topdog/supersonic
1
0
Fork 0
mirror of https://github.com/tencentmusic/supersonic.git synced 2026-05-07 10:24:21 +08:00
Code Issues Actions Packages Projects Releases Wiki Activity

[Improvement][headless] Only expose OPEN_AI/OLLAMA/DIFY chat model providers. (#2237)

Browse Source
This commit is contained in:
beat4ocean
2025-05-05 15:51:26 +08:00
committed by GitHub
parent acffc03c79
commit c4992501bd
26 changed files with 25 additions and 980 deletions
Download Patch File Download Diff File Expand all files Collapse all files

5
common/src/main/java/dev/langchain4j/model/openai/OpenAiChatModel.java
View File

@@ -66,7 +66,6 @@ import static java.util.Collections.singletonList;
@Slf4j
public class OpenAiChatModel implements ChatLanguageModel, TokenCountEstimator {
public static final String ZHIPU = "bigmodel";
private final OpenAiClient client;
private final String baseUrl;
private final String modelName;
@@ -192,9 +191,7 @@ public class OpenAiChatModel implements ChatLanguageModel, TokenCountEstimator {
.responseFormat(responseFormat).seed(seed).user(user)
.parallelToolCalls(parallelToolCalls);
if (!(baseUrl.contains(ZHIPU))) {
requestBuilder.temperature(temperature);
}
requestBuilder.temperature(temperature);
if (toolSpecifications != null && !toolSpecifications.isEmpty()) {
requestBuilder.tools(toTools(toolSpecifications, strictTools));

16
common/src/main/java/dev/langchain4j/model/zhipu/ChatCompletionModel.java
View File

@@ -1,16 +0,0 @@
package dev.langchain4j.model.zhipu;
public enum ChatCompletionModel {
GLM_4("glm-4"), GLM_3_TURBO("glm-3-turbo"), CHATGLM_TURBO("chatglm_turbo");
private final String value;
ChatCompletionModel(String value) {
this.value = value;
}
@Override
public String toString() {
return this.value;
}
}

100
common/src/main/java/dev/langchain4j/model/zhipu/ZhipuAiChatModel.java
View File

@@ -1,100 +0,0 @@
package dev.langchain4j.model.zhipu;
import dev.langchain4j.agent.tool.ToolSpecification;
import dev.langchain4j.data.message.AiMessage;
import dev.langchain4j.data.message.ChatMessage;
import dev.langchain4j.model.chat.ChatLanguageModel;
import dev.langchain4j.model.output.Response;
import dev.langchain4j.model.zhipu.chat.ChatCompletionRequest;
import dev.langchain4j.model.zhipu.chat.ChatCompletionResponse;
import dev.langchain4j.model.zhipu.spi.ZhipuAiChatModelBuilderFactory;
import lombok.Builder;
import java.util.List;
import static dev.langchain4j.internal.RetryUtils.withRetry;
import static dev.langchain4j.internal.Utils.getOrDefault;
import static dev.langchain4j.internal.Utils.isNullOrEmpty;
import static dev.langchain4j.internal.ValidationUtils.ensureNotEmpty;
import static dev.langchain4j.model.zhipu.DefaultZhipuAiHelper.aiMessageFrom;
import static dev.langchain4j.model.zhipu.DefaultZhipuAiHelper.finishReasonFrom;
import static dev.langchain4j.model.zhipu.DefaultZhipuAiHelper.toTools;
import static dev.langchain4j.model.zhipu.DefaultZhipuAiHelper.toZhipuAiMessages;
import static dev.langchain4j.model.zhipu.DefaultZhipuAiHelper.tokenUsageFrom;
import static dev.langchain4j.model.zhipu.chat.ToolChoiceMode.AUTO;
import static dev.langchain4j.spi.ServiceHelper.loadFactories;
import static java.util.Collections.singletonList;
/**
* Represents an ZhipuAi language model with a chat completion interface, such as glm-3-turbo and
* glm-4. You can find description of parameters
* <a href="https://open.bigmodel.cn/dev/api">here</a>.
*/
public class ZhipuAiChatModel implements ChatLanguageModel {
private final String baseUrl;
private final Double temperature;
private final Double topP;
private final String model;
private final Integer maxRetries;
private final Integer maxToken;
private final ZhipuAiClient client;
@Builder
public ZhipuAiChatModel(String baseUrl, String apiKey, Double temperature, Double topP,
String model, Integer maxRetries, Integer maxToken, Boolean logRequests,
Boolean logResponses) {
this.baseUrl = getOrDefault(baseUrl, "https://open.bigmodel.cn/");
this.temperature = getOrDefault(temperature, 0.7);
this.topP = topP;
this.model = getOrDefault(model, ChatCompletionModel.GLM_4.toString());
this.maxRetries = getOrDefault(maxRetries, 3);
this.maxToken = getOrDefault(maxToken, 512);
this.client = ZhipuAiClient.builder().baseUrl(this.baseUrl).apiKey(apiKey)
.logRequests(getOrDefault(logRequests, false))
.logResponses(getOrDefault(logResponses, false)).build();
}
public static ZhipuAiChatModelBuilder builder() {
for (ZhipuAiChatModelBuilderFactory factories : loadFactories(
ZhipuAiChatModelBuilderFactory.class)) {
return factories.get();
}
return new ZhipuAiChatModelBuilder();
}
@Override
public Response<AiMessage> generate(List<ChatMessage> messages) {
return generate(messages, (ToolSpecification) null);
}
@Override
public Response<AiMessage> generate(List<ChatMessage> messages,
List<ToolSpecification> toolSpecifications) {
ensureNotEmpty(messages, "messages");
ChatCompletionRequest.Builder requestBuilder =
ChatCompletionRequest.builder().model(this.model).maxTokens(maxToken).stream(false)
.topP(topP).toolChoice(AUTO).messages(toZhipuAiMessages(messages));
if (!isNullOrEmpty(toolSpecifications)) {
requestBuilder.tools(toTools(toolSpecifications));
}
ChatCompletionResponse response =
withRetry(() -> client.chatCompletion(requestBuilder.build()), maxRetries);
return Response.from(aiMessageFrom(response), tokenUsageFrom(response.getUsage()),
finishReasonFrom(response.getChoices().get(0).getFinishReason()));
}
@Override
public Response<AiMessage> generate(List<ChatMessage> messages,
ToolSpecification toolSpecification) {
return generate(messages,
toolSpecification != null ? singletonList(toolSpecification) : null);
}
public static class ZhipuAiChatModelBuilder {
public ZhipuAiChatModelBuilder() {}
}
}