package com.bruce.modelprovider.gateway;

import com.bruce.modelprovider.client.OpenAiCompatibleModelClient;
import com.bruce.modelprovider.entity.ModelCallLog;
import com.bruce.modelprovider.entity.ModelConfig;
import com.bruce.modelprovider.entity.ModelProvider;
import com.bruce.modelprovider.enums.ModelCallStatusEnum;
import com.bruce.modelprovider.route.ModelRouteContext;
import com.bruce.modelprovider.route.ModelRouteDecision;
import com.bruce.modelprovider.service.IModelCallLogService;
import com.bruce.modelprovider.service.IModelProviderService;
import com.bruce.modelprovider.service.IModelRouteService;
import lombok.RequiredArgsConstructor;
import org.springframework.stereotype.Component;
import org.springframework.util.DigestUtils;

import java.nio.charset.StandardCharsets;
import java.util.List;
import java.util.UUID;

/**
 * Embedding 网关实现。
 * <p>
 * 主要职责：
 * 1. 将业务请求转换为统一路由上下文并完成模型决策；
 * 2. 调用上游 Embedding 接口并执行结果校验（数量、维度）；
 * 3. 记录调用日志（成功/失败、耗时、错误摘要、请求哈希）；
 * 4. 在主模型失败时按备用模型顺序执行兜底调用。
 */
@Component
@RequiredArgsConstructor
/**
 * EmbeddingModelGatewayImpl，负责模型平台对应层的职责。
 */
public class EmbeddingModelGatewayImpl implements EmbeddingModelGateway {

    private final IModelRouteService modelRouteService;
    private final IModelProviderService modelProviderService;
    private final IModelCallLogService modelCallLogService;
    private final OpenAiCompatibleModelClient openAiCompatibleModelClient;

    /**
     * 统一 Embedding 调用入口。
     *
     * @param request 向量化请求，包含文本、任务类型和业务上下文
     * @return 向量化结果（模型信息 + 向量数组 + 调用日志）
     */
    @Override
    /**
     * 方法 embed，用于执行业务逻辑处理。
     */
    public EmbeddingResult embed(EmbeddingRequest request) {
        long start = System.currentTimeMillis();
        ModelCallLog callLog = new ModelCallLog();
        callLog.setRequestId(UUID.randomUUID().toString().replace("-", ""));
        callLog.setTaskType(request.getTaskType());
        callLog.setBizType(request.getBizType());
        callLog.setBizId(request.getBizId());
        callLog.setCallType("EMBEDDING");
        callLog.setRequestHash(DigestUtils.md5DigestAsHex(String.join("|", request.getTexts()).getBytes(StandardCharsets.UTF_8)));

        try {
            ModelRouteContext routeContext = new ModelRouteContext();
            routeContext.setTaskType(request.getTaskType());
            routeContext.setMatchScope(request.getMatchScope());
            routeContext.setScopeId(request.getScopeId());
            routeContext.setRequiredModelType("EMBEDDING");
            routeContext.setRequiredEmbeddingDimension(request.getExpectedDimension());
            routeContext.setBizType(request.getBizType());
            routeContext.setBizId(request.getBizId());
            ModelRouteDecision decision = modelRouteService.route(routeContext);

            ModelConfig model = decision.getPrimaryModel();
            ModelProvider provider = modelProviderService.getById(model.getProviderId());
            if (provider == null || !Boolean.TRUE.equals(provider.getEnabled())) {
                throw new IllegalStateException("模型服务商不可用");
            }

            List<List<Double>> vectors = executeWithFallback(provider, model, decision.getFallbackModels(), request.getTexts(), request.getExpectedDimension());
            if (vectors.size() != request.getTexts().size()) {
                throw new IllegalStateException("向量数量与输入文本数量不一致");
            }
            Integer dimension = vectors.isEmpty() ? 0 : vectors.getFirst().size();
            if (request.getExpectedDimension() != null && !request.getExpectedDimension().equals(dimension)) {
                throw new IllegalStateException("向量维度不匹配，expected=" + request.getExpectedDimension() + ", actual=" + dimension);
            }

            callLog.setProviderId(provider.getId());
            callLog.setModelId(model.getId());
            callLog.setStatus(ModelCallStatusEnum.SUCCESS.name());
            callLog.setDurationMs((int) (System.currentTimeMillis() - start));
            modelCallLogService.save(callLog);

            EmbeddingResult result = new EmbeddingResult();
            result.setModelId(model.getId());
            result.setModelName(model.getModelName());
            result.setDimension(dimension);
            result.setVectors(vectors);
            result.setCallLog(callLog);
            return result;
        } catch (Exception ex) {
            callLog.setStatus(ModelCallStatusEnum.FAILED.name());
            callLog.setDurationMs((int) (System.currentTimeMillis() - start));
            callLog.setErrorCode("EMBEDDING_FAILED");
            String msg = ex.getMessage();
            callLog.setErrorMessage(msg == null ? "unknown" : msg.substring(0, Math.min(msg.length(), 1000)));
            modelCallLogService.save(callLog);
            throw ex;
        }
    }

    /**
     * 主模型优先调用，失败后按备用模型顺序重试。
     */
    private List<List<Double>> executeWithFallback(ModelProvider primaryProvider,
                                                    ModelConfig primaryModel,
                                                    List<ModelConfig> fallbackModels,
                                                    List<String> texts,
                                                    Integer expectedDimension) {
        try {
            return openAiCompatibleModelClient.embeddings(primaryProvider, primaryModel, texts, expectedDimension);
        } catch (Exception primaryEx) {
            for (ModelConfig fallback : fallbackModels) {
                try {
                    ModelProvider provider = modelProviderService.getById(fallback.getProviderId());
                    if (provider == null || !Boolean.TRUE.equals(provider.getEnabled())) {
                        continue;
                    }
                    return openAiCompatibleModelClient.embeddings(provider, fallback, texts, expectedDimension);
                } catch (Exception ignored) {
                    // continue fallback chain
                }
            }
            throw primaryEx;
        }
    }
}