小说推文AgentExample

examples/novel_to_video/actions/get_novel.py

@@ -0,0 +1,49 @@
+# -*- coding: utf-8 -*-
+# @Date    : 2024/2/5 16:28
+# @Author  : 宏伟（散人）
+# @Desc    :
+
+from MetaGPT.metagpt.actions import Action
+
+
+class Novel:
+    id = 0
+    title = ''
+    alias = ''
+    chapters = []
+
+    def __init__(self, id, title, alias, chapters):
+        self.id = id
+        self.title = title
+        self.alias = alias
+        self.chapters = chapters
+
+
+class Chapter:
+    title = ''
+    content = ''
+
+    def __init__(self, title, content):
+        self.title = title
+        self.content = content
+
+def GetNovelById(nid):
+    # return Novel.query.get(nid)  #此处可以对接到数据库
+    print('nid:', nid)
+    chapter1 = Chapter(title='Chapter 1', content='This is the content of Chapter 1.')
+    chapter2 = Chapter(title='Chapter 2', content='This is the content of Chapter 2.')
+    novel = Novel(id=1, title='Novel 1', alias='N1', chapters=[chapter1, chapter2])
+    return novel
+
+
+class GetNovel(Action):
+    def __init__(self, name: str = "", *args, **kwargs):
+        super().__init__(**kwargs)
+
+    async def run(self, novel_id: str = '', *args, **kwargs) -> Novel:
+        novel = GetNovelById(novel_id)
+        return novel
+
+
+
+

examples/novel_to_video/actions/make_video.py

@@ -0,0 +1,116 @@
+# -*- coding: utf-8 -*-
+# @Date    : 2024/2/5 16:28
+# @Author  : 宏伟（散人）
+# @Desc    :
+
+import asyncio
+import os
+
+import cv2
+from moviepy.audio.AudioClip import concatenate_audioclips
+from moviepy.audio.io.AudioFileClip import AudioFileClip
+from moviepy.video.io.VideoFileClip import VideoFileClip
+from metagpt.actions import Action
+
+from MetaGPT.examples.novel_to_video.roles.ai_novel_video import WORKSPACE_DIR
+
+
+class MakeVideo(Action):
+    # 定义工作文件夹路径
+    workspace: str = ""
+
+    def __init__(self, name: str = "", workspace: str = '', *args, **kwargs):
+        super().__init__(**kwargs)
+        self.workspace = workspace
+
+    async def run(self, *args, **kwargs) -> str:
+        image_folder = self.workspace+'/image'
+        audio_folder = self.workspace+'/tts'
+        # 获取图片和音频文件的名称
+        # print('image_folder:', image_folder)
+        # print('audio_folder:', audio_folder)
+        image_names = os.listdir(image_folder)
+        audio_names = os.listdir(audio_folder)
+        output_file = self.workspace + '/output_temp.mp4'
+        output_file_final = self.workspace + '/output_final.mp4'
+        # 初始化视频写入器
+        frame_width = 768
+        frame_height = 512
+
+        fourcc = cv2.VideoWriter_fourcc(*'mp4v')
+        output_video = cv2.VideoWriter(output_file, fourcc, 30, (frame_width, frame_height), True)
+
+        for i in range(len(audio_names)):
+            print(f'Processing {i}/{len(audio_names)}')
+            img_path = os.path.join(image_folder, image_names[i])
+            audio_path = os.path.join(audio_folder, audio_names[i])
+            print('img_path:', img_path)
+            print('audio_path:', audio_path)
+
+            # 读取图片和音频
+            audio = AudioFileClip(audio_path)
+            audio_duration = audio.duration  # 音频的时长（秒）
+            audio_fps = audio.fps  # 音频的采样率（Hz）
+            # print('audio_duration:', audio_duration)
+            # print('audio_fps:', audio_fps)  # 一般来说，音频的采样率是44100Hz，但可能因文件而异
+
+            num_frames = int(audio_duration * 30)
+
+            for _ in range(num_frames):
+                img = cv2.imread(img_path)
+                output_video.write(img)
+
+        # 释放写入器
+        output_video.release()
+
+        # 写入音频
+        audio_clips = []
+        previous_end_time = 0
+        video_clip = VideoFileClip(output_file)
+        for i in range(len(audio_names)):
+            # print(f'Processing {i}/{len(audio_names)}')
+            audio_path = os.path.join(audio_folder, audio_names[i])
+            # print('audio_path:', audio_path)
+
+            # 读取图片和音频
+            audio_clip = AudioFileClip(audio_path)
+            audio_duration = audio_clip.duration  # 音频的时长（秒）
+            audio_fps = audio_clip.fps  # 音频的采样率（Hz）
+            # print('audio_duration:', audio_duration)
+            # print('audio_fps:', audio_fps)  # 一般来说，音频的采样率是44100Hz，但可能因文件而异
+
+            if i == 0:
+                audio_clip = audio_clip.set_start(0)
+            else:
+                audio_clip = audio_clip.set_start(previous_end_time)
+
+            previous_end_time += audio_duration
+
+            # 将音频片段存储在列表中
+            audio_clips.append(audio_clip)
+
+        # 将音频合成到视频
+        # 使用 concatenate_audio clips 方法连接音频片段
+        final_audio = concatenate_audioclips(audio_clips)
+
+        # 将最终音频附加到视频
+        video_clip = video_clip.set_audio(final_audio)
+
+        # 保存输出视频
+        video_clip.write_videofile(output_file_final, codec='libx264', audio_codec='aac')
+
+        # 删除临时文件
+        if os.path.exists(output_file):
+            os.remove(output_file)
+
+        return output_file_final
+
+
+async def main():
+    action = MakeVideo(workspace=WORKSPACE_DIR)
+    res = await action.run()
+    print(res)
+
+
+if __name__ == '__main__':
+    asyncio.run(main())