {
         assertEquals("Test value", ((StringFieldValue)adapter.values.get("myOutputString")).getString());
     }
 
+    @Test
+    public void testToUri() {
+        var tester = new ScriptTester();
+        var expression = tester.expressionFrom("input myString | to_uri | attribute myUri");
+
+        SimpleTestAdapter adapter = new SimpleTestAdapter();
+        var myString = new Field("myString", DataType.STRING);
+        adapter.createField(myString);
+        adapter.setValue("myString", new StringFieldValue("https://vespa.ai"));
+        adapter.createField(new Field("myUri", DataType.URI));
+
+        expression.verify(adapter);
+        ExecutionContext context = new ExecutionContext(adapter);
+        expression.execute(context);
+        assertEquals(new UriFieldValue("https://vespa.ai"), adapter.values.get("myUri"));
+    }
+
 }
diff --git a/...n/schema-language-server/language-server/src/main/ccc/indexinglanguage/IndexingParser.ccc b/...n/schema-language-server/language-server/src/main/ccc/indexinglanguage/IndexingParser.ccc
@@ -229,16 +229,17 @@ TOKEN :
     <THIS: "this"> |
     <TOKENIZE: "tokenize"> |
     <TO_ARRAY: "to_array"> |
+    <TO_BOOL: "to_bool"> |
     <TO_BYTE: "to_byte"> |
     <TO_DOUBLE: "to_double"> |
+    <TO_EPOCH_SECOND: "to_epoch_second"> |
     <TO_FLOAT: "to_float"> |
     <TO_INT: "to_int"> |
     <TO_LONG: "to_long"> |
     <TO_POS: "to_pos"> |
-    <TO_EPOCH_SECOND: "to_epoch_second"> |
     <TO_STRING: "to_string"> |
+    <TO_URI: "to_uri"> |
     <TO_WSET: "to_wset"> |
-    <TO_BOOL: "to_bool"> |
     <TRIM: "trim"> |
     <ZCURVE: "zcurve"> |
     <TRUE: "true" > |
@@ -381,16 +382,17 @@ Expression value() :
       val = thisExp()               |
       val = tokenizeExp()           |
       val = toArrayExp()            |
+      val = toBoolExp()             |
       val = toByteExp()             |
       val = toDoubleExp()           |
+      val = toEpochSecondExp()      |
       val = toFloatExp()            |
       val = toIntExp()              |
       val = toLongExp()             |
       val = toPosExp()              |
-      val = toEpochSecondExp()      |
       val = toStringExp()           |
+      val = toUriExp()              |
       val = toWsetExp()             |
-      val = toBoolExp()             |
       val = trimExp()               |
       val = literalBoolExp()        |
       val = zcurveExp()             |
@@ -785,6 +787,12 @@ Expression toArrayExp() : { }
     { return new ToArrayExpression(); }
 ;
 
+Expression toBoolExp() : { }
+
+    ( <TO_BOOL> )
+    { return new ToBoolExpression(); }
+;
+
 Expression toByteExp() : { }
 
     ( <TO_BYTE> )
@@ -797,6 +805,12 @@ Expression toDoubleExp() : { }
     { return new ToDoubleExpression(); }
 ;
 
+Expression toEpochSecondExp() : { }
+
+    ( <TO_EPOCH_SECOND> )
+    { return new ToEpochSecondExpression(); }
+;
+
 Expression toFloatExp() : { }
 
     ( <TO_FLOAT> )
@@ -821,18 +835,18 @@ Expression toPosExp() : { }
     { return new ToPositionExpression(); }
 ;
 
-Expression toEpochSecondExp() : { }
-
-    ( <TO_EPOCH_SECOND> )
-    { return new ToEpochSecondExpression(); }
-;
-
 Expression toStringExp() : { }
 
     ( <TO_STRING> )
     { return new ToStringExpression(); }
 ;
 
+Expression toUriExp() : { }
+
+    ( <TO_URI> )
+    { return new ToUriExpression(); }
+;
+
 Expression toWsetExp() :
 {
     boolean createIfNonExistent = false;
@@ -844,12 +858,6 @@ Expression toWsetExp() :
     { return new ToWsetExpression(createIfNonExistent, removeIfZero); }
 ;
 
-Expression toBoolExp() : { }
-
-    ( <TO_BOOL> )
-    { return new ToBoolExpression(); }
-;
-
 Expression trimExp() : { }
 
     ( <TRIM> )
@@ -928,14 +936,17 @@ String identifierStr() :
         <SWITCH>                 |
         <THIS>                   |
         <TO_ARRAY>               |
+        <TO_BOOL>                |
+        <TO_BYTE>                |
         <TO_DOUBLE>              |
+        <TO_EPOCH_SECOND>        |
         <TO_FLOAT>               |
         <TO_INT>                 |
         <TO_LONG>                |
         <TO_POS>                 |
         <TO_STRING>              |
+        <TO_URI>                 |
         <TO_WSET>                |
-        <TO_BOOL>                |
         <TOKENIZE>               |
         <TRIM>                   |
         <TRUE>                   |

diff --git a/...ema-language-server/language-server/src/test/java/ai/vespa/schemals/ParserTokensTest.java b/...ema-language-server/language-server/src/test/java/ai/vespa/schemals/ParserTokensTest.java
@@ -34,6 +34,7 @@
  * Tests that the set of tokens declared in JavaCC parsers are also present in CongoCC parsers.
  */
 public class ParserTokensTest {
+
     // These are special
     public static Set<String> javaCCSpecialTokens = Set.of(
         "DEFAULT", // Lexical State in JavaCC parser
@@ -81,7 +82,6 @@ public void testSchemaTokenList() {
         Field[] javaCCFields = SchemaParserConstants.class.getDeclaredFields();
 
         Set<String> congoCCTokenStrings = new HashSet<>();
-
         for (var tokenType : SchemaParserLexer.getRegularTokens()) {
             congoCCTokenStrings.add(tokenType.toString());
         }
@@ -96,7 +96,6 @@ public void testIndexingTokenList() {
         Field[] javaCCFields = IndexingParserConstants.class.getDeclaredFields();
 
         Set<String> congoCCTokenStrings = new HashSet<>();
-
         for (var tokenType : IndexingParserLexer.getRegularTokens()) {
             congoCCTokenStrings.add(tokenType.toString());
         }
@@ -110,7 +109,6 @@ public void testRankingExpressionTokenList() {
         Field[] javaCCFields = RankingExpressionParserConstants.class.getDeclaredFields();
 
         Set<String> congoCCTokenStrings = new HashSet<>();
-
         for (var tokenType : RankingExpressionParserLexer.getRegularTokens()) {
             congoCCTokenStrings.add(tokenType.toString());
         }
@@ -162,6 +160,6 @@ public void testYQLPlusTokenList() {
         }
 
         assertEquals(0, missing.size(), "Missing yqlplus tokens in CongoCC: " + String.join(", ", missing));
-
     }
+
 }