客户管理功能调整， 音频模型调整
This commit is contained in:
spllzh
@@ -12,6 +12,8 @@ import lombok.extern.slf4j.Slf4j;
|
||||
import org.springframework.beans.factory.annotation.Autowired;
|
||||
import org.springframework.beans.factory.annotation.Value;
|
||||
import org.springframework.stereotype.Service;
|
||||
import com.fasterxml.jackson.databind.ObjectMapper;
|
||||
import com.fasterxml.jackson.databind.JsonNode;
|
||||
|
||||
import java.time.LocalDateTime;
|
||||
import java.util.ArrayList;
|
||||
@@ -171,54 +173,83 @@ public class TtsRequestLogServiceImpl implements ITtsRequestLogService {
|
||||
log.info("语音识别完成 - result: {} ", result);
|
||||
long processingTime = System.currentTimeMillis() - startTime;
|
||||
|
||||
if (result.getOutput() != null) {
|
||||
// 解析识别结果 - 从JSON中提取文本
|
||||
if (result.getResults() != null && !result.getResults().isEmpty()) {
|
||||
// 解析识别结果 - 从results中获取transcriptionUrl并下载识别结果
|
||||
String fullText = "";
|
||||
Double duration = 0.0;
|
||||
List<AsrResponse.AsrResult> asrResults = new ArrayList<>();
|
||||
|
||||
try {
|
||||
// 从输出中提取文本和时长信息
|
||||
if (result.getOutput().toString().contains("\"text\"")) {
|
||||
// 简单解析JSON获取文本
|
||||
String outputStr = result.getOutput().toString();
|
||||
int textStart = outputStr.indexOf("\"text\":\"") + 8;
|
||||
int textEnd = outputStr.indexOf("\"", textStart);
|
||||
if (textStart > 7 && textEnd > textStart) {
|
||||
fullText = outputStr.substring(textStart, textEnd);
|
||||
// 从usage中获取时长信息
|
||||
if (result.getUsage() != null) {
|
||||
// 从usage对象中获取duration，usage通常是一个JsonObject
|
||||
try {
|
||||
String usageStr = result.getUsage().toString();
|
||||
if (usageStr.contains("\"duration\"")) {
|
||||
int durationStart = usageStr.indexOf("\"duration\":") + 11;
|
||||
int durationEnd = usageStr.indexOf(",", durationStart);
|
||||
if (durationEnd == -1) {
|
||||
durationEnd = usageStr.indexOf("}", durationStart);
|
||||
}
|
||||
if (durationStart > 10 && durationEnd > durationStart) {
|
||||
duration = Double.parseDouble(usageStr.substring(durationStart, durationEnd).trim());
|
||||
}
|
||||
}
|
||||
} catch (Exception e) {
|
||||
log.warn("解析usage中的duration失败: {}", e.getMessage());
|
||||
}
|
||||
}
|
||||
|
||||
// 提取时长信息
|
||||
if (result.getOutput().toString().contains("\"duration\"")) {
|
||||
String outputStr = result.getOutput().toString();
|
||||
int durationStart = outputStr.indexOf("\"duration\":") + 11;
|
||||
int durationEnd = outputStr.indexOf(",", durationStart);
|
||||
if (durationEnd == -1) {
|
||||
durationEnd = outputStr.indexOf("}", durationStart);
|
||||
}
|
||||
if (durationStart > 10 && durationEnd > durationStart) {
|
||||
// 遍历所有结果，获取transcriptionUrl并下载识别文本
|
||||
for (TranscriptionTaskResult taskResult : result.getResults()) {
|
||||
if (taskResult.getTranscriptionUrl() != null) {
|
||||
try {
|
||||
duration = Double.parseDouble(outputStr.substring(durationStart, durationEnd).trim());
|
||||
} catch (NumberFormatException e) {
|
||||
duration = 0.0;
|
||||
// 从transcriptionUrl下载识别结果
|
||||
String transcriptionResult = downloadTranscriptionResult(taskResult.getTranscriptionUrl());
|
||||
log.info("transcriptionResult: {}" , transcriptionResult);
|
||||
|
||||
if (transcriptionResult != null && !transcriptionResult.trim().isEmpty()) {
|
||||
// 解析transcription结果JSON
|
||||
String extractedText = parseTranscriptionResult(transcriptionResult);
|
||||
if (extractedText != null && !extractedText.trim().isEmpty()) {
|
||||
fullText += extractedText;
|
||||
|
||||
// 输出识别文本到控制台
|
||||
System.out.println("=== 语音识别结果 ===");
|
||||
System.out.println("任务ID: " + taskId);
|
||||
System.out.println("识别文本: " + extractedText);
|
||||
System.out.println("==================");
|
||||
|
||||
// 创建结果对象
|
||||
AsrResponse.AsrResult asrResult = new AsrResponse.AsrResult();
|
||||
asrResult.setText(extractedText);
|
||||
asrResult.setStartTime(0.0);
|
||||
asrResult.setEndTime(duration);
|
||||
asrResults.add(asrResult);
|
||||
}
|
||||
}
|
||||
} catch (Exception e) {
|
||||
log.warn("下载transcription结果失败: {}", e.getMessage());
|
||||
}
|
||||
}
|
||||
}
|
||||
|
||||
} catch (Exception e) {
|
||||
log.warn("解析ASR结果时出错: {}", e.getMessage());
|
||||
}
|
||||
|
||||
List<AsrResponse.AsrResult> asrResults = new ArrayList<>();
|
||||
|
||||
// 创建简单的结果对象
|
||||
AsrResponse.AsrResult asrResult = new AsrResponse.AsrResult();
|
||||
asrResult.setText(fullText);
|
||||
asrResult.setStartTime(0.0);
|
||||
asrResult.setEndTime(duration);
|
||||
asrResults.add(asrResult);
|
||||
if (fullText.trim().isEmpty()) {
|
||||
log.warn("ASR识别结果为空 - TaskId: {}", taskId);
|
||||
return AsrResponse.error("语音识别结果为空");
|
||||
}
|
||||
|
||||
log.info("语音识别完成 - 识别文本长度: {}, 处理时间: {}ms", fullText.length(), processingTime);
|
||||
|
||||
// 如果提供了TTS日志ID，则保存识别文本到数据库
|
||||
if (request.getId() != null && !request.getId().trim().isEmpty()) {
|
||||
saveRecognizedTextToDatabase(request.getId(), fullText);
|
||||
}
|
||||
|
||||
return AsrResponse.success(fullText, asrResults, duration, processingTime, taskId, requestId);
|
||||
} else {
|
||||
log.warn("ASR识别结果为空 - TaskId: {}", taskId);
|
||||
@@ -235,4 +266,135 @@ public class TtsRequestLogServiceImpl implements ITtsRequestLogService {
|
||||
public boolean isAsrServiceAvailable() {
|
||||
return asrEnabled && apiKey != null && !apiKey.trim().isEmpty();
|
||||
}
|
||||
|
||||
/**
|
||||
* 下载transcription结果
|
||||
*/
|
||||
private String downloadTranscriptionResult(String transcriptionUrl) {
|
||||
try {
|
||||
log.info("开始下载transcription结果，URL: {}", transcriptionUrl);
|
||||
|
||||
// 使用简单的URL连接下载transcription结果，避免HttpClient依赖问题
|
||||
java.net.URL url = new java.net.URL(transcriptionUrl);
|
||||
java.net.HttpURLConnection connection = (java.net.HttpURLConnection) url.openConnection();
|
||||
|
||||
// 设置请求头
|
||||
connection.setRequestMethod("GET");
|
||||
connection.setRequestProperty("User-Agent", "Mozilla/5.0 (Windows NT 10.0; Win64; x64) AppleWebKit/537.36");
|
||||
connection.setRequestProperty("Accept", "application/json");
|
||||
connection.setConnectTimeout(30000); // 30秒连接超时
|
||||
connection.setReadTimeout(60000); // 60秒读取超时
|
||||
|
||||
int responseCode = connection.getResponseCode();
|
||||
if (responseCode == 200) {
|
||||
// 读取响应内容
|
||||
StringBuilder result = new StringBuilder();
|
||||
try (java.io.BufferedReader reader = new java.io.BufferedReader(
|
||||
new java.io.InputStreamReader(connection.getInputStream(), java.nio.charset.StandardCharsets.UTF_8))) {
|
||||
String line;
|
||||
while ((line = reader.readLine()) != null) {
|
||||
result.append(line).append("\n");
|
||||
}
|
||||
}
|
||||
|
||||
String responseBody = result.toString().trim();
|
||||
log.info("成功下载transcription结果，长度: {} 字符", responseBody.length());
|
||||
return responseBody;
|
||||
} else {
|
||||
log.warn("下载transcription结果失败，状态码: {}", responseCode);
|
||||
return null;
|
||||
}
|
||||
} catch (Exception e) {
|
||||
log.error("下载transcription结果异常: {}", e.getMessage(), e);
|
||||
return null;
|
||||
}
|
||||
}
|
||||
|
||||
/**
|
||||
* 解析transcription结果JSON，提取识别文本
|
||||
*/
|
||||
private String parseTranscriptionResult(String transcriptionJson) {
|
||||
try {
|
||||
// 使用Jackson解析JSON
|
||||
ObjectMapper mapper = new ObjectMapper();
|
||||
JsonNode rootNode = mapper.readTree(transcriptionJson);
|
||||
|
||||
log.info("开始解析transcription JSON");
|
||||
|
||||
// 根据阿里云ASR返回的JSON结构解析
|
||||
// 阿里云返回的结构: {"transcripts": [{"text": "识别的文本", "sentences": [...]}]}
|
||||
JsonNode transcriptsNode = rootNode.get("transcripts");
|
||||
if (transcriptsNode != null && transcriptsNode.isArray() && transcriptsNode.size() > 0) {
|
||||
JsonNode firstTranscript = transcriptsNode.get(0);
|
||||
JsonNode textNode = firstTranscript.get("text");
|
||||
if (textNode != null && !textNode.isNull()) {
|
||||
String text = textNode.asText();
|
||||
log.info("成功提取识别文本: {}", text);
|
||||
return text;
|
||||
}
|
||||
}
|
||||
|
||||
// 如果没有transcripts字段，尝试其他可能的字段结构
|
||||
JsonNode resultsNode = rootNode.get("results");
|
||||
if (resultsNode != null && resultsNode.isArray()) {
|
||||
StringBuilder textBuilder = new StringBuilder();
|
||||
for (JsonNode resultNode : resultsNode) {
|
||||
JsonNode textNode = resultNode.get("text");
|
||||
if (textNode != null && !textNode.isNull()) {
|
||||
String text = textNode.asText();
|
||||
textBuilder.append(text).append(" ");
|
||||
log.info("提取到文本片段: {}", text);
|
||||
}
|
||||
}
|
||||
String fullText = textBuilder.toString().trim();
|
||||
if (!fullText.isEmpty()) {
|
||||
log.info("完整识别文本: {}", fullText);
|
||||
return fullText;
|
||||
}
|
||||
}
|
||||
|
||||
// 如果没有results字段，尝试直接获取text字段
|
||||
JsonNode textNode = rootNode.get("text");
|
||||
if (textNode != null && !textNode.isNull()) {
|
||||
String text = textNode.asText();
|
||||
log.info("直接获取文本: {}", text);
|
||||
return text;
|
||||
}
|
||||
|
||||
// 如果都没有，记录JSON结构用于调试
|
||||
log.warn("未找到预期的文本字段，JSON结构: {}", rootNode.toString());
|
||||
return null;
|
||||
} catch (Exception e) {
|
||||
log.error("解析transcription结果JSON失败: {}", e.getMessage(), e);
|
||||
return null;
|
||||
}
|
||||
}
|
||||
|
||||
/**
|
||||
* 保存识别文本到数据库
|
||||
*/
|
||||
private void saveRecognizedTextToDatabase(String ttsLogId, String recognizedText) {
|
||||
try {
|
||||
// 根据ID查询TTS请求日志
|
||||
TtsRequestLog ttsLog = ttsRequestLogMapper.selectById(ttsLogId);
|
||||
if (ttsLog != null) {
|
||||
// 将识别文本保存到inputText字段
|
||||
ttsLog.setInputText(recognizedText);
|
||||
ttsLog.setInputLength(recognizedText.length());
|
||||
ttsLog.setUpdateTime(LocalDateTime.now());
|
||||
|
||||
// 保存到数据库
|
||||
int updateResult = ttsRequestLogMapper.updateById(ttsLog);
|
||||
if (updateResult > 0) {
|
||||
log.info("成功保存识别文本到数据库，TTS日志ID: {}, 文本长度: {}", ttsLogId, recognizedText.length());
|
||||
} else {
|
||||
log.warn("保存识别文本到数据库失败，TTS日志ID: {}", ttsLogId);
|
||||
}
|
||||
} else {
|
||||
log.warn("未找到对应的TTS请求日志，ID: {}", ttsLogId);
|
||||
}
|
||||
} catch (Exception e) {
|
||||
log.error("保存识别文本到数据库异常: {}", e.getMessage(), e);
|
||||
}
|
||||
}
|
||||
}
|
||||
|
||||
Reference in New Issue
Block a user