diff --git a/dsLightRag/B1_Mp4ToWav.py b/dsLightRag/B1_Mp4ToWav.py
new file mode 100644
index 00000000..50b1e30b
--- /dev/null
+++ b/dsLightRag/B1_Mp4ToWav.py
@@ -0,0 +1,49 @@
+import subprocess
+from pathlib import Path
+
+
+def convert_mp4_to_wav(
+        input_mp4: str,
+        ffmpeg_path: str = r"d:\ffmpeg\ffmpeg.exe",
+        output_path: str = None  # 新增输出路径参数
+):
+    """
+    将指定MP4文件转换为WAV格式
+    参数：
+    input_mp4 - 输入的MP4文件路径
+    ffmpeg_path - FFmpeg可执行文件路径
+    output_path - 可选输出路径（可指定完整路径或目录）
+    """
+    input_path = Path(input_mp4)
+
+    # 处理输出路径逻辑
+    if output_path:
+        output_path = Path(output_path)
+        if output_path.is_dir():  # 如果传入的是目录
+            output_path = output_path / f"{input_path.stem}.wav"
+        output_path.parent.mkdir(parents=True, exist_ok=True)
+    else:
+        # 默认输出到wav目录
+        output_dir = Path("wav")
+        output_dir.mkdir(exist_ok=True)
+        output_path = output_dir / f"{input_path.stem}.wav"
+
+    cmd = [
+        ffmpeg_path,
+        "-i", str(input_path),
+        "-acodec", "pcm_s16le",
+        "-ac", "1",
+        "-ar", "16000",
+        "-y",
+        str(output_path)  # 使用处理后的输出路径
+    ]
+
+    try:
+        # 执行转换命令
+        subprocess.run(cmd, check=True, capture_output=True)
+        print(f"转换成功：{input_path} -> {output_path}")
+    except subprocess.CalledProcessError as e:
+        print(f"转换失败：{e.stderr.decode('gbk')}")
+    except FileNotFoundError:
+        print(f"FFmpeg未找到，请确认路径是否正确：{ffmpeg_path}")
+
diff --git a/dsLightRag/B1_Start.py b/dsLightRag/B1_Start.py
new file mode 100644
index 00000000..6a35f918
--- /dev/null
+++ b/dsLightRag/B1_Start.py
@@ -0,0 +1,21 @@
+from B1_Mp4ToWav import *
+from B2_WavToText import *
+from B3_TextSummarize import *
+
+if __name__ == '__main__':
+    # 1、根据云校的视频课程下载视频文件
+
+    # 2、调用ffmpeg将视频文件转成wav文件
+    # convert_mp4_to_wav(r"D:\backup\七年级第三单元复习课.mp4",
+    #                   output_path=r"D:\backup\123.wav")
+    # 3、将转换完成的WAV上传到阿里云
+
+    # 4、调用阿里云的语音识别API，将wav文件转成文字
+    #audio_url = "https://ylt.oss-cn-hangzhou.aliyuncs.com/HuangHai/123.wav"
+    #ShiBie(audio_url, Path("识别结果.txt"))
+    # 2.5元一小时，本音频文件45分钟，约2元
+
+    # 6、生成总结
+    input_file = Path(r"D:\dsWork\QingLong\AI\音频文本.txt")
+    output_file = Path(r"D:\dsWork\QingLong\AI\分析结果.txt")
+    analyzer_action(input_file, output_file)
diff --git a/dsLightRag/B2_WavToText.py b/dsLightRag/B2_WavToText.py
new file mode 100644
index 00000000..b69f0f16
--- /dev/null
+++ b/dsLightRag/B2_WavToText.py
@@ -0,0 +1,189 @@
+# -*- coding: utf-8 -*-
+import json
+import time
+from typing import Dict, Tuple
+from pathlib import Path
+from aliyunsdkcore.acs_exception.exceptions import ClientException, ServerException
+from aliyunsdkcore.client import AcsClient
+from aliyunsdkcore.request import CommonRequest
+from Config import *
+from Config.Config import ALY_AK, ALY_SK
+
+# 服务常量配置
+CONFIG = {
+    "REGION_ID": "cn-shanghai",
+    "PRODUCT": "nls-filetrans",
+    "DOMAIN": "filetrans.cn-shanghai.aliyuncs.com",
+    "API_VERSION": "2018-08-17",
+    "MAX_RETRIES": 20,
+    "POLL_INTERVAL": 10
+}
+
+
+class TranscriptionError(Exception):
+    """自定义语音识别异常基类"""
+    pass
+
+
+class APIConnectionError(TranscriptionError):
+    """API连接异常"""
+    pass
+
+
+class TaskSubmissionError(TranscriptionError):
+    """任务提交异常"""
+    pass
+
+
+class TaskTimeoutError(TranscriptionError):
+    """任务超时异常"""
+    pass
+
+
+def submit_transcription_task(client: AcsClient, app_key: str, file_link: str) -> str:
+    """提交语音识别任务
+
+    :param client: 阿里云客户端实例
+    :param app_key: 应用密钥
+    :param file_link: 音频文件URL
+    :return: 任务ID
+    :raises TaskSubmissionError: 任务提交失败时抛出
+    """
+    task_config = {
+        "appkey": app_key,
+        "file_link": file_link,
+        "version": "4.0",
+        "enable_words": False
+    }
+
+    request = CommonRequest()
+    request.set_domain(CONFIG["DOMAIN"])
+    request.set_version(CONFIG["API_VERSION"])
+    request.set_product(CONFIG["PRODUCT"])
+    request.set_action_name("SubmitTask")
+    request.set_method('POST')
+    request.add_body_params("Task", json.dumps(task_config))
+
+    try:
+        response = client.do_action_with_exception(request)
+        response_data = json.loads(response)
+
+        if response_data.get("StatusText") != "SUCCESS":
+            raise TaskSubmissionError(f"任务提交失败: {response_data.get('StatusText', '未知错误')}")
+
+        return response_data["TaskId"]
+
+    except (ServerException, ClientException) as e:
+        raise APIConnectionError(f"API连接异常: {str(e)}") from e
+    except KeyError as e:
+        raise TaskSubmissionError("响应缺少TaskId字段") from e
+
+
+def poll_transcription_result(client: AcsClient, task_id: str) -> Dict:
+    """轮询识别任务结果
+
+    :param client: 阿里云客户端实例
+    :param task_id: 任务ID
+    :return: 识别结果字典
+    :raises TaskTimeoutError: 超时未完成时抛出
+    """
+    request = CommonRequest()
+    request.set_domain(CONFIG["DOMAIN"])
+    request.set_version(CONFIG["API_VERSION"])
+    request.set_product(CONFIG["PRODUCT"])
+    request.set_action_name("GetTaskResult")
+    request.set_method('GET')
+    request.add_query_param("TaskId", task_id)
+
+    retries = 0
+    while retries < CONFIG["MAX_RETRIES"]:
+        try:
+            response = client.do_action_with_exception(request)
+            result = json.loads(response)
+            status = result.get("StatusText", "")
+
+            if status == "SUCCESS":
+                return result.get("Result", {})
+            if status in ("RUNNING", "QUEUEING"):
+                time.sleep(CONFIG["POLL_INTERVAL"])
+                retries += 1
+            else:
+                raise TaskSubmissionError(f"识别失败，状态: {status}")
+
+        except (ServerException, ClientException) as e:
+            raise APIConnectionError(f"查询异常: {str(e)}") from e
+
+    raise TaskTimeoutError(f"超过最大重试次数({CONFIG['MAX_RETRIES']})，任务未完成")
+
+
+def save_transcription_result(result: Dict, output_path: Path) -> Path:
+    """保存识别结果到文件
+
+    :param result: 识别结果字典
+    :param output_path: 输出文件路径
+    :return: 实际保存路径
+    :raises IOError: 文件保存失败时抛出
+    """
+    if not result.get('Sentences'):
+        raise ValueError("识别结果为空")
+
+    try:
+        output_path.parent.mkdir(parents=True, exist_ok=True)
+        output_path.write_text(str(result['Sentences']), encoding='utf-8')
+        return output_path.resolve()
+    except (IOError, PermissionError) as e:
+        raise IOError(f"文件保存失败: {str(e)}") from e
+
+
+def transcribe_audio_file(
+        access_key_id: str,
+        access_key_secret: str,
+        app_key: str,
+        audio_url: str,
+        output_path: Path = Path("识别结果.txt")
+) -> Tuple[bool, Path, str]:
+    """语音识别主流程
+
+    :param access_key_id: 阿里云访问密钥ID
+    :param access_key_secret: 阿里云访问密钥
+    :param app_key: 应用密钥
+    :param audio_url: 音频文件URL
+    :param output_path: 输出文件路径
+    :return: (成功标志, 输出路径, 错误信息)
+    """
+    client = AcsClient(access_key_id, access_key_secret, CONFIG["REGION_ID"])
+
+    try:
+        task_id = submit_transcription_task(client, app_key, audio_url)
+        print(f"任务已提交，ID: {task_id}")
+
+        result = poll_transcription_result(client, task_id)
+        saved_path = save_transcription_result(result, output_path)
+
+        return True, saved_path, ""
+
+    except TranscriptionError as e:
+        return False, Path(), str(e)
+    except Exception as e:
+        return False, Path(), f"未处理的异常: {str(e)}"
+
+def ShiBie(audio_url, output_path):
+    # 配置参数（应通过环境变量或配置文件获取）
+    CREDENTIALS = {
+        "access_key_id": ALY_AK,
+        "access_key_secret": ALY_SK,
+        "app_key": "OIpiw501l4o6MYEe",
+        "audio_url":audio_url,
+        "output_path": output_path
+    }
+
+    # 执行识别
+    success, path, error = transcribe_audio_file(**CREDENTIALS)
+
+    if success:
+        print(f"✅ 识别成功，文件已保存至: {path}")
+    else:
+        print(f"❌ 识别失败: {error}")
+        exit(1)
+
+
diff --git a/dsLightRag/B3_TextSummarize.py b/dsLightRag/B3_TextSummarize.py
new file mode 100644
index 00000000..67213bac
--- /dev/null
+++ b/dsLightRag/B3_TextSummarize.py
@@ -0,0 +1,173 @@
+# -*- coding: utf-8 -*-
+from typing import Optional, Tuple, Iterator
+from openai import OpenAI, APIError, APITimeoutError
+import time
+import httpx
+from pathlib import Path
+from Config.Config import LLM_API_KEY, LLM_BASE_URL, LLM_MODEL_NAME
+
+
+class ContentAnalyzer:
+    """课程内容分析器（流式版本）"""
+
+    def __init__(
+            self,
+            api_key: str = LLM_API_KEY,
+            base_url: str = LLM_BASE_URL,
+            model: str = LLM_MODEL_NAME,
+            max_retries: int = 10,
+            initial_timeout: int = 300
+    ):
+        self._show_progress("🔧", "初始化分析器...", level=0)
+        self.client = OpenAI(api_key=api_key, base_url=base_url)
+        self.model = model
+        self.max_retries = max_retries
+        self.initial_timeout = initial_timeout
+        self._check_network()
+        self._show_progress("✅", "分析器准备就绪", level=0)
+
+    def _show_progress(self, emoji: str, message: str, level: int = 1):
+        indent = "  " * level
+        timestamp = time.strftime("%H:%M:%S")
+        print(f"{indent}{emoji} [{timestamp}] {message}")
+
+    def _check_network(self):
+        try:
+            with httpx.Client(timeout=30) as client:
+                client.get("https://dashscope.aliyuncs.com")
+            self._show_progress("🌐", "网络连接正常", level=1)
+        except Exception as e:
+            self._show_progress("❌", f"网络异常: {str(e)}", level=1)
+            raise
+
+    def _retry_delay(self, attempt: int) -> int:
+        """指数退避延迟"""
+        return min(2 ** attempt, 60)  # 最大延迟60秒
+
+    def analyze_content_stream(
+            self,
+            content: str,
+            prompt_template: str = "帮我梳理：这节课分了几个部分，每部分的名称和开始的时间是多少：{}"
+    ) -> Iterator[Tuple[bool, str]]:
+        """流式分析内容"""
+        for attempt in range(self.max_retries + 1):
+            try:
+                current_timeout = self.initial_timeout + attempt * 5
+                self._show_progress("⏱️", f"尝试 {attempt + 1}/{self.max_retries} (超时: {current_timeout}s)", level=2)
+
+                full_prompt = prompt_template.format(content)
+                stream = self.client.chat.completions.create(
+                    model=self.model,
+                    messages=[{'role': 'user', 'content': full_prompt}],
+                    timeout=current_timeout,
+                    stream=True  # 启用流式模式
+                )
+
+                buffer = []
+                for chunk in stream:
+                    if chunk.choices and chunk.choices[0].delta.content:
+                        content_chunk = chunk.choices[0].delta.content
+                        buffer.append(content_chunk)
+                        yield True, content_chunk  # 实时返回每个片段
+
+                # 返回完整结果
+                if buffer:
+                    yield True, ''.join(buffer)
+                return
+
+            except APITimeoutError as e:
+                if attempt < self.max_retries:
+                    delay = self._retry_delay(attempt)
+                    self._show_progress("⏳", f"{delay}s后重试...", level=2)
+                    time.sleep(delay)
+                else:
+                    yield False, f"API请求超时，已重试{self.max_retries}次"
+                    return
+
+            except APIError as e:
+                yield False, f"API错误: {str(e)}"
+                return
+
+            except Exception as e:
+                yield False, f"未处理的异常: {str(e)}"
+                return
+
+    def analyze_file(
+            self,
+            file_path: Path,
+            output_path: Optional[Path] = None,
+            encoding: str = 'utf-8'
+    ) -> Tuple[bool, str]:
+        """处理文件全流程（流式版本）"""
+        try:
+            self._show_progress("📂", f"开始处理文件: {file_path}", level=0)
+
+            # 文件验证
+            self._show_progress("🔍", "验证文件...", level=1)
+            if not file_path.exists():
+                self._show_progress("❌", "文件不存在", level=2)
+                return False, f"文件不存在: {file_path}"
+            if file_path.stat().st_size > 10 * 1024 * 1024:
+                self._show_progress("⚠️", "注意：大文件可能影响处理速度", level=2)
+
+            # 读取内容
+            self._show_progress("📖", "读取文件内容...", level=1)
+            try:
+                content = file_path.read_text(encoding=encoding)
+            except UnicodeDecodeError:
+                self._show_progress("🔠", "解码失败，尝试GBK编码...", level=2)
+                content = file_path.read_text(encoding='gbk')
+
+            # 流式分析
+            self._show_progress("🧠", "开始流式分析...", level=1)
+            result_buffer = []
+            has_error = False
+            error_msg = ""
+
+            for status, chunk in self.analyze_content_stream(content):
+                if not status:
+                    has_error = True
+                    error_msg = chunk
+                    break
+                print(chunk, end='', flush=True)  # 实时输出
+                result_buffer.append(chunk)
+
+            if has_error:
+                self._show_progress("❌", f"分析失败: {error_msg}", level=1)
+                return False, error_msg
+
+            final_result = ''.join(result_buffer)
+
+            # 保存结果
+            if output_path:
+                self._show_progress("💾", f"保存到: {output_path}", level=1)
+                try:
+                    output_path.parent.mkdir(parents=True, exist_ok=True)
+                    output_path.write_text(final_result, encoding=encoding)
+                    self._show_progress("✅", "保存成功", level=2)
+                except Exception as e:
+                    self._show_progress("❌", f"保存失败: {str(e)}", level=2)
+                    return False, f"结果保存失败: {str(e)}"
+
+            self._show_progress("🎉", "处理完成！", level=0)
+            return True, final_result
+
+        except Exception as e:
+            self._show_progress("💣", f"严重错误: {str(e)}", level=1)
+            return False, f"文件处理失败: {str(e)}"
+
+
+def analyzer_action(input_file, output_file):
+    print("\n" + "=" * 50)
+    print("  🚀 长春云校视频课程智能打标记系统  ".center(50, "✨"))
+    print("=" * 50)
+
+    analyzer = ContentAnalyzer(initial_timeout=300)
+    success, result = analyzer.analyze_file(Path(input_file), Path(output_file))
+
+    print("\n" + "=" * 50)
+    if success:
+        print("\n✅ 分析成功！结果已保存至：", output_file)
+    else:
+        print(f"\n❌ 分析失败：{result}")
+    print("=" * 50)
\ No newline at end of file
diff --git a/dsLightRag/Config/Config.py b/dsLightRag/Config/Config.py
index 5b1a6daa..46f3b4d5 100644
--- a/dsLightRag/Config/Config.py
+++ b/dsLightRag/Config/Config.py
@@ -1,3 +1,7 @@
+# 阿里云的配置信息
+ALY_AK = 'LTAI5tE4tgpGcKWhbZg6C4bh'
+ALY_SK = 'oizcTOZ8izbGUouboC00RcmGE8vBQ1'
+
 # 大模型 【DeepSeek深度求索官方】
 # LLM_API_KEY = "sk-44ae895eeb614aa1a9c6460579e322f1"
 # LLM_BASE_URL = "https://api.deepseek.com"
diff --git a/dsLightRag/FlagSystem/视频课程打标记人工与大模型的对比.png b/dsLightRag/FlagSystem/视频课程打标记人工与大模型的对比.png
new file mode 100644
index 00000000..b6151fb3
Binary files /dev/null and b/dsLightRag/FlagSystem/视频课程打标记人工与大模型的对比.png differ
diff --git a/dsLightRag/FlagSystem/长春云校视频课程智能打标记点系统.txt b/dsLightRag/FlagSystem/长春云校视频课程智能打标记点系统.txt
new file mode 100644
index 00000000..ab61451d
--- /dev/null
+++ b/dsLightRag/FlagSystem/长春云校视频课程智能打标记点系统.txt
@@ -0,0 +1,82 @@
+# 视频课程
+# 七年级第三单元复习课
+https://yx.ccsjy.cn/ChangChunCloudSchool/index.html#/course-detail/99f50f9c61bd52f61d359e15e3903153
+
+
+D:\anaconda3\envs\py310\python.exe D:\dsWork\QingLong\AI\T3_TextSummarize.py
+
+==================================================
+✨✨✨✨✨✨✨✨✨✨✨✨✨✨  🚀 长春云校视频课程智能打标记系统  ✨✨✨✨✨✨✨✨✨✨✨✨✨✨✨
+==================================================
+🔧 [14:13:57] 初始化分析器...
+  🌐 [14:13:58] 网络连接正常
+✅ [14:13:58] 分析器准备就绪
+📂 [14:13:58] 开始处理文件: D:\dsWork\QingLong\AI\音频文本.txt
+  🔍 [14:13:58] 验证文件...
+  📖 [14:13:58] 读取文件内容...
+  🧠 [14:13:58] 开始分析...
+  ⏱️ [14:13:58] 尝试 1/10 (超时: 300s)
+    ✅ [14:16:57] 请求成功
+  💾 [14:16:57] 保存到: D:\dsWork\QingLong\AI\分析结果.txt
+    ✅ [14:16:57] 保存成功
+🎉 [14:16:57] 处理完成！
+
+==================================================
+✅ 分析成功！结果如下：
+
+以下是本节课的结构梳理，按时间顺序分为5大部分：
+
+**1. 课程导入与目标说明（8,940-73,100 ms）**
+- 单元复习导入（8,940-23,219）
+- 单元目标说明（73,100-100,259）
+  一、总结学习之道
+  二、积累成语运用
+  三、掌握默读技巧
+
+**2. 任务一：温故知新（106,400-417,500 ms）**
+- 核心内容：
+  - 四篇课文学习之道分析
+    •《从百草园到三味书屋》
+    •《往事依依》 
+    •《再塑生命的人》
+    •《论语十二章》
+  - 系统归纳表格梳理（366,160-417,500）
+
+**3. 任务二：成语积累（421,840-807,480 ms）**
+- 活动一：成语展示（475,800-606,440）
+  - 课文成语解析
+  - 论语演化成语
+- 活动二：成语运用（646,860-807,480）
+  - 语段填空练习
+  - 片段写作训练
+
+**4. 任务三：阅读方法（936,420-1,655,800 ms）**
+- 默读技巧讲解（936,420-1,177,420）
+  - 三到原则（眼到/心到/手到）
+  - 关键信息捕捉
+- 朗读vs默读对比（1,188,780-1,405,180）
+  - 特点/作用/适用场景差异
+- 综合训练（1,407,180-1,655,800）
+  - 文章分析《读书声最美》
+  - 阅读方法实践
+
+**5. 总结与作业（1,625,580-1,959,179 ms）**
+- 单元要点回顾（1,625,580-1,656,080）
+- 拓展作业布置（1,656,166-1,959,179）
+  - 阅读《悬崖边的树》
+  - 总结学习之道
+
+**时间轴完整结构：**
+
+| 部分名称               | 开始时间（ms） | 主要内容                                                                 |
+|------------------------|----------------|--------------------------------------------------------------------------|
+| 课程导入与目标说明     | 8,940          | 单元概述、教学目标设定                                                   |
+| 任务一：温故知新       | 106,400        | 四篇课文学习之道深度解析                                                 |
+| 任务二：成语积累       | 421,840        | 成语知识系统梳理与运用实践                                               |
+| 任务三：阅读方法       | 936,420        | 默读技巧训练/朗读与默读对比分析                                          |
+| 总结与作业             | 1,625,580      | 单元知识结构化总结/拓展阅读《悬崖边的树》                                |
+
+注：时间单位为毫秒（ms），完整课程时长约32分钟（1,959,179 ms ≈ 1,959秒）
+==================================================
+
+进程已结束，退出代码为 0