feat: Implement skipCache functionality across conversation handling

src/main/java/de/extio/lmlib/agent/AgentContext.java

@@ -26,6 +26,8 @@ public final class AgentContext {
 	
 	private volatile boolean skipNextCompletion;
 
+	private volatile boolean skipCache;
+	
 	private volatile Conversation conversation;
 	
 	private final AgentRequestStatistic requestStatistic;
@@ -74,6 +76,7 @@ public final class AgentContext {
 		this.agentContextUpdateConsumer = other.agentContextUpdateConsumer;
 		this.error = other.error;
 		this.skipNextCompletion = other.skipNextCompletion;
+		this.skipCache = other.skipCache;
 		this.streaming = other.streaming;
 		synchronized (other.graph) {
 			this.graph.addAll(other.graph);
@@ -202,6 +205,14 @@ public final class AgentContext {
 		this.skipNextCompletion = skipNextCompletion;
 	}
 
+	public boolean isSkipCache() {
+		return this.skipCache;
+	}
+
+	public void setSkipCache(final boolean skipCache) {
+		this.skipCache = skipCache;
+	}
+	
 	Consumer<AgentContext> getAgentContextUpdateConsumer() {
 		return agentContextUpdateConsumer;
 	}

src/main/java/de/extio/lmlib/agent/BaseAgent.java

@@ -103,6 +103,8 @@ public interface BaseAgent {
 					split.context().getGraph().add(this.name());
 					
 					if (!skipCompletion) {
+						final boolean skipCache = split.context().isSkipCache();
+						split.context().setSkipCache(false);
 						final var conversation = this.setupConversation(split);
 						
 						boolean parseable = false;
@@ -126,21 +128,21 @@ public interface BaseAgent {
 								};
 								if (modelProfile != null) {
 									requestStart = System.currentTimeMillis();
-									completion = client.streamConversation(modelProfile, conversation, chunkConsumer);
+									completion = client.streamConversation(modelProfile, conversation, chunkConsumer, skipCache);
 								}
 								else {
 									requestStart = System.currentTimeMillis();
-									completion = client.streamConversation(this.modelCategory(split.context()), conversation, chunkConsumer);
+									completion = client.streamConversation(this.modelCategory(split.context()), conversation, chunkConsumer, skipCache);
 								}
 							}
 							else {
 								if (modelProfile != null) {
 									requestStart = System.currentTimeMillis();
-									completion = client.conversation(modelProfile, conversation);
+									completion = client.conversation(modelProfile, conversation, skipCache);
 								}
 								else {
 									requestStart = System.currentTimeMillis();
-									completion = client.conversation(this.modelCategory(split.context()), conversation);
+									completion = client.conversation(this.modelCategory(split.context()), conversation, skipCache);
 								}
 							}
 							final var durationMs = System.currentTimeMillis() - requestStart;

src/main/java/de/extio/lmlib/client/Client.java

@@ -8,19 +8,13 @@ import de.extio.lmlib.profile.ModelProfile.ModelProvider;
 
 public interface Client {
 	
-	/**
+	Completion conversation(ModelCategory modelCategory, Conversation conversation, boolean skipCache);
-	 * @deprecated use {@link #conversation(ModelCategory, Conversation)} instead. Conversation can be converted from a system and text by Conversation.create(system, text)
-	 */
-	@Deprecated
-	Completion completion(ModelCategory modelCategory, String system, String text);
 	
-	Completion conversation(ModelCategory modelCategory, Conversation conversation);
+	Completion conversation(ModelProfile modelProfile, Conversation conversation, boolean skipCache);
 
-	Completion conversation(ModelProfile modelProfile, Conversation conversation);
+	Completion streamConversation(ModelCategory modelCategory, Conversation conversation, Consumer<Chunk> chunkConsumer, boolean skipCache);
 	
-	Completion streamConversation(ModelCategory modelCategory, Conversation conversation, Consumer<Chunk> chunkConsumer);
+	Completion streamConversation(ModelProfile modelProfile, Conversation conversation, Consumer<Chunk> chunkConsumer, boolean skipCache);
-	
-	Completion streamConversation(ModelProfile modelProfile, Conversation conversation, Consumer<Chunk> chunkConsumer);
 	
 	ModelProvider getModelProvider();
 }

src/main/java/de/extio/lmlib/client/ClientService.java

@@ -57,29 +57,24 @@ public class ClientService {
 		}
 		
 		@Override
-		public Completion completion(final ModelCategory modelCategory, final String system, final String text) {
+		public Completion conversation(final ModelCategory modelCategory, final Conversation conversation_, final boolean skipCache) {
-			return this.client.completion(modelCategory, system, text);
-		}
-		
-		@Override
-		public Completion conversation(final ModelCategory modelCategory, final Conversation conversation_) {
 			final var modelProfile = this.modelProfileService.getModelProfile(modelCategory.getModelProfile(), modelCategory);
-			return this.conversation(modelProfile, conversation_);
+			return this.conversation(modelProfile, conversation_, skipCache);
 		}
 		
 		@Override
-		public Completion streamConversation(final ModelCategory modelCategory, final Conversation conversation_, final Consumer<Chunk> chunkConsumer) {
+		public Completion streamConversation(final ModelCategory modelCategory, final Conversation conversation_, final Consumer<Chunk> chunkConsumer, final boolean skipCache) {
 			final var modelProfile = this.modelProfileService.getModelProfile(modelCategory.getModelProfile(), modelCategory);
-			return this.streamConversation(modelProfile, conversation_, chunkConsumer);
+			return this.streamConversation(modelProfile, conversation_, chunkConsumer, skipCache);
 		}
 
 		@Override
-		public Completion conversation(final ModelProfile modelProfile, final Conversation conversation_) {
+		public Completion conversation(final ModelProfile modelProfile, final Conversation conversation_, final boolean skipCache) {
 			Conversation conversation = conversation_;
 			for (final var interceptor : this.completionInterceptors) {
 				conversation = interceptor.before(modelProfile, conversation);
 			}
-			var completion = this.client.conversation(modelProfile, conversation);
+			var completion = this.client.conversation(modelProfile, conversation, skipCache);
 			for (final var interceptor : this.completionInterceptors) {
 				completion = interceptor.after(modelProfile, conversation, completion);
 			}
@@ -87,12 +82,12 @@ public class ClientService {
 		}
 		
 		@Override
-		public Completion streamConversation(final ModelProfile modelProfile, final Conversation conversation_, final Consumer<Chunk> chunkConsumer) {
+		public Completion streamConversation(final ModelProfile modelProfile, final Conversation conversation_, final Consumer<Chunk> chunkConsumer, final boolean skipCache) {
 			Conversation conversation = conversation_;
 			for (final var interceptor : this.completionInterceptors) {
 				conversation = interceptor.before(modelProfile, conversation);
 			}
-			var completion = this.client.streamConversation(modelProfile, conversation, chunkConsumer);
+			var completion = this.client.streamConversation(modelProfile, conversation, chunkConsumer, skipCache);
 			for (final var interceptor : this.completionInterceptors) {
 				completion = interceptor.after(modelProfile, conversation, completion);
 			}

src/main/java/de/extio/lmlib/client/cached/CachedClient.java

@@ -6,7 +6,6 @@ import java.security.NoSuchAlgorithmException;
 import java.time.Duration;
 import java.time.OffsetDateTime;
 import java.util.HexFormat;
-import java.util.Objects;
 import java.util.function.Consumer;
 import java.util.function.Supplier;
 
@@ -36,23 +35,18 @@ public class CachedClient implements Client {
 	}
 	
 	@Override
-	public Completion completion(final ModelCategory modelCategory, final String system, final String text) {
+	public Completion conversation(final ModelCategory modelCategory, final Conversation conversation, final boolean skipCache) {
-		return this.getCachedResponse(modelCategory, null, () -> this.client.completion(modelCategory, system, text), system, text);
+		return this.getCachedResponse(modelCategory, null, skipCache, () -> this.client.conversation(modelCategory, conversation, skipCache), conversation.getConversation().toString());
 	}
 	
 	@Override
-	public Completion conversation(final ModelCategory modelCategory, final Conversation conversation) {
+	public Completion conversation(final ModelProfile modelProfile, final Conversation conversation, final boolean skipCache) {
-		return this.getCachedResponse(modelCategory, null, () -> this.client.conversation(modelCategory, conversation), conversation.getConversation().toString());
+		return this.getCachedResponse(null, modelProfile, skipCache, () -> this.client.conversation(modelProfile, conversation, skipCache), conversation.getConversation().toString());
 	}
 
 	@Override
-	public Completion conversation(final ModelProfile modelProfile, final Conversation conversation) {
+	public Completion streamConversation(final ModelCategory modelCategory, final Conversation conversation, final Consumer<Chunk> chunkConsumer, final boolean skipCache) {
-		return this.getCachedResponse(null, modelProfile, () -> this.client.conversation(modelProfile, conversation), conversation.getConversation().toString());
+		final var response = this.getCachedResponse(modelCategory, null, skipCache, () -> this.client.streamConversation(modelCategory, conversation, chunkConsumer, skipCache), conversation.getConversation().toString());
-	}
-
-	@Override
-	public Completion streamConversation(final ModelCategory modelCategory, final Conversation conversation, final Consumer<Chunk> chunkConsumer) {
-		final var response = this.getCachedResponse(modelCategory, null, () -> this.client.streamConversation(modelCategory, conversation, chunkConsumer), conversation.getConversation().toString());
 		if (response.statistics().cached()) {
 			chunkConsumer.accept(new Chunk(response.response(), response.reasoning()));
 		}
@@ -60,8 +54,8 @@ public class CachedClient implements Client {
 	}
 	
 	@Override
-	public Completion streamConversation(final ModelProfile modelProfile, final Conversation conversation, final Consumer<Chunk> chunkConsumer) {
+	public Completion streamConversation(final ModelProfile modelProfile, final Conversation conversation, final Consumer<Chunk> chunkConsumer, final boolean skipCache) {
-		final var response = this.getCachedResponse(null, modelProfile, () -> this.client.streamConversation(modelProfile, conversation, chunkConsumer), conversation.getConversation().toString());
+		final var response = this.getCachedResponse(null, modelProfile, skipCache, () -> this.client.streamConversation(modelProfile, conversation, chunkConsumer, skipCache), conversation.getConversation().toString());
 		if (response.statistics().cached()) {
 			chunkConsumer.accept(new Chunk(response.response(), response.reasoning()));
 		}
@@ -73,10 +67,10 @@ public class CachedClient implements Client {
 		return this.client.getModelProvider();
 	}
 	
-	private Completion getCachedResponse(final ModelCategory modelCategory, final ModelProfile modelProfile, final Supplier<Completion> supplier, final String... keys) {
+	private Completion getCachedResponse(final ModelCategory modelCategory, final ModelProfile modelProfile, final boolean skipCache, final Supplier<Completion> supplier, final String... keys) {
 		final var key = this.createCacheKey(modelCategory, modelProfile, keys);
 		
-		final var cachedCompletion = this.cacheRepository.get(key);
+		final var cachedCompletion = skipCache ? null : this.cacheRepository.get(key);
 		if (cachedCompletion != null) {
 			return new Completion(cachedCompletion.response(), cachedCompletion.reasoning(), CompletionFinishReason.DONE, new CompletionStatistics(0, Duration.ZERO, cachedCompletion.inTokens(), cachedCompletion.cachedInTokens(), cachedCompletion.outTokens(), cachedCompletion.reasoningOutTokens(), BigDecimal.ZERO, true));
 		}

src/main/java/de/extio/lmlib/client/oai/completion/AbstractCompletionClient.java

@@ -82,29 +82,24 @@ public abstract class AbstractCompletionClient implements Client, DisposableBean
 	}
 	
 	@Override
-	public Completion completion(final ModelCategory modelCategory, final String system, final String text) {
+	public Completion conversation(final ModelCategory modelCategory, final Conversation conversation, final boolean skipCache) {
-		return this.conversation(modelCategory, Conversation.create(system, text));
+		return this.streamConversation(modelCategory, conversation, null, skipCache);
 	}
 	
 	@Override
-	public Completion conversation(final ModelCategory modelCategory, final Conversation conversation) {
+	public Completion conversation(final ModelProfile modelProfile, final Conversation conversation, final boolean skipCache) {
-		return this.streamConversation(modelCategory, conversation, null);
+		return this.streamConversation(modelProfile, conversation, null, skipCache);
 	}
 	
 	@Override
-	public Completion conversation(final ModelProfile modelProfile, final Conversation conversation) {
+	public Completion streamConversation(final ModelCategory modelCategory_, final Conversation conversation, final Consumer<Chunk> chunkConsumer, final boolean skipCache) {
-		return this.streamConversation(modelProfile, conversation, null);
-	}
-	
-	@Override
-	public Completion streamConversation(final ModelCategory modelCategory_, final Conversation conversation, final Consumer<Chunk> chunkConsumer) {
 		final ModelCategory modelCategory = (modelCategory_ == null) ? ModelCategory.MEDIUM : modelCategory_;
 		final var modelProfile = this.modelProfileService.getModelProfile(modelCategory.getModelProfile(), modelCategory);
-		return this.streamConversation(modelProfile, conversation, chunkConsumer);
+		return this.streamConversation(modelProfile, conversation, chunkConsumer, skipCache);
 	}
 	
 	@Override
-	public Completion streamConversation(final ModelProfile modelProfile, final Conversation conversation, final Consumer<Chunk> chunkConsumer) {
+	public Completion streamConversation(final ModelProfile modelProfile, final Conversation conversation, final Consumer<Chunk> chunkConsumer, final boolean skipCache) {
 		if (modelProfile == null || (modelProfile.modelProvider() != ModelProvider.OAI_TEXT_COMPLETION && modelProfile.modelProvider() != ModelProvider.OAI_CHAT_COMPLETION)) {
 			throw new IllegalArgumentException("Invalid ModelProfile");
 		}

src/main/java/de/extio/lmlib/grader/Grader.java

@@ -5,6 +5,7 @@ import org.slf4j.LoggerFactory;
 
 import de.extio.lmlib.client.Client;
 import de.extio.lmlib.client.ClientService;
+import de.extio.lmlib.client.Conversation;
 import de.extio.lmlib.profile.ModelCategory;
 
 /**
@@ -25,15 +26,17 @@ public class Grader {
 		
 		var score = 0;
 		while (Math.abs(score) < 3) {
-			final var completion = client.completion(
+			final var completion = client.conversation(
 					ModelCategory.MEDIUM,
+					Conversation.create(
 						"""
 								You are a grader assessing the truthfulness of a given text to a user question.
 								Please provide a binary response 'true' or 'false' for the following text.
 								'true' means that the text provides a truthful answer to the question, while 'false' means that it does not.
 								Only provide the response as a single word and no preamble and no explanation.""",
 						"Here is the text:\n" + text + "\n\n" +
-					"Here is the user question: " + question);
+						"Here is the user question: " + question),
+					true);
 			LOGGER.info(completion.response());
 			score += Boolean.parseBoolean(completion.response().strip()) ? 1 : -1;
 			if (Math.abs(score) == 2) {

src/main/java/de/extio/lmlib/grader/Grader2.java

@@ -49,7 +49,8 @@ public class Grader2 {
 					- 'true': The text provides a truthful answer.
 					- 'false': The text does not provide a truthful answer.
 					Only provide the response as a single word. Do not include any preamble or explanation.
-					""".formatted(question, text)));
+					""".formatted(question, text)),
+					true);
 			LOGGER.info(completion.response());
 			score += Boolean.parseBoolean(completion.response().strip()) ? 1 : -1;
 			if (Math.abs(score) == 2) {

src/main/java/de/extio/lmlib/grader/agent/AnswerGrader.java

@@ -31,7 +31,9 @@ public class AnswerGrader {
 						The answer is: {{answer}}
 						""",
 				new JsonAgentResponseHandler(),
-				null,
+				context -> {
+					context.setSkipCache(true);
+				},
 				context -> {
 					if (!context.getContext().containsKey("qaRating")) {
 						context.setStringValue("qaRating", "0");

src/main/java/de/extio/lmlib/grader/agent/AnswerGrader2.java

@@ -53,6 +53,7 @@ public class AnswerGrader2 {
 					}
 					context.getContext().remove("qaRating");
 					context.getContext().remove("qaExplanation");
+					context.setSkipCache(true);
 				},
 				context -> {
 					if (!context.getContext().containsKey("qaRating")) {

src/main/java/de/extio/lmlib/grader/agent/AnswerGraderBinary.java

@@ -44,6 +44,7 @@ public class AnswerGraderBinary {
 					}
 					context.getContext().remove("qaPassed");
 					context.getContext().remove("qaExplanation");
+					context.setSkipCache(true);
 				},
 				context -> {
 					if (!context.getContext().containsKey("qaPassed")) {

src/test/java/de/extio/lmlib/client/cached/CachedClientTest.java

@@ -47,29 +47,29 @@ public class CachedClientTest {
 	
 	@Test
 	void cacheCompletion() throws Exception {
-		var completion = this.clientService.getClient(ModelCategory.SMALL).completion(
+		var completion = this.clientService.getClient(ModelCategory.SMALL).conversation(
 				ModelCategory.SMALL,
-				"You are a funny person",
+				Conversation.create("You are a funny person", "Tell me a cat joke"),
-				"Tell me a cat joke");
+				false);
 		LOGGER.info(completion.toString());
 		assertFalse(completion.statistics().cached());
 		assertEquals(1, completion.statistics().requests());
 		final var firstResponse = completion.response();
 		assertFalse(firstResponse.isBlank());
 		
-		completion = this.clientService.getClient(ModelCategory.SMALL).completion(
+		completion = this.clientService.getClient(ModelCategory.SMALL).conversation(
 				ModelCategory.SMALL,
-				"You are a funny person",
+				Conversation.create("You are a funny person", "Tell me a cat joke"),
-				"Tell me a cat joke");
+				false);
 		LOGGER.info(completion.toString());
 		assertTrue(completion.statistics().cached());
 		assertEquals(0, completion.statistics().requests());
 		assertEquals(firstResponse, completion.response());
 		
-		completion = this.clientService.getClient(ModelCategory.SMALL).completion(
+		completion = this.clientService.getClient(ModelCategory.SMALL).conversation(
 				ModelCategory.SMALL,
-				"You are a funny person",
+				Conversation.create("You are a funny person", "Tell me a dog joke"),
-				"Tell me a dog joke");
+				false);
 		LOGGER.info(completion.toString());
 		assertFalse(completion.statistics().cached());
 		assertEquals(1, completion.statistics().requests());
@@ -80,14 +80,14 @@ public class CachedClientTest {
 	void cacheChat() throws Exception {
 		final var conversation = Conversation.create("You are an assistant who stutters", "Hi there, I would like to know the climate in Munich.");
 		
-		var completion = this.clientService.getClient(ModelCategory.MEDIUM).conversation(ModelCategory.MEDIUM, conversation);
+		var completion = this.clientService.getClient(ModelCategory.MEDIUM).conversation(ModelCategory.MEDIUM, conversation, false);
 		LOGGER.info(completion.toString());
 		assertFalse(completion.statistics().cached());
 		assertEquals(1, completion.statistics().requests());
 		final var firstResponse = completion.response();
 		assertFalse(firstResponse.isBlank());
 		
-		completion = this.clientService.getClient(ModelCategory.MEDIUM).conversation(ModelCategory.MEDIUM, conversation);
+		completion = this.clientService.getClient(ModelCategory.MEDIUM).conversation(ModelCategory.MEDIUM, conversation, false);
 		LOGGER.info(completion.toString());
 		assertTrue(completion.statistics().cached());
 		assertEquals(0, completion.statistics().requests());
@@ -96,13 +96,37 @@ public class CachedClientTest {
 		conversation.addTurn(new Turn(TurnType.ASSISTANT,
 				"T-t-the climate in B-b-berlin, um, it's, uh, quite, c-c-cool, actually. B-b-berlin has a, um, temperate climate, with, uh, four distinct s-s-seasons. S-s-spring is, um, pretty mild, with, uh, temperatures ranging from, uh, 10 to 20 degrees Celsius. S-s-summer is, um, warm, with, uh, highs around 25 degrees Celsius. F-f-fall is, um, cool, with, uh, temperatures ranging from 5 to 15 degrees Celsius. And, um, w-w-winter is, uh, quite, c-c-cold, with, uh, temperatures ranging from -2 to 5 degrees Celsius."));
 		conversation.addTurn(new Turn(TurnType.USER, "Great! How is the weather in summer?"));
-		completion = this.clientService.getClient(ModelCategory.MEDIUM).conversation(ModelCategory.MEDIUM, conversation);
+		completion = this.clientService.getClient(ModelCategory.MEDIUM).conversation(ModelCategory.MEDIUM, conversation, false);
 		LOGGER.info(completion.toString());
 		assertFalse(completion.statistics().cached());
 		assertEquals(1, completion.statistics().requests());
 		assertNotEquals(firstResponse, completion.response());
 	}
 
+	@Test
+	void bypassCacheStillStoresFreshResult() throws Exception {
+		final var conversation = Conversation.create("You are a witty assistant", "Give me a short joke about caches.");
+
+		final var initialCompletion = this.clientService.getClient(ModelCategory.MEDIUM).conversation(ModelCategory.MEDIUM, conversation, false);
+		LOGGER.info(initialCompletion.toString());
+		assertFalse(initialCompletion.statistics().cached());
+		assertEquals(1, initialCompletion.statistics().requests());
+		assertFalse(initialCompletion.response().isBlank());
+
+		final var bypassedCompletion = this.clientService.getClient(ModelCategory.MEDIUM).conversation(ModelCategory.MEDIUM, conversation, true);
+		LOGGER.info(bypassedCompletion.toString());
+		assertFalse(bypassedCompletion.statistics().cached());
+		assertEquals(1, bypassedCompletion.statistics().requests());
+		assertFalse(bypassedCompletion.response().isBlank());
+
+		final var cachedCompletion = this.clientService.getClient(ModelCategory.MEDIUM).conversation(ModelCategory.MEDIUM, conversation, false);
+		LOGGER.info(cachedCompletion.toString());
+		assertTrue(cachedCompletion.statistics().cached());
+		assertEquals(0, cachedCompletion.statistics().requests());
+		assertEquals(bypassedCompletion.response(), cachedCompletion.response());
+		assertEquals(bypassedCompletion.reasoning(), cachedCompletion.reasoning());
+	}
+
 	@Test
 	void cacheStream() throws Exception {
 		final var conversation = Conversation.create("You are an assistant who talks extremely like a pirate", "Hi there, I would like to know the climate in the red sea.");
@@ -116,7 +140,7 @@ public class CachedClientTest {
 				sb.append(chunk.content());
 				System.out.print(chunk.content());
 			}
-		});
+		}, false);
 		assertFalse(completion.statistics().cached());
 		assertEquals(1, completion.statistics().requests());
 		assertTrue(!sb.isEmpty());
@@ -134,7 +158,7 @@ public class CachedClientTest {
 				sb2.append(chunk.content());
 				System.out.print(chunk.content());
 			}
-		});
+		}, false);
 		assertTrue(completion.statistics().cached());
 		assertEquals(sb.toString(), sb2.toString());
 		assertEquals(sb2.toString(), completion.response());

src/test/java/de/extio/lmlib/client/oai/CompletionClientTest.java

@@ -27,6 +27,7 @@ import de.extio.lmlib.client.Conversation.TurnType;
 import de.extio.lmlib.client.oai.completion.chat.ChatCompletionClient;
 import de.extio.lmlib.client.oai.completion.text.TextCompletionClient;
 import de.extio.lmlib.grader.Grader2;
+import de.extio.lmlib.profile.ModelCategory;
 import de.extio.lmlib.profile.ModelProfileService;
 
 @Disabled("This test requires a running Llama server")
@@ -55,7 +56,8 @@ public class CompletionClientTest {
 		final var modelProfile = this.modelProfileService.getModelProfile("profile.model.textcompletion");
 		final var completion = this.textCompletionClient.conversation(
 				modelProfile,
-				Conversation.create("You are a helpful assistant", "Say the color is green"));
+				Conversation.create("You are a helpful assistant", "Say the color is green"),
+				false);
 		
 		LOGGER.info(completion.reasoning());
 		LOGGER.info(completion.response());
@@ -69,7 +71,7 @@ public class CompletionClientTest {
 		final var conversation = Conversation.create("You are a helpful assistant", "I would like to paint my canvas in a solid color.");
 		conversation.addTurn(new Turn(TurnType.ASSISTANT, "Which color do you want to see?"));
 		conversation.addTurn(new Turn(TurnType.USER, "You choose!"));
-		final var completion = this.textCompletionClient.conversation(modelProfile, conversation);
+		final var completion = this.textCompletionClient.conversation(modelProfile, conversation, false);
 		
 		LOGGER.info(completion.reasoning());
 		LOGGER.info(completion.response());
@@ -92,7 +94,7 @@ public class CompletionClientTest {
 				System.out.print(chunk.content());
 				sb.append(chunk.content());
 			}
-		});
+		}, false);
 		
 		if (completion.reasoning() != null) {
 			// Prompt template does not separate reasoning and response from the stream at the moment
@@ -112,7 +114,7 @@ public class CompletionClientTest {
 		final var conversation = Conversation.create("You are a helpful assistant", "I would like to paint my canvas in a solid color.");
 		conversation.addTurn(new Turn(TurnType.ASSISTANT, "Which color do you want to see?"));
 		conversation.addTurn(new Turn(TurnType.USER, "You choose!"));
-		final var completion = this.chatCompletionclient.conversation(modelProfile, conversation);
+		final var completion = this.chatCompletionclient.conversation(modelProfile, conversation, false);
 		
 		LOGGER.info(completion.reasoning());
 		LOGGER.info(completion.response());
@@ -135,7 +137,7 @@ public class CompletionClientTest {
 				System.out.print(chunk.content());
 				sb.append(chunk.content());
 			}
-		});
+		}, false);
 		
 		assertEquals(completion.response(), sb.toString());
 		assertTrue(Grader2.assessScoreBinary("Does the text explain a chess opening?", completion.response(), modelProfile, this.clientService));
@@ -151,10 +153,10 @@ public class CompletionClientTest {
 		}
 		sb.delete(sb.length() - 2, sb.length());
 		
-		final var completion = this.textCompletionClient.completion(
+		final var completion = this.textCompletionClient.conversation(
-				null,
+				ModelCategory.MEDIUM,
-				"You are counting machine",
+				Conversation.create("You are counting machine", "How many numbers do you count?\n" + sb.toString()),
-				"How many numbers do you count?\n" + sb.toString());
+				false);
 		LOGGER.info(completion.response());
 	}
 	
@@ -199,10 +201,10 @@ public class CompletionClientTest {
 	private Tasks createTasks() {
 		return new Tasks(() -> {
 			try {
-				final var completion = this.textCompletionClient.completion(
+				final var completion = this.textCompletionClient.conversation(
-						null,
+						ModelCategory.MEDIUM,
-						"You are a helpful assistant",
+						Conversation.create("You are a helpful assistant", "How do I calculate the annual profit margin?"),
-						"How do I calculate the annual profit margin?");
+						false);
 				LOGGER.info(completion.response());
 			}
 			catch (final Exception e) {
@@ -210,10 +212,10 @@ public class CompletionClientTest {
 			}
 		}, () -> {
 			try {
-				final var completion = this.textCompletionClient.completion(
+				final var completion = this.textCompletionClient.conversation(
-						null,
+						ModelCategory.MEDIUM,
-						"You are a helpful assistant",
+						Conversation.create("You are a helpful assistant", "How do I calculate the operating profit ratio?"),
-						"How do I calculate the operating profit ratio?");
+						false);
 				LOGGER.info(completion.response());
 			}
 			catch (final Exception e) {
@@ -221,10 +223,10 @@ public class CompletionClientTest {
 			}
 		}, () -> {
 			try {
-				final var completion = this.textCompletionClient.completion(
+				final var completion = this.textCompletionClient.conversation(
-						null,
+						ModelCategory.MEDIUM,
-						"You are a helpful assistant",
+						Conversation.create("You are a helpful assistant", "How do I calculate the return of investment?"),
-						"How do I calculate the return of investment?");
+						false);
 				LOGGER.info(completion.response());
 			}
 			catch (final Exception e) {
@@ -232,10 +234,10 @@ public class CompletionClientTest {
 			}
 		}, () -> {
 			try {
-				final var completion = this.textCompletionClient.completion(
+				final var completion = this.textCompletionClient.conversation(
-						null,
+						ModelCategory.MEDIUM,
-						"You are a helpful assistant",
+						Conversation.create("You are a helpful assistant", "How do I calculate the return on net worth?"),
-						"How do I calculate the return on net worth?");
+						false);
 				LOGGER.info(completion.response());
 			}
 			catch (final Exception e) {