updating python and adding send on enter

.gitignore

@@ -75,4 +75,5 @@ build/
 .idea
 bin/
 dist/
-application-local.yaml
+application-local.yaml
+service/python/config.json

K8S.md

@@ -48,7 +48,7 @@ Change to the new folder:
 cd oci-generative-ai-jet-ui
 ```
 
-Install Node.js 16 on Cloud Shell.
+Install Node.js 18 on Cloud Shell.
 
 ```bash
 nvm install 18 && nvm use 18
@@ -130,6 +130,7 @@ Run `get deploy` a few times:
 
 ```bash
 kubectl get deploy -n backend
+kubectl get pods -n backend
 ```
 
 Wait for all deployments to be `Ready` and `Available`.

LOCAL.md

@@ -1,4 +1,13 @@
-# Run Local
+# Running Java Backend for Local Development
+
+This guide provides step-by-step instructions for running the Java backend application locally for development purposes.
+
+Prerequisites
+Ensure you have the following installed on your system:
+Java Development Kit (JDK) 11 or later
+Gradle 6.8 or later
+Oracle JDBC Driver (if using an Oracle database)
+Familiarize yourself with the project structure and configuration files.
 
 ## Run components
 

app/package.json

@@ -1,6 +1,6 @@
 {
   "name": "JETGenAI",
-  "version": "1.0.0",
+  "version": "1.0.1",
   "description": "Sample Client app showing communication with OCI Generative AI services via Websocket",
   "dependencies": {
     "@oracle/oraclejet": "~16.1.0",

app/src/components/content/chat.tsx

@@ -68,6 +68,14 @@ export const Chat = ({ testId, data, questionChanged, question }: Props) => {
     );
   };
 
+  const handleQuestionChange = (event: any) => {
+    const newValue = event.detail.value.trim();
+    if (newValue !== "") {
+      questionChanged(event);
+      question.current = ""; // Clear the input field after adding the question
+    }
+  };
+
   return (
     <>
       <div class="oj-flex-item">
@@ -91,8 +99,8 @@ export const Chat = ({ testId, data, questionChanged, question }: Props) => {
         value={question?.current}
         placeholder="ask me anything..."
         aria-label="enter a question"
-        onojValueAction={questionChanged}
+        onojValueAction={handleQuestionChange}
       ></oj-input-search>
     </>
   );
-};
+};

app/src/components/content/settings.tsx

@@ -81,9 +81,8 @@ export const Settings = (props: Props) => {
       const json = await response.json();
       const result = json.filter((model: Model) => {
         if (
-          model.capabilities.includes("TEXT_GENERATION") &&
-          (model.vendor == "cohere" || model.vendor == "") &&
-          model.version != "14.2"
+          model.capabilities.includes("CHAT") &&
+          (model.vendor == "cohere" || model.vendor == "meta")
         )
           return model;
       });

backend/src/main/java/dev/victormartin/oci/genai/backend/backend/controller/GenAIController.java

@@ -8,6 +8,7 @@
 import com.oracle.bmc.generativeai.responses.ListEndpointsResponse;
 import dev.victormartin.oci.genai.backend.backend.dao.GenAiModel;
 import dev.victormartin.oci.genai.backend.backend.dao.GenAiEndpoint;
+import dev.victormartin.oci.genai.backend.backend.service.GenAIModelsService;
 import dev.victormartin.oci.genai.backend.backend.service.GenAiClientService;
 import org.slf4j.Logger;
 import org.slf4j.LoggerFactory;
@@ -29,19 +30,16 @@ public class GenAIController {
     @Autowired
     private GenAiClientService generativeAiClientService;
 
+    @Autowired
+    private GenAIModelsService genAIModelsService;
+
     @GetMapping("/api/genai/models")
     public List<GenAiModel> getModels() {
         logger.info("getModels()");
-        ListModelsRequest listModelsRequest = ListModelsRequest.builder().compartmentId(COMPARTMENT_ID).build();
-        GenerativeAiClient client = generativeAiClientService.getClient();
-        ListModelsResponse response = client.listModels(listModelsRequest);
-        return response.getModelCollection().getItems().stream().map(m -> {
-            List<String> capabilities = m.getCapabilities().stream().map(ModelCapability::getValue)
-                    .collect(Collectors.toList());
-            GenAiModel model = new GenAiModel(m.getId(), m.getDisplayName(), m.getVendor(), m.getVersion(),
-                    capabilities, m.getTimeCreated());
-            return model;
-        }).collect(Collectors.toList());
+        List<GenAiModel> models = genAIModelsService.getModels();
+        return models.stream()
+                .filter(m -> m.capabilities().contains("CHAT"))
+                .collect(Collectors.toList());
     }
 
     @GetMapping("/api/genai/endpoints")

backend/src/main/java/dev/victormartin/oci/genai/backend/backend/controller/PromptController.java

@@ -61,7 +61,7 @@ public Answer handlePrompt(Prompt prompt) {
                 throw new InvalidPromptRequest();
             }
             saved.setDatetimeResponse(new Date());
-            String responseFromGenAI = genAI.resolvePrompt(promptEscaped, activeModel, finetune);
+            String responseFromGenAI = genAI.resolvePrompt(promptEscaped, activeModel, finetune, false);
             saved.setResponse(responseFromGenAI);
             interactionRepository.save(saved);
             return new Answer(responseFromGenAI, "");

backend/src/main/java/dev/victormartin/oci/genai/backend/backend/service/GenAIModelsService.java

@@ -0,0 +1,44 @@
+package dev.victormartin.oci.genai.backend.backend.service;
+
+import com.oracle.bmc.generativeai.GenerativeAiClient;
+import com.oracle.bmc.generativeai.model.ModelCapability;
+import com.oracle.bmc.generativeai.requests.ListModelsRequest;
+import com.oracle.bmc.generativeai.responses.ListModelsResponse;
+import dev.victormartin.oci.genai.backend.backend.dao.GenAiModel;
+import org.slf4j.Logger;
+import org.slf4j.LoggerFactory;
+import org.springframework.beans.factory.annotation.Autowired;
+import org.springframework.beans.factory.annotation.Value;
+import org.springframework.stereotype.Service;
+
+import java.util.List;
+import java.util.stream.Collectors;
+
+@Service
+public class GenAIModelsService {
+    Logger log = LoggerFactory.getLogger(GenAIModelsService.class);
+
+    @Value("${genai.compartment_id}")
+    private String COMPARTMENT_ID;
+
+    @Autowired
+    private GenAiClientService generativeAiClientService;
+
+    public List<GenAiModel> getModels() {
+        log.info("getModels()");
+        ListModelsRequest listModelsRequest = ListModelsRequest.builder()
+                .compartmentId(COMPARTMENT_ID)
+                .build();
+        GenerativeAiClient client = generativeAiClientService.getClient();
+        ListModelsResponse response = client.listModels(listModelsRequest);
+        return response.getModelCollection().getItems().stream()
+                .map(m -> {
+                    List<String> capabilities = m.getCapabilities().stream()
+                            .map(ModelCapability::getValue).collect(Collectors.toList());
+                    GenAiModel model = new GenAiModel(
+                            m.getId(), m.getDisplayName(), m.getVendor(),
+                            m.getVersion(), capabilities, m.getTimeCreated());
+                    return model;
+                }).collect(Collectors.toList());
+    }
+}

backend/src/main/java/dev/victormartin/oci/genai/backend/backend/service/OCIGenAIService.java

@@ -1,48 +1,94 @@
 package dev.victormartin.oci.genai.backend.backend.service;
 
-import com.oracle.bmc.generativeaiinference.GenerativeAiInferenceClient;
 import com.oracle.bmc.generativeaiinference.model.*;
+import com.oracle.bmc.generativeaiinference.model.Message;
 import com.oracle.bmc.generativeaiinference.requests.ChatRequest;
-import com.oracle.bmc.generativeaiinference.requests.GenerateTextRequest;
-import com.oracle.bmc.generativeaiinference.requests.SummarizeTextRequest;
 import com.oracle.bmc.generativeaiinference.responses.ChatResponse;
-import com.oracle.bmc.generativeaiinference.responses.GenerateTextResponse;
-import com.oracle.bmc.generativeaiinference.responses.SummarizeTextResponse;
-import com.oracle.bmc.http.client.jersey.WrappedResponseInputStream;
-import org.hibernate.boot.archive.scan.internal.StandardScanner;
+import dev.victormartin.oci.genai.backend.backend.dao.GenAiModel;
+import org.slf4j.Logger;
+import org.slf4j.LoggerFactory;
 import org.springframework.beans.factory.annotation.Autowired;
 import org.springframework.beans.factory.annotation.Value;
 import org.springframework.stereotype.Service;
 
-import java.io.*;
-import java.nio.charset.StandardCharsets;
+import java.util.ArrayList;
 import java.util.List;
-import java.util.stream.Collectors;
 
 @Service
 public class OCIGenAIService {
+
+        Logger log = LoggerFactory.getLogger(OCIGenAIService.class);
+
         @Value("${genai.compartment_id}")
         private String COMPARTMENT_ID;
 
         @Autowired
         private GenAiInferenceClientService generativeAiInferenceClientService;
 
-        public String resolvePrompt(String input, String modelId, boolean finetune) {
-                CohereChatRequest cohereChatRequest = CohereChatRequest.builder()
-                        .message(input)
-                        .maxTokens(600)
-                        .temperature((double) 1)
-                        .frequencyPenalty((double) 0)
-                        .topP((double) 0.75)
-                        .topK(0)
-                        .isStream(false) // TODO websockets and streams
-                        .build();
+        @Autowired
+        private GenAIModelsService genAIModelsService;
 
-                ChatDetails chatDetails = ChatDetails.builder()
-                        .servingMode(OnDemandServingMode.builder().modelId(modelId).build())
-                        .compartmentId(COMPARTMENT_ID)
-                        .chatRequest(cohereChatRequest)
-                        .build();
+        public String resolvePrompt(String input, String modelId, boolean finetune, boolean summarization) {
+
+                List<GenAiModel> models = genAIModelsService.getModels();
+                GenAiModel currentModel = models.stream()
+                        .filter(m-> modelId.equals(m.id()))
+                        .findFirst()
+                        .orElseThrow();
+
+                log.info("Model {} with finetune {}", currentModel.name(), finetune? "yes" : "no");
+
+                double temperature = summarization?0.0:0.5;
+
+                String inputText = summarization?"Summarize this text:\n" + input: input;
+
+                        ChatDetails chatDetails;
+                switch (currentModel.vendor()) {
+                        case "cohere":
+                                CohereChatRequest cohereChatRequest = CohereChatRequest.builder()
+                                        .message(inputText)
+                                        .maxTokens(600)
+                                        .temperature(temperature)
+                                        .frequencyPenalty((double) 0)
+                                        .topP(0.75)
+                                        .topK(0)
+                                        .isStream(false) // TODO websockets and streams
+                                        .build();
+
+                                chatDetails = ChatDetails.builder()
+                                        .servingMode(OnDemandServingMode.builder().modelId(currentModel.id()).build())
+                                        .compartmentId(COMPARTMENT_ID)
+                                        .chatRequest(cohereChatRequest)
+                                        .build();
+                                break;
+                        case "meta":
+                                ChatContent content = TextContent.builder()
+                                        .text(inputText)
+                                        .build();
+                                List<ChatContent> contents = new ArrayList<>();
+                                contents.add(content);
+                                List<Message> messages = new ArrayList<>();
+                                Message message = new UserMessage(contents, "user");
+                                messages.add(message);
+                                GenericChatRequest genericChatRequest = GenericChatRequest.builder()
+                                        .messages(messages)
+                                        .maxTokens(600)
+                                        .temperature((double)1)
+                                        .frequencyPenalty((double)0)
+                                        .presencePenalty((double)0)
+                                        .topP(0.75)
+                                        .topK(-1)
+                                        .isStream(false)
+                                        .build();
+                                chatDetails = ChatDetails.builder()
+                                        .servingMode(OnDemandServingMode.builder().modelId(currentModel.id()).build())
+                                        .compartmentId(COMPARTMENT_ID)
+                                        .chatRequest(genericChatRequest)
+                                        .build();
+                                break;
+                        default:
+                                throw new IllegalStateException("Unexpected value: " + currentModel.vendor());
+                }
 
                 ChatRequest request = ChatRequest.builder()
                         .chatDetails(chatDetails)
@@ -65,7 +111,7 @@ public String resolvePrompt(String input, String modelId, boolean finetune) {
         }
 
         public String summaryText(String input, String modelId, boolean finetuned) {
-                String response = resolvePrompt("Summarize this:\n" + input, modelId, finetuned);
+                String response = resolvePrompt(input, modelId, finetuned, true);
                 return response;
         }
 }

scripts/release.mjs

@@ -53,7 +53,7 @@ config.set("ocir_user_email", ocir_user_email);
 config.set("ocir_user_token", ocir_user_token);
 
 await containerLogin(namespace, ocir_user, ocir_user_token, ocirUrl);
-await releaseWeb();
+// await releaseWeb();
 await releaseApp();
 await releaseBackend();
 

service/python/config.json.txt

@@ -0,0 +1,8 @@
+{
+    "compartment_id": "<your_compartment_id>",
+    "config_profile": "<your_config_profile>",
+    "service_endpoint": "https://inference.generativeai.us-chicago-1.oci.oraclecloud.com",
+    "model_type": "<cohere|llama>",
+    "cohere_model_id": "<your_cohere_model_id>",
+    "llama_model_id": "<your_llama_model_id>"
+}

service/python/requirements.txt

@@ -4,7 +4,7 @@ cached-property==1.5.2
 certifi==2024.7.4
 cffi==1.16.0
 circuitbreaker==1.4.0
-cryptography==43.0.1
+cryptography==42.0.6
 oci==2.126.2
 pycparser==2.21
 pyOpenSSL==24.1.0