从URL到视频：用Python和AI构建自动化内容讲解视频生成管道

# requirements: requests, beautifulsoup4, readability-lxml
import requests
from readability import Document

def extract_text_from_url(url: str) -> (str, str):
    """从URL提取文章标题和纯文本内容"""
    try:
        response = requests.get(url, headers={'User-Agent': 'Mozilla/5.0'})
        response.raise_for_status()
        
        doc = Document(response.content)
        title = doc.title()
        content_html = doc.summary()
        
        # 使用BeautifulSoup从HTML中提取纯文本
        from bs4 import BeautifulSoup
        soup = BeautifulSoup(content_html, 'lxml')
        content_text = soup.get_text(separator='\n', strip=True)
        
        return title, content_text
    except requests.RequestException as e:
        print(f"Error fetching URL {url}: {e}")
        return None, None

# # 示例
# url = "一篇公开的飞书文档或博客文章链接"
# title, text = extract_text_from_url(url)
# if text:
#     print(f"Title: {title}")
#     print(f"Content length: {len(text)}")

# 角色
你是一位专业的视频脚本编剧和课程设计师。

# 任务
我将提供一段长文本，请你将其改编成一个结构化的视频讲解脚本。脚本应该被分解为多个“幻灯片”(slides)。

# 输入文本
"""
{{long_text}}
"""

# 输出要求
请严格按照以下JSON格式输出，不要有任何额外解释。每一张幻灯片包含一个标题(title)和要讲解的核心内容(content)，内容应简洁明了，适合口语化讲解。
[
  {
    "slide_number": 1,
    "title": "幻灯片1的标题",
    "content": "这是第一张幻灯片要讲解的具体内容..."
  },
  {
    "slide_number": 2,
    "title": "幻灯片2的标题",
    "content": "这是第二张幻灯片要讲解的具体内容..."
  },
  // ... 更多幻灯片
]

# 假设使用一个TTS库或API
def generate_audio_for_slide(slide_text: str, output_path: str):
    # 此处为调用具体TTS服务的代码，例如：
    # response = openai.audio.speech.create(
    #     model="tts-1",
    #     voice="alloy",
    #     input=slide_text
    # )
    # response.stream_to_file(output_path)
    print(f"Generated audio at: {output_path}")
    # 注意：实际应用中需要处理API调用、认证和错误
    pass

# requirements: Pillow
from PIL import Image, ImageDraw, ImageFont

def create_slide_image(title: str, content: str, output_path: str, width=1920, height=1080):
    """使用Pillow创建一张包含标题和内容的简单幻灯片图片"""
    img = Image.new('RGB', (width, height), color = (25, 25, 112)) # 深蓝色背景
    draw = ImageDraw.Draw(img)

    # 简化处理，实际应用中字体路径和大小需要更精细的控制
    try:
        title_font = ImageFont.truetype("msyh.ttc", 80) # 微软雅黑
        content_font = ImageFont.truetype("msyh.ttc", 50)
    except IOError:
        title_font = ImageFont.load_default()
        content_font = ImageFont.load_default()

    # 绘制标题
    draw.text((width/2, height*0.2), title, font=title_font, fill=(255, 255, 255), anchor="ms")
    
    # 绘制内容 (简化处理，未做自动换行)
    draw.text((width/2, height*0.4), content, font=content_font, fill=(255, 255, 255), anchor="ma")

    img.save(output_path)
    print(f"Generated image at: {output_path}")

# requirements: moviepy
from moviepy.editor import ImageClip, AudioFileClip, concatenate_videoclips

def synthesize_video(slides_data: list, audio_paths: list, image_paths: list, output_path: str):
    """
    将图片序列和音频序列合成为视频
    :param slides_data: 包含脚本信息的列表
    :param audio_paths: 每个slide对应音频文件的路径列表
    :param image_paths: 每个slide对应图片文件的路径列表
    :param output_path: 最终视频的输出路径
    """
    clips = []
    for i, slide in enumerate(slides_data):
        try:
            audio_clip = AudioFileClip(audio_paths[i])
            # 视频片段的长度由音频的长度决定
            image_clip = ImageClip(image_paths[i]).set_duration(audio_clip.duration)
            # 将音频附加到视频片段上
            video_clip = image_clip.set_audio(audio_clip)
            clips.append(video_clip)
        except Exception as e:
            print(f"Error processing slide {i+1}: {e}")
            continue

    if not clips:
        print("No valid clips to concatenate.")
        return

    final_video = concatenate_videoclips(clips, method="compose")
    final_video.write_videofile(output_path, fps=24, codec="libx264")
    print(f"Final video saved to: {output_path}")