luigidemasi commented on code in PR #27494:
URL: https://github.com/apache/camel/pull/27494#discussion_r4216999298
##########
core/camel-core-languages/src/main/java/org/apache/camel/language/simple/ast/UnaryExpression.java:
##########
@@ -102,10 +103,7 @@ public void init(CamelContext context) {
@Override
public <T> T evaluate(Exchange exchange, Class<T> type) {
- Object value = exp.evaluate(exchange, Object.class);
- // the same rule the language uses for a predicate on its own,
where ${body} is true and a missing
- // header is false, so !${body} and !${header.foo} answer the
opposite of those (CAMEL-24984)
- boolean matches = ObjectHelper.evaluateValuePredicate(value);
+ boolean matches =
ExpressionToPredicateAdapter.toPredicate(exp).matches(exchange);
Review Comment:
Fixed in
[2ad0b3a75251](https://github.com/apache/camel/commit/2ad0b3a75251700121823e954c49f9f975ef170d).
`createNotExpression()` creates the predicate once, outside `evaluate()`, and
retains the CAMEL-24984 explanation. Existing expressions keep value-based
negation, including body and missing-header behavior; only the new
`SimplePredicateExpression` opt-in uses the predicate contract.
`SimpleNotOperatorTest` covers compatibility.
_AI-generated by Codex on behalf of @luigidemasi._
##########
core/camel-core-languages/src/main/java/org/apache/camel/language/simple/ast/SimpleFunctionExpression.java:
##########
@@ -148,9 +149,8 @@ public void init(CamelContext context) {
@Override
public <T> T evaluate(Exchange exchange, Class<T> type) {
- Object value = exp.evaluate(exchange, Object.class);
- // the same rule the language uses for a predicate on its own
(CAMEL-24984)
- boolean matches = ObjectHelper.evaluateValuePredicate(value);
+ // Preserve a function's predicate contract, including
semantic's boolean-only guard.
+ boolean matches =
ExpressionToPredicateAdapter.toPredicate(exp).matches(exchange);
Review Comment:
Fixed in
[2ad0b3a75251](https://github.com/apache/camel/commit/2ad0b3a75251700121823e954c49f9f975ef170d).
`createNegatedExpression()` builds its predicate once and reuses it for
evaluation. The CAMEL-24984 comment and existing value-based semantics are
retained; semantic expressions opt into predicate semantics through
`SimplePredicateExpression`.
_AI-generated by Codex on behalf of @luigidemasi._
##########
components/camel-ai/camel-typesafe-ai/src/main/java/org/apache/camel/component/typesafeai/TypeSafeAiSemanticAdapter.java:
##########
@@ -120,10 +187,47 @@ private SemanticResult result(SemanticQuestion question,
JsonObject answer, Json
if (answer.get("probabilities") instanceof Map<?, ?> values) {
values.forEach((key, value) -> probabilities.put((String) key,
((Number) value).doubleValue()));
}
- return new SemanticResult(
- question.getType() == SemanticQuestion.Type.BOOLEAN ? null :
answer.get(type(question)),
- question.getType() == SemanticQuestion.Type.BOOLEAN ?
answer.getDouble("noul") : null,
+ Double probability = question.getType() ==
SemanticQuestion.Type.BOOLEAN ? answer.getDouble("noul") : null;
+ Object value = switch (question.getType()) {
+ case BOOLEAN -> booleanDecision(question, probability);
Review Comment:
The HTTP client rejects a missing `noul` with `IOException: Invalid TypeSafe
AI response: Expected a number` before this mapping code runs.
[96e6d465f016](https://github.com/apache/camel/commit/96e6d465f016856fb43c7e95b8e67818cda8112d)
adds adapter-level HTTP regressions for missing `noul`, the named answer and
`answers`. No duplicate null guard is needed here. Resolving following your
confirmation in [your latest
review](https://github.com/apache/camel/pull/27494#pullrequestreview-5453768092).
_AI-generated by Codex on behalf of @luigidemasi._
##########
components/camel-ai/camel-semantic/src/main/java/org/apache/camel/semantic/SemanticQuestion.java:
##########
@@ -16,118 +16,178 @@
*/
package org.apache.camel.semantic;
+import java.math.BigDecimal;
+import java.math.BigInteger;
+import java.util.ArrayList;
import java.util.Collections;
import java.util.LinkedHashMap;
import java.util.List;
+import java.util.Locale;
import java.util.Map;
-import java.util.Objects;
+import java.util.Set;
-/**
- * Immutable question definition. Probabilities and thresholds are
provider-specific evidence, not accuracy guarantees.
- */
+/** Immutable evaluation declaration. The expert defines the operation,
parameter vocabulary and result semantics. */
public final class SemanticQuestion {
+ private static final Set<Class<?>> NUMBER_TYPES = Set.of(Byte.class,
Short.class, Integer.class, Long.class,
+ Float.class, Double.class, BigInteger.class, BigDecimal.class);
+
public enum Type {
Review Comment:
Addressed in
[2ad0b3a75251](https://github.com/apache/camel/commit/2ad0b3a75251700121823e954c49f9f975ef170d).
The generic declaration is now `SemanticEvaluation`, containing only
operation, expert, state and immutable parameters. The generic
TypeSafe-specific enum, convenience constructors and policy getters were
removed; operation mapping, policy defaults and interpretation live in
`TypeSafeAiSemanticAdapter`.
_AI-generated by Codex on behalf of @luigidemasi._
--
This is an automated message from the Apache Git Service.
To respond to the message, please log on to GitHub and use the
URL above to go to the specific comment.
To unsubscribe, e-mail: [email protected]
For queries about this service, please contact Infrastructure at:
[email protected]
