[maven-release-plugin] prepare release v1.0.56

Adds streaming feature to generate API

docs/docs/apis-extras/_category_.json

@@ -1,6 +1,6 @@
 {
   "label": "APIs - Extras",
-  "position": 10,
+  "position": 4,
   "link": {
     "type": "generated-index",
     "description": "Details of APIs to handle bunch of extra stuff."

docs/docs/apis-generate/_category_.json

@@ -1,6 +1,6 @@
 {
-  "label": "APIs - Ask",
-  "position": 10,
+  "label": "APIs - Generate",
+  "position": 3,
   "link": {
     "type": "generated-index",
     "description": "Details of APIs to interact with LLMs."

docs/docs/apis-generate/generate.md

@@ -41,6 +41,41 @@ You will get a response similar to:
 > require
 > natural language understanding and generation capabilities.
 
+## Try asking a question, receiving the answer streamed
+
+```java
+public class Main {
+
+    public static void main(String[] args) {
+
+        String host = "http://localhost:11434/";
+
+        OllamaAPI ollamaAPI = new OllamaAPI(host);
+        // define a stream handler (Consumer<String>)
+        OllamaStreamHandler streamHandler = (s) -> {
+           System.out.println(s);
+        };
+
+        // Should be called using seperate thread to gain non blocking streaming effect.
+        OllamaResult result = ollamaAPI.generate(config.getModel(),
+          "What is the capital of France? And what's France's connection with Mona Lisa?",
+          new OptionsBuilder().build(), streamHandler);
+        
+        System.out.println("Full response: " +result.getResponse());
+    }
+}
+```
+You will get a response similar to:
+
+> The
+> The capital
+> The capital of
+> The capital of France
+> The capital of France is 
+> The capital of France is Paris
+> The capital of France is Paris.
+> Full response: The capital of France is Paris.
+
 ## Try asking a question from general topics.
 
 ```java

docs/docs/apis-model-management/_category_.json

@@ -1,6 +1,6 @@
 {
   "label": "APIs - Model Management",
-  "position": 4,
+  "position": 2,
   "link": {
     "type": "generated-index",
     "description": "Details of APIs to manage LLMs."

pom.xml

@@ -4,7 +4,7 @@
 
     <groupId>io.github.amithkoujalgi</groupId>
     <artifactId>ollama4j</artifactId>
-    <version>1.0.54</version>
+    <version>1.0.56</version>
 
     <name>Ollama4j</name>
     <description>Java library for interacting with Ollama API.</description>
@@ -39,7 +39,7 @@
         <connection>scm:git:git@github.com:amithkoujalgi/ollama4j.git</connection>
         <developerConnection>scm:git:https://github.com/amithkoujalgi/ollama4j.git</developerConnection>
         <url>https://github.com/amithkoujalgi/ollama4j</url>
-        <tag>v1.0.54</tag>
+        <tag>v1.0.56</tag>
     </scm>
 
     <build>

src/main/java/io/github/amithkoujalgi/ollama4j/core/OllamaAPI.java

@@ -342,13 +342,24 @@ public class OllamaAPI {
    * @param options the Options object - <a
    *     href="https://github.com/jmorganca/ollama/blob/main/docs/modelfile.md#valid-parameters-and-values">More
    *     details on the options</a>
+   * @param streamHandler optional callback consumer that will be applied every time a streamed response is received. If not set, the stream parameter of the request is set to false.
    * @return OllamaResult that includes response text and time taken for response
    */
-  public OllamaResult generate(String model, String prompt, Options options)
+  public OllamaResult generate(String model, String prompt, Options options, OllamaStreamHandler streamHandler)
       throws OllamaBaseException, IOException, InterruptedException {
     OllamaGenerateRequestModel ollamaRequestModel = new OllamaGenerateRequestModel(model, prompt);
     ollamaRequestModel.setOptions(options.getOptionsMap());
-    return generateSyncForOllamaRequestModel(ollamaRequestModel);
+    return generateSyncForOllamaRequestModel(ollamaRequestModel,streamHandler);
+  }
+
+  /**
+   * Convenience method to call Ollama API without streaming responses.
+   * 
+   * Uses {@link #generate(String, String, Options, OllamaStreamHandler)}
+   */
+  public OllamaResult generate(String model, String prompt, Options options)
+  throws OllamaBaseException, IOException, InterruptedException {
+    return generate(model, prompt, options,null);
   }
 
   /**
@@ -381,10 +392,11 @@ public class OllamaAPI {
    * @param options the Options object - <a
    *     href="https://github.com/jmorganca/ollama/blob/main/docs/modelfile.md#valid-parameters-and-values">More
    *     details on the options</a>
+   * @param streamHandler optional callback consumer that will be applied every time a streamed response is received. If not set, the stream parameter of the request is set to false.
    * @return OllamaResult that includes response text and time taken for response
    */
   public OllamaResult generateWithImageFiles(
-      String model, String prompt, List<File> imageFiles, Options options)
+      String model, String prompt, List<File> imageFiles, Options options, OllamaStreamHandler streamHandler)
       throws OllamaBaseException, IOException, InterruptedException {
     List<String> images = new ArrayList<>();
     for (File imageFile : imageFiles) {
@@ -392,9 +404,20 @@ public class OllamaAPI {
     }
     OllamaGenerateRequestModel ollamaRequestModel = new OllamaGenerateRequestModel(model, prompt, images);
     ollamaRequestModel.setOptions(options.getOptionsMap());
-    return generateSyncForOllamaRequestModel(ollamaRequestModel);
+    return generateSyncForOllamaRequestModel(ollamaRequestModel,streamHandler);
   }
 
+   /**
+   * Convenience method to call Ollama API without streaming responses.
+   * 
+   * Uses {@link #generateWithImageFiles(String, String, List, Options, OllamaStreamHandler)}
+   */
+  public OllamaResult generateWithImageFiles(
+    String model, String prompt, List<File> imageFiles, Options options)
+    throws OllamaBaseException, IOException, InterruptedException{
+      return generateWithImageFiles(model, prompt, imageFiles, options, null);
+}
+
   /**
    * With one or more image URLs, ask a question to a model running on Ollama server. This is a
    * sync/blocking call.
@@ -405,10 +428,11 @@ public class OllamaAPI {
    * @param options the Options object - <a
    *     href="https://github.com/jmorganca/ollama/blob/main/docs/modelfile.md#valid-parameters-and-values">More
    *     details on the options</a>
+   * @param streamHandler optional callback consumer that will be applied every time a streamed response is received. If not set, the stream parameter of the request is set to false.
    * @return OllamaResult that includes response text and time taken for response
    */
   public OllamaResult generateWithImageURLs(
-      String model, String prompt, List<String> imageURLs, Options options)
+      String model, String prompt, List<String> imageURLs, Options options, OllamaStreamHandler streamHandler)
       throws OllamaBaseException, IOException, InterruptedException, URISyntaxException {
     List<String> images = new ArrayList<>();
     for (String imageURL : imageURLs) {
@@ -416,7 +440,18 @@ public class OllamaAPI {
     }
     OllamaGenerateRequestModel ollamaRequestModel = new OllamaGenerateRequestModel(model, prompt, images);
     ollamaRequestModel.setOptions(options.getOptionsMap());
-    return generateSyncForOllamaRequestModel(ollamaRequestModel);
+    return generateSyncForOllamaRequestModel(ollamaRequestModel,streamHandler);
+  }
+
+  /**
+   * Convenience method to call Ollama API without streaming responses.
+   * 
+   * Uses {@link #generateWithImageURLs(String, String, List, Options, OllamaStreamHandler)}
+   */
+  public OllamaResult generateWithImageURLs(String model, String prompt, List<String> imageURLs,
+      Options options)
+      throws OllamaBaseException, IOException, InterruptedException, URISyntaxException {
+    return generateWithImageURLs(model, prompt, imageURLs, options, null);
   }
 
 
@@ -487,10 +522,19 @@ public class OllamaAPI {
     return Base64.getEncoder().encodeToString(bytes);
   }
 
-  private OllamaResult generateSyncForOllamaRequestModel(OllamaGenerateRequestModel ollamaRequestModel)
+  private OllamaResult generateSyncForOllamaRequestModel(
+      OllamaGenerateRequestModel ollamaRequestModel, OllamaStreamHandler streamHandler)
       throws OllamaBaseException, IOException, InterruptedException {
-        OllamaGenerateEndpointCaller requestCaller = new OllamaGenerateEndpointCaller(host, basicAuth, requestTimeoutSeconds, verbose);
-        return requestCaller.callSync(ollamaRequestModel);
+    OllamaGenerateEndpointCaller requestCaller =
+        new OllamaGenerateEndpointCaller(host, basicAuth, requestTimeoutSeconds, verbose);
+    OllamaResult result;
+    if (streamHandler != null) {
+      ollamaRequestModel.setStream(true);
+      result = requestCaller.call(ollamaRequestModel, streamHandler);
+    } else {
+      result = requestCaller.callSync(ollamaRequestModel);
+    }
+    return result;
   }
 
   /**

src/main/java/io/github/amithkoujalgi/ollama4j/core/models/Model.java

@@ -1,6 +1,8 @@
 package io.github.amithkoujalgi.ollama4j.core.models;
 
 import com.fasterxml.jackson.annotation.JsonProperty;
+import com.fasterxml.jackson.core.JsonProcessingException;
+import io.github.amithkoujalgi.ollama4j.core.utils.Utils;
 import lombok.Data;
 
 @Data
@@ -34,4 +36,13 @@ public class Model {
     return name.split(":")[1];
   }
 
+    @Override
+  public String toString() {
+    try {
+      return Utils.getObjectMapper().writerWithDefaultPrettyPrinter().writeValueAsString(this);
+    } catch (JsonProcessingException e) {
+      throw new RuntimeException(e);
+    }
+  }
+
 }

src/main/java/io/github/amithkoujalgi/ollama4j/core/models/ModelDetail.java

@@ -2,7 +2,8 @@ package io.github.amithkoujalgi.ollama4j.core.models;
 
 import com.fasterxml.jackson.annotation.JsonIgnoreProperties;
 import com.fasterxml.jackson.annotation.JsonProperty;
-import java.util.Map;
+import com.fasterxml.jackson.core.JsonProcessingException;
+import io.github.amithkoujalgi.ollama4j.core.utils.Utils;
 import lombok.Data;
 
 @Data
@@ -16,5 +17,14 @@ public class ModelDetail {
   private String parameters;
   private String template;
   private String system;
-  private Map<String, String> details;
+  private ModelMeta details;
+
+    @Override
+  public String toString() {
+    try {
+      return Utils.getObjectMapper().writerWithDefaultPrettyPrinter().writeValueAsString(this);
+    } catch (JsonProcessingException e) {
+      throw new RuntimeException(e);
+    }
+  }
 }

src/main/java/io/github/amithkoujalgi/ollama4j/core/models/ModelMeta.java

@@ -2,6 +2,8 @@ package io.github.amithkoujalgi.ollama4j.core.models;
 
 import com.fasterxml.jackson.annotation.JsonIgnoreProperties;
 import com.fasterxml.jackson.annotation.JsonProperty;
+import com.fasterxml.jackson.core.JsonProcessingException;
+import io.github.amithkoujalgi.ollama4j.core.utils.Utils;
 import lombok.Data;
 
 @Data
@@ -21,4 +23,13 @@ public class ModelMeta {
 
   @JsonProperty("quantization_level")
   private String quantizationLevel;
+
+    @Override
+  public String toString() {
+    try {
+      return Utils.getObjectMapper().writerWithDefaultPrettyPrinter().writeValueAsString(this);
+    } catch (JsonProcessingException e) {
+      throw new RuntimeException(e);
+    }
+  }
 }

src/main/java/io/github/amithkoujalgi/ollama4j/core/models/chat/OllamaChatStreamObserver.java

@@ -2,10 +2,8 @@ package io.github.amithkoujalgi.ollama4j.core.models.chat;
 
 import java.util.ArrayList;
 import java.util.List;
-import java.util.stream.Collectors;
 
 import io.github.amithkoujalgi.ollama4j.core.OllamaStreamHandler;
-import lombok.NonNull;
 
 public class OllamaChatStreamObserver {
 
@@ -13,7 +11,7 @@ public class OllamaChatStreamObserver {
 
     private List<OllamaChatResponseModel> responseParts = new ArrayList<>();
 
-    private String message;
+    private String message = "";
 
     public OllamaChatStreamObserver(OllamaStreamHandler streamHandler) {
         this.streamHandler = streamHandler;
@@ -25,8 +23,7 @@ public class OllamaChatStreamObserver {
     }
     
     protected void handleCurrentResponsePart(OllamaChatResponseModel currentResponsePart){
-        List<@NonNull String> allResponsePartsByNow = responseParts.stream().map(r -> r.getMessage().getContent()).collect(Collectors.toList());
-        message = String.join("", allResponsePartsByNow);
+        message = message + currentResponsePart.getMessage().getContent();
         streamHandler.accept(message);
     }
 

src/main/java/io/github/amithkoujalgi/ollama4j/core/models/generate/OllamaGenerateStreamObserver.java

@@ -0,0 +1,31 @@
+package io.github.amithkoujalgi.ollama4j.core.models.generate;
+
+import java.util.ArrayList;
+import java.util.List;
+
+import io.github.amithkoujalgi.ollama4j.core.OllamaStreamHandler;
+
+public class OllamaGenerateStreamObserver {
+
+    private OllamaStreamHandler streamHandler;
+
+    private List<OllamaGenerateResponseModel> responseParts = new ArrayList<>();
+
+    private String message = "";
+
+    public OllamaGenerateStreamObserver(OllamaStreamHandler streamHandler) {
+        this.streamHandler = streamHandler;
+    }
+
+    public void notify(OllamaGenerateResponseModel currentResponsePart){
+        responseParts.add(currentResponsePart);
+        handleCurrentResponsePart(currentResponsePart);
+    }
+    
+    protected void handleCurrentResponsePart(OllamaGenerateResponseModel currentResponsePart){
+        message = message + currentResponsePart.getResponse();
+        streamHandler.accept(message);
+    }
+
+
+}

src/main/java/io/github/amithkoujalgi/ollama4j/core/models/request/OllamaGenerateEndpointCaller.java

@@ -1,18 +1,25 @@
 package io.github.amithkoujalgi.ollama4j.core.models.request;
 
+import java.io.IOException;
 import org.slf4j.Logger;
 import org.slf4j.LoggerFactory;
 
 import com.fasterxml.jackson.core.JsonProcessingException;
-
+import io.github.amithkoujalgi.ollama4j.core.OllamaStreamHandler;
+import io.github.amithkoujalgi.ollama4j.core.exceptions.OllamaBaseException;
 import io.github.amithkoujalgi.ollama4j.core.models.BasicAuth;
+import io.github.amithkoujalgi.ollama4j.core.models.OllamaResult;
 import io.github.amithkoujalgi.ollama4j.core.models.generate.OllamaGenerateResponseModel;
+import io.github.amithkoujalgi.ollama4j.core.models.generate.OllamaGenerateStreamObserver;
+import io.github.amithkoujalgi.ollama4j.core.utils.OllamaRequestBody;
 import io.github.amithkoujalgi.ollama4j.core.utils.Utils;
 
 public class OllamaGenerateEndpointCaller extends OllamaEndpointCaller{
 
     private static final Logger LOG = LoggerFactory.getLogger(OllamaGenerateEndpointCaller.class);
 
+    private OllamaGenerateStreamObserver streamObserver;
+
     public OllamaGenerateEndpointCaller(String host, BasicAuth basicAuth, long requestTimeoutSeconds, boolean verbose) {
         super(host, basicAuth, requestTimeoutSeconds, verbose);   
     }
@@ -27,6 +34,9 @@ public class OllamaGenerateEndpointCaller extends OllamaEndpointCaller{
                 try {
                     OllamaGenerateResponseModel ollamaResponseModel = Utils.getObjectMapper().readValue(line, OllamaGenerateResponseModel.class);
                     responseBuffer.append(ollamaResponseModel.getResponse());
+                    if(streamObserver != null) {
+                        streamObserver.notify(ollamaResponseModel);
+                    }
                     return ollamaResponseModel.isDone();
                 } catch (JsonProcessingException e) {
                     LOG.error("Error parsing the Ollama chat response!",e);
@@ -34,7 +44,11 @@ public class OllamaGenerateEndpointCaller extends OllamaEndpointCaller{
                 }         
     }
 
-    
+    public OllamaResult call(OllamaRequestBody body, OllamaStreamHandler streamHandler)
+        throws OllamaBaseException, IOException, InterruptedException {
+    streamObserver = new OllamaGenerateStreamObserver(streamHandler);
+    return super.callSync(body);
+    }
     
     
 }

src/test/java/io/github/amithkoujalgi/ollama4j/integrationtests/TestRealAPIs.java

@@ -4,6 +4,7 @@ import static org.junit.jupiter.api.Assertions.*;
 
 import io.github.amithkoujalgi.ollama4j.core.OllamaAPI;
 import io.github.amithkoujalgi.ollama4j.core.exceptions.OllamaBaseException;
+import io.github.amithkoujalgi.ollama4j.core.models.ModelDetail;
 import io.github.amithkoujalgi.ollama4j.core.models.OllamaResult;
 import io.github.amithkoujalgi.ollama4j.core.models.chat.OllamaChatMessageRole;
 import io.github.amithkoujalgi.ollama4j.core.models.chat.OllamaChatRequestBuilder;
@@ -91,6 +92,19 @@ class TestRealAPIs {
     }
   }
 
+  @Test
+  @Order(3)
+  void testListDtails() {
+    testEndpointReachability();
+    try {
+      ModelDetail modelDetails = ollamaAPI.getModelDetails(config.getModel());
+      assertNotNull(modelDetails);
+      System.out.println(modelDetails);
+    } catch (IOException | OllamaBaseException | InterruptedException | URISyntaxException e) {
+      throw new RuntimeException(e);
+    }
+  }
+
   @Test
   @Order(3)
   void testAskModelWithDefaultOptions() {
@@ -109,6 +123,32 @@ class TestRealAPIs {
     }
   }
 
+  @Test
+  @Order(3)
+  void testAskModelWithDefaultOptionsStreamed() {
+    testEndpointReachability();
+    try {
+
+      StringBuffer sb = new StringBuffer("");
+
+      OllamaResult result = ollamaAPI.generate(config.getModel(),
+          "What is the capital of France? And what's France's connection with Mona Lisa?",
+          new OptionsBuilder().build(), (s) -> {
+            LOG.info(s);
+            String substring = s.substring(sb.toString().length(), s.length());
+            LOG.info(substring);
+            sb.append(substring);
+          });
+
+      assertNotNull(result);
+      assertNotNull(result.getResponse());
+      assertFalse(result.getResponse().isEmpty());
+      assertEquals(sb.toString().trim(), result.getResponse().trim());
+    } catch (IOException | OllamaBaseException | InterruptedException e) {
+      throw new RuntimeException(e);
+    }
+  }
+
   @Test
   @Order(3)
   void testAskModelWithOptions() {
@@ -183,7 +223,7 @@ class TestRealAPIs {
 
       OllamaChatResult chatResult = ollamaAPI.chat(requestModel,(s) -> {
         LOG.info(s);
-        String substring = s.substring(sb.toString().length(), s.length()-1);
+        String substring = s.substring(sb.toString().length(), s.length());
         LOG.info(substring);
         sb.append(substring);
       });
@@ -262,6 +302,30 @@ class TestRealAPIs {
     }
   }
 
+  @Test
+  @Order(3)
+  void testAskModelWithOptionsAndImageFilesStreamed() {
+    testEndpointReachability();
+    File imageFile = getImageFileFromClasspath("dog-on-a-boat.jpg");
+    try {
+      StringBuffer sb = new StringBuffer("");
+
+      OllamaResult result = ollamaAPI.generateWithImageFiles(config.getImageModel(),
+          "What is in this image?", List.of(imageFile), new OptionsBuilder().build(), (s) -> {
+            LOG.info(s);
+            String substring = s.substring(sb.toString().length(), s.length());
+            LOG.info(substring);
+            sb.append(substring);
+          });
+      assertNotNull(result);
+      assertNotNull(result.getResponse());
+      assertFalse(result.getResponse().isEmpty());
+      assertEquals(sb.toString().trim(), result.getResponse().trim());
+    } catch (IOException | OllamaBaseException | InterruptedException e) {
+      throw new RuntimeException(e);
+    }
+  }
+
   @Test
   @Order(3)
   void testAskModelWithOptionsAndImageURLs() {