Popiai-skill/social-media/hilo-promo/scripts/tutorial_synthesis.py

#!/usr/bin/env python3
"""
教程合成脚本 - 制作教程段素材（高亮框 + 箭头 + 上传动画 + 预拼接）

使用方式：
    # 1. 生成高亮框教程截图（tut_002）
    python tutorial_synthesis.py make-highlight \\
        --template-screenshot tut_002_raw.jpg \\
        --output tut_002.png \\
        --box-x 100 --box-y 200 --box-w 300 --box-h 150

    # 2. 生成上传动画（tut_003 → tut_004）
    python tutorial_synthesis.py upload-anim \\
        --before tut_003.png --after tut_004.png \\
        --output tut_upload_anim.mp4 \\
        --duration 2.5

    # 3. 预拼接教程段（tut_001 + tut_002 + 上传动画）
    python tutorial_synthesis.py combine \\
        --segments tut1_norm.mp4 tut_002_norm.jpg tut_upload_anim.mp4 \\
        --output tut_combined.mp4

    # 4. 完整流水线（一次性执行上述所有步骤）
    python tutorial_synthesis.py pipeline \\
        --tut001 tut_001.mp4 \\
        --tut002-raw tut_002_raw.jpg \\
        --tut003 tut_003.png \\
        --tut004 tut_004.png \\
        --output-dir ./
"""

import argparse
import subprocess
import sys
from pathlib import Path
from typing import Any, List, Optional

from PIL import Image, ImageDraw, ImageFilter


# ─────────────────────────── 配置 ───────────────────────────

CANVAS_WIDTH = 1080
CANVAS_HEIGHT = 1920
TARGET_FPS = 30

# 高亮框参数
HIGHLIGHT_FILL = (255, 100, 150, 60)  # 半透明粉色
HIGHLIGHT_OUTLINE = (255, 80, 130, 220)  # 深粉色描边
HIGHLIGHT_LINE_WIDTH = 4


# ─────────────────────────── 工具函数 ───────────────────────────


def rounded_rectangle(
    draw: Any,
    xy: tuple,
    radius: int = 20,
    fill: Optional[tuple] = None,
    outline: Optional[tuple] = None,
    width: int = 1,
) -> None:
    """绘制圆角矩形"""
    x1, y1, x2, y2 = xy

    # 画圆弧
    draw.arc(
        [x1, y1, x1 + radius * 2, y1 + radius * 2], 180, 270, fill or outline, width
    )
    draw.arc(
        [x2 - radius * 2, y1, x2, y1 + radius * 2], 270, 360, fill or outline, width
    )
    draw.arc(
        [x1, y2 - radius * 2, x1 + radius * 2, y2], 90, 180, fill or outline, width
    )
    draw.arc([x2 - radius * 2, y2 - radius * 2, x2, y2], 0, 90, fill or outline, width)

    # 画矩形主体
    draw.rectangle([x1 + radius, y1, x2 - radius, y2], fill=fill)
    draw.rectangle([x1, y1 + radius, x2, y2 - radius], fill=fill)

    # 画边框
    draw.rectangle([x1, y1 + radius, x1 + width, y2 - radius], fill=outline)
    draw.rectangle([x2 - width, y1 + radius, x2, y2 - radius], fill=outline)
    draw.rectangle([x1 + radius, y1, x2 - radius, y1 + width], fill=outline)
    draw.rectangle([x1 + radius, y2 - width, x2 - radius, y2], fill=outline)


def draw_arrow(
    draw: Any,
    x: int,
    y: int,
    size: int = 30,
    angle: float = 0,
    color: tuple = (255, 80, 130),
    width: int = 4,
) -> None:
    """绘制手绘风格箭头（简化版：向下箭头）"""
    # 向下箭头
    cx, cy = x, y
    tip = (cx, cy + size)
    left = (cx - size // 2, cy)
    right = (cx + size // 2, cy)
    draw.line([tip, left], fill=color, width=width)
    draw.line([tip, right], fill=color, width=width)
    draw.line([left, (left[0], left[1] - size // 2)], fill=color, width=width)
    draw.line([right, (right[0], right[1] - size // 2)], fill=color, width=width)


def ensure_dir(path: Path) -> None:
    path.parent.mkdir(parents=True, exist_ok=True)


def run_ffmpeg(cmd: List[str], check: bool = True) -> None:
    """执行 ffmpeg 命令"""
    print(f"🔧 {' '.join(cmd[:6])}...")
    result = subprocess.run(cmd, capture_output=True, text=True)
    if result.returncode != 0:
        print(f"❌ ffmpeg 错误：\n{result.stderr[-500:]}")
        if check:
            sys.exit(1)


# ─────────────────────────── 子命令实现 ───────────────────────────


def cmd_make_highlight(args: argparse.Namespace) -> None:
    """生成带高亮框的教程截图"""
    print(f"🎨 生成高亮框教程截图...")

    # 加载原图
    img = Image.open(args.template_screenshot).convert("RGBA")

    # 缩放到目标分辨率（如果需要）
    LANCZOS = getattr(Image, "LANCZOS", Image.BICUBIC)
    if img.size != (CANVAS_WIDTH, CANVAS_HEIGHT):
        img = img.resize((CANVAS_WIDTH, CANVAS_HEIGHT), LANCZOS)

    draw = ImageDraw.Draw(img)

    # 绘制圆角高亮框
    box = (args.box_x, args.box_y, args.box_x + args.box_w, args.box_y + args.box_h)
    rounded_rectangle(
        draw,
        box,
        radius=15,
        fill=HIGHLIGHT_FILL,
        outline=HIGHLIGHT_OUTLINE,
        width=HIGHLIGHT_LINE_WIDTH,
    )

    # 绘制箭头（指向高亮区域）
    arrow_x = args.box_x + args.box_w // 2
    arrow_y = args.box_y + args.box_h + 20
    draw_arrow(draw, arrow_x, arrow_y, size=25, color=HIGHLIGHT_OUTLINE, width=3)

    # 保存
    ensure_dir(Path(args.output))
    img.save(args.output, "PNG")
    print(f"✅ 输出：{args.output}")


def cmd_upload_anim(args: argparse.Namespace) -> None:
    """生成上传动画（xfade 过渡）"""
    print(f"🎬 生成上传动画...")

    output = Path(args.output)
    ensure_dir(output)

    # xfade 参数
    duration = args.duration
    transition = args.transition  # slideup / slideleft / fadeblack
    xfade_dur = duration * 0.3  # 转场占 30%
    offset = duration - xfade_dur  # 第二张图出现时间

    cmd = [
        "ffmpeg",
        "-y",
        "-loop",
        "1",
        "-t",
        str(duration),
        "-i",
        str(args.before),
        "-loop",
        "1",
        "-t",
        str(duration),
        "-i",
        str(args.after),
        "-filter_complex",
        f"[0:v]fps={TARGET_FPS},format=yuv420p[v0];"
        f"[1:v]fps={TARGET_FPS},format=yuv420p[v1];"
        f"[v0][v1]xfade=transition={transition}:duration={xfade_dur}:offset={offset}",
        "-c:v",
        "libx264",
        "-pix_fmt",
        "yuv420p",
        "-t",
        str(duration),
        str(output),
    ]
    run_ffmpeg(cmd)
    print(f"✅ 输出：{output}")


def cmd_combine(args: argparse.Namespace) -> None:
    """预拼接教程段（统一编码 + concat）"""
    print(f"🔗 预拼接教程段...")

    output = Path(args.output)
    ensure_dir(output)

    temp_files = []
    tmp_dir = Path("_temp_tut_combine")
    tmp_dir.mkdir(exist_ok=True)

    try:
        # 1. 统一所有素材的分辨率/帧率/像素格式
        for i, seg in enumerate(args.segments):
            seg_path = Path(seg)
            tmp_out = tmp_dir / f"seg_{i:02d}_norm.mp4"

            if seg_path.suffix in [".jpg", ".png"]:
                # 图片：循环播放 1.5s
                cmd = [
                    "ffmpeg",
                    "-y",
                    "-loop",
                    "1",
                    "-t",
                    "1.5",
                    "-i",
                    str(seg_path),
                    "-vf",
                    f"scale={CANVAS_WIDTH}:{CANVAS_HEIGHT},"
                    f"fps={TARGET_FPS},format=yuv420p,setsar=1",
                    "-c:v",
                    "libx264",
                    "-pix_fmt",
                    "yuv420p",
                    "-t",
                    "1.5",
                    str(tmp_out),
                ]
            else:
                # 视频：统一编码
                cmd = [
                    "ffmpeg",
                    "-y",
                    "-i",
                    str(seg_path),
                    "-vf",
                    f"scale={CANVAS_WIDTH}:{CANVAS_HEIGHT},"
                    f"fps={TARGET_FPS},format=yuv420p,setsar=1",
                    "-an",
                    "-c:v",
                    "libx264",
                    "-pix_fmt",
                    "yuv420p",
                    str(tmp_out),
                ]
            run_ffmpeg(cmd)
            temp_files.append(str(tmp_out))

        # 2. concat demuxer 拼接（不用 concat filter）
        concat_file = tmp_dir / "concat_list.txt"
        with open(concat_file, "w") as f:
            for tf in temp_files:
                f.write(f"file '{tf}'\n")

        cmd = [
            "ffmpeg",
            "-y",
            "-f",
            "concat",
            "-safe",
            "0",
            "-i",
            str(concat_file),
            "-c",
            "copy",
            str(output),
        ]
        run_ffmpeg(cmd)

    finally:
        # 清理临时文件
        import shutil

        shutil.rmtree(tmp_dir, ignore_errors=True)

    print(f"✅ 输出：{output}")


def cmd_pipeline(args: argparse.Namespace) -> None:
    """完整流水线：生成 tut_002 高亮框 → 上传动画 → 预拼接"""
    import shutil

    tmp_dir = Path("_temp_tut_pipeline")
    if tmp_dir.exists():
        shutil.rmtree(tmp_dir)
    tmp_dir.mkdir(parents=True)

    output_dir = Path(args.output_dir)
    output_dir.mkdir(parents=True, exist_ok=True)

    # Step 1: 生成 tut_002（高亮框）
    tut002 = output_dir / "tut_002.png"
    print("\n📸 Step 1: 生成 tut_002 高亮框...")
    img = Image.open(args.tut002_raw).convert("RGBA")
    if img.size != (CANVAS_WIDTH, CANVAS_HEIGHT):
        img = img.resize((CANVAS_WIDTH, CANVAS_HEIGHT), Image.LANCZOS)
    draw = ImageDraw.Draw(img)
    # 默认高亮框（居中偏上）
    box = (200, 400, 880, 900)
    rounded_rectangle(
        draw,
        box,
        radius=15,
        fill=HIGHLIGHT_FILL,
        outline=HIGHLIGHT_OUTLINE,
        width=HIGHLIGHT_LINE_WIDTH,
    )
    draw_arrow(draw, 540, 920, size=25, color=HIGHLIGHT_OUTLINE, width=3)
    img.save(tut002, "PNG")
    print(f"   ✅ {tut002}")

    # Step 2: 生成上传动画
    tut_upload = tmp_dir / "tut_upload_anim.mp4"
    print("\n🎬 Step 2: 生成上传动画...")
    cmd = [
        "ffmpeg",
        "-y",
        "-loop",
        "1",
        "-t",
        "2.5",
        "-i",
        str(args.tut003),
        "-loop",
        "1",
        "-t",
        "2.5",
        "-i",
        str(args.tut004),
        "-filter_complex",
        f"[0:v]fps={TARGET_FPS},format=yuv420p[v0];"
        f"[1:v]fps={TARGET_FPS},format=yuv420p[v1];"
        f"[v0][v1]xfade=transition=slideup:duration=0.8:offset=1.7",
        "-c:v",
        "libx264",
        "-pix_fmt",
        "yuv420p",
        "-t",
        "2.5",
        str(tut_upload),
    ]
    run_ffmpeg(cmd)

    # Step 3: 统一编码所有片段
    print("\n🔗 Step 3: 统一编码...")
    norm_files = []
    for i, (src, dur) in enumerate(
        [
            (Path(args.tut001), None),  # tut_001 视频
            (tut002, "1.2"),  # tut_002 图片
            (tut_upload, "2.5"),  # 上传动画
        ]
    ):
        tmp_out = tmp_dir / f"norm_{i}.mp4"
        if src.suffix in [".jpg", ".png"]:
            cmd = [
                "ffmpeg",
                "-y",
                "-loop",
                "1",
                "-t",
                dur,
                "-i",
                str(src),
                "-vf",
                f"scale={CANVAS_WIDTH}:{CANVAS_HEIGHT},"
                f"fps={TARGET_FPS},format=yuv420p,setsar=1",
                "-c:v",
                "libx264",
                "-pix_fmt",
                "yuv420p",
                "-t",
                dur,
                str(tmp_out),
            ]
        else:
            cmd = [
                "ffmpeg",
                "-y",
                "-i",
                str(src),
                "-vf",
                f"scale={CANVAS_WIDTH}:{CANVAS_HEIGHT},"
                f"fps={TARGET_FPS},format=yuv420p,setsar=1",
                "-an",
                "-c:v",
                "libx264",
                "-pix_fmt",
                "yuv420p",
                str(tmp_out),
            ]
        run_ffmpeg(cmd)
        norm_files.append(str(tmp_out))

    # Step 4: concat 拼接
    print("\n🔗 Step 4: 拼接...")
    tut_combined = output_dir / "tut_combined.mp4"
    concat_file = tmp_dir / "concat.txt"
    with open(concat_file, "w") as f:
        for nf in norm_files:
            f.write(f"file '{nf}'\n")

    cmd = [
        "ffmpeg",
        "-y",
        "-f",
        "concat",
        "-safe",
        "0",
        "-i",
        str(concat_file),
        "-c",
        "copy",
        str(tut_combined),
    ]
    run_ffmpeg(cmd)

    shutil.rmtree(tmp_dir)
    print(f"\n✅ 流水线完成！")
    print(f"   tut_002.png      → {tut002}")
    print(f"   tut_upload_anim.mp4 → {tmp_dir / 'tut_upload_anim.mp4'}")
    print(f"   tut_combined.mp4   → {tut_combined}")


# ─────────────────────────── 主入口 ───────────────────────────


def main():
    parser = argparse.ArgumentParser(
        description="教程合成脚本：高亮框 + 上传动画 + 预拼接",
        formatter_class=argparse.RawDescriptionHelpFormatter,
        epilog=__doc__,
    )
    subparsers = parser.add_subparsers(dest="cmd", required=True)

    # make-highlight
    p_highlight = subparsers.add_parser("make-highlight", help="生成带高亮框的教程截图")
    p_highlight.add_argument("--template-screenshot", required=True, help="原教程截图")
    p_highlight.add_argument("--output", required=True, help="输出 PNG 路径")
    p_highlight.add_argument("--box-x", type=int, default=200, help="高亮框 X")
    p_highlight.add_argument("--box-y", type=int, default=400, help="高亮框 Y")
    p_highlight.add_argument("--box-w", type=int, default=680, help="高亮框宽度")
    p_highlight.add_argument("--box-h", type=int, default=300, help="高亮框高度")

    # upload-anim
    p_anim = subparsers.add_parser("upload-anim", help="生成上传动画")
    p_anim.add_argument("--before", required=True, help="上传前状态图片")
    p_anim.add_argument("--after", required=True, help="上传后状态图片")
    p_anim.add_argument("--output", required=True, help="输出 MP4 路径")
    p_anim.add_argument("--duration", type=float, default=2.5, help="动画总时长（秒）")
    p_anim.add_argument(
        "--transition",
        default="slideup",
        choices=["slideup", "slideleft", "fadeblack"],
        help="转场类型",
    )

    # combine
    p_combine = subparsers.add_parser("combine", help="预拼接教程段")
    p_combine.add_argument(
        "--segments", nargs="+", required=True, help="素材列表（视频/图片）"
    )
    p_combine.add_argument("--output", required=True, help="输出 MP4 路径")

    # pipeline
    p_pipeline = subparsers.add_parser("pipeline", help="完整流水线")
    p_pipeline.add_argument("--tut001", required=True, help="tut_001 视频")
    p_pipeline.add_argument("--tut002-raw", required=True, help="tut_002 原图")
    p_pipeline.add_argument("--tut003", required=True, help="上传前状态 tut_003")
    p_pipeline.add_argument("--tut004", required=True, help="上传后状态 tut_004")
    p_pipeline.add_argument("--output-dir", default=".", help="输出目录")

    args = parser.parse_args()

    if args.cmd == "make-highlight":
        cmd_make_highlight(args)
    elif args.cmd == "upload-anim":
        cmd_upload_anim(args)
    elif args.cmd == "combine":
        cmd_combine(args)
    elif args.cmd == "pipeline":
        cmd_pipeline(args)


if __name__ == "__main__":
    main()