讯飞星火模型-语音转文字实现

news2024/10/7 18:25:08

项目结构

准备音频

接口Demo

准备代码（完整修改后）

测试提取中文文字代码

结果

下载链接：

这是上周打算试试，提取视频文字之后，制作视频字幕，从而想用大模型来实现，基本的demo可以在讯飞对应官网找到，之后就是api对接即可

项目结构

准备音频

接口Demo

示例demo请点击这里下载。

准备代码（完整修改后）

Ifasrdemo.java


package cn.xfyun;

import cn.hutool.json.JSONUtil;
import cn.xfyun.sign.LfasrSignature;
import cn.xfyun.utils.HttpUtil;
import com.google.gson.Gson;
import org.apache.commons.lang.StringEscapeUtils;

import java.io.*;
import java.net.URISyntaxException;
import java.nio.charset.StandardCharsets;
import java.nio.file.Files;
import java.nio.file.Paths;
import java.security.SignatureException;
import java.util.HashMap;


public class Ifasrdemo {
    private static final String HOST = "https://raasr.xfyun.cn";
    private static String AUDIO_FILE_PATH;
    private static final String appid = "xxxxx";
    private static final String keySecret = "xxxxxxxx";

    private static final Gson gson = new Gson();

    static {
        try {
            AUDIO_FILE_PATH = Ifasrdemo.class.getResource("/").toURI().getPath() + "/audio/合成音频.wav";
        } catch (URISyntaxException e) {
            e.printStackTrace();
        }
    }


    public static void main(String[] args) throws Exception {
        String result = upload();
        String jsonStr = StringEscapeUtils.unescapeJavaScript(result);
        String orderId = String.valueOf(JSONUtil.getByPath(JSONUtil.parse(jsonStr), "content.orderId"));
        getResult(orderId);



    }

    private static String upload() throws SignatureException, FileNotFoundException {
        HashMap<String, Object> map = new HashMap<>(16);
        File audio = new File(AUDIO_FILE_PATH);
        String fileName = audio.getName();
        long fileSize = audio.length();
        map.put("appId", appid);
        map.put("fileSize", fileSize);
        map.put("fileName", fileName);
        map.put("duration", "200");
        LfasrSignature lfasrSignature = new LfasrSignature(appid, keySecret);
        map.put("signa", lfasrSignature.getSigna());
        map.put("ts", lfasrSignature.getTs());

        String paramString = HttpUtil.parseMapToPathParam(map);
        System.out.println("upload paramString:" + paramString);

        String url = HOST + "/v2/api/upload" + "?" + paramString;
        System.out.println("upload_url:" + url);
        String response = HttpUtil.iflyrecUpload(url, new FileInputStream(audio));

        System.out.println("upload response:" + response);
        return response;
    }

    private static String getResult(String orderId) throws SignatureException, InterruptedException, IOException {
        HashMap<String, Object> map = new HashMap<>(16);
        map.put("orderId", orderId);
        LfasrSignature lfasrSignature = new LfasrSignature(appid, keySecret);
        map.put("signa", lfasrSignature.getSigna());
        map.put("ts", lfasrSignature.getTs());
        map.put("appId", appid);
        map.put("resultType", "transfer,predict");
        String paramString = HttpUtil.parseMapToPathParam(map);
        String url = HOST + "/v2/api/getResult" + "?" + paramString;
        System.out.println("\nget_result_url:" + url);
        while (true) {
            String response = HttpUtil.iflyrecGet(url);
            JsonParse jsonParse = gson.fromJson(response, JsonParse.class);
            if (jsonParse.content.orderInfo.status == 4 || jsonParse.content.orderInfo.status == -1) {
                System.out.println("订单完成:" + response);
                write(response);
                return response;
            } else {
                System.out.println("进行中...，状态为:" + jsonParse.content.orderInfo.status);
                //建议使用回调的方式查询结果，查询接口有请求频率限制
                Thread.sleep(7000);
            }
        }
    }

   
    public static void write(String resp) throws IOException {
        // 使用 try-with-resources 确保资源被正确关闭
        try (BufferedWriter bw = new BufferedWriter(new OutputStreamWriter(
                new FileOutputStream("src/main/resources/output/test.txt"), "UTF-8"))) {
            String ss = resp;
            // 用 UTF-8 写入
            bw.write(ss);
            System.out.println("写入txt成功");
        } catch (IOException e) {
            e.printStackTrace();
        }
    }




    class JsonParse {
        Content content;
    }

    class Content {
        OrderInfo orderInfo;
    }

    class OrderInfo {
        Integer status;
    }
}

测试提取中文文字代码

ChineseTextExtractor.java

package cn.xfyun;

import java.io.*;
import java.util.regex.Matcher;
import java.util.regex.Pattern;

public class ChineseTextExtractor {

    public static void main(String[] args) {
        String inputFilePath = "src/main/resources/output/test.txt"; // 输入文件路径
        String outputFilePath = "src/main/resources/output/out.txt"; // 输出文件路径
        extractAndWriteChineseText(inputFilePath, outputFilePath);
    }

    public static void extractAndWriteChineseText(String inputFilePath, String outputFilePath) {
        Pattern pattern = Pattern.compile("[\\u4e00-\\u9fa5]+"); // 中文字符的正则表达式
        try (BufferedReader br = new BufferedReader(new FileReader(inputFilePath));
             BufferedWriter bw = new BufferedWriter(new FileWriter(outputFilePath))) {

            String line;
            while ((line = br.readLine()) != null) {
                Matcher matcher = pattern.matcher(line);
                while (matcher.find()) {
                    bw.write(matcher.group()); // 写入匹配的中文字符
                    bw.newLine(); // 每个匹配项后面添加一个新行
                }
            }
            System.out.println("中文字符提取完成，已写入 " + outputFilePath);
        } catch (IOException e) {
            e.printStackTrace();
        }
    }
}