@remixmate/cli 0.1.0 → 0.1.2

package/skills/gen-digital-human/scripts/gen_digital_human_video.py

@@ -1,574 +0,0 @@
-#!/usr/bin/env python3
-"""
-数字人口播视频脚本 - 三合一工具
-支持列出形象、生成视频、查询状态。
-
-默认行为:
-  生成完成后直接输出视频 URL，供调用方（如 Claude）直接展示，无需落盘。
-
-用法:
-  python gen_digital_human_video.py --list-avatars
-  python gen_digital_human_video.py --avatar-id 7 --text "大家好"
-  python gen_digital_human_video.py --avatar-id 39 --audio-url "https://..."
-  python gen_digital_human_video.py --check-status --generation-id 593
-
-手动配置（命令行，优先级高于环境变量）:
-  --priv-token TOKEN  PrivToken
-
-环境变量:
-  MM_API_BASE_URL              - 后端 API 地址（默认: http://localhost:3001/api）
-  PRIV_TOKEN                   - PrivToken（优先读取；未配置时提示手动输入）
-"""
-
-import argparse
-import json
-import os
-import sys
-import time
-import urllib.request
-import urllib.error
-
-API_BASE_URL = os.environ.get("MM_API_BASE_URL", "http://localhost:3001/api")
-PRIVATE_TOKEN = ""  # 在 main() 中通过 resolve_token() 初始化
-SKILL_NAME = "gen-digital-human"
-AGENT_NAME = os.environ.get("AGENT_NAME", "")
-
-
-def resolve_token() -> str:
-    """优先从 PRIV_TOKEN 环境变量读取 token，取不到则报错退出"""
-    token = os.environ.get("PRIV_TOKEN", "")
-    if not token:
-        if sys.stdin.isatty():
-            token = input("请输入 PrivToken: ").strip()
-        else:
-            print("❌ PRIV_TOKEN env var not configured; cannot authenticate", file=sys.stderr)
-            sys.exit(1)
-    return token
-
-
-VALID_ASPECT_RATIOS = ["9:16", "16:9", "3:4", "1:1"]
-VALID_SOURCES = ["jimeng", "hifly"]
-
-STATUS_GENERATING = "generating"
-STATUS_COMPLETED = "completed"
-STATUS_FAILED = "failed"
-
-
-def api_request(path: str, payload: dict) -> dict:
-    """
-    通用 API 请求函数（统一 POST 方法）
-    """
-    if not PRIVATE_TOKEN:
-        print("❌ PrivToken not configured: set the PRIV_TOKEN env var", file=sys.stderr)
-        sys.exit(1)
-
-    url = f"{API_BASE_URL}{path}"
-    data = json.dumps(payload).encode("utf-8")
-    headers = {
-        "Content-Type": "application/json",
-        "X-Priv-Token": PRIVATE_TOKEN,
-        "x-invoke-skill": SKILL_NAME,
-    }
-    if AGENT_NAME:
-        headers["x-invoke-agent"] = AGENT_NAME
-
-    req = urllib.request.Request(url, data=data, headers=headers, method="POST")
-
-    try:
-        with urllib.request.urlopen(req, timeout=60) as resp:
-            return json.loads(resp.read().decode("utf-8"))
-    except urllib.error.HTTPError as e:
-        body = e.read().decode("utf-8")
-        print(f"❌ API request failed (HTTP {e.code}): {body}", file=sys.stderr)
-        sys.exit(1)
-    except urllib.error.URLError as e:
-        print(f"❌ network error: {e.reason}", file=sys.stderr)
-        sys.exit(1)
-
-
-# ---------------------------------------------------------------------------
-# 模式 1: 列出形象
-# ---------------------------------------------------------------------------
-
-def list_avatars(
-    source: str | None = None,
-    gender: str | None = None,
-    mine: bool = False,
-    name: str | None = None,
-) -> None:
-    """
-    查询可用数字人形象列表
-    """
-    payload = {"current": 1, "pageSize": 100}
-    if source:
-        payload["source"] = source
-    if gender:
-        payload["gender"] = gender
-    if mine:
-        payload["mine"] = True
-    if name:
-        payload["name"] = name
-
-    print("🔍 Fetching digital-human avatar list...")
-
-    result = api_request("/digital-human/avatar/page", payload)
-
-    if result.get("code") != 0:
-        msg = result.get("msg", "unknown error")
-        print(f"❌ API returned an error: {msg}", file=sys.stderr)
-        sys.exit(1)
-
-    avatar_list = (result.get("data") or {}).get("list", [])
-
-    if not avatar_list:
-        print("⚠️  No avatars available")
-        return
-
-    # Tabular output
-    print(f"\n{'ID':<8} {'Name':<16} {'Source':<10} {'Gender':<8} {'Ratio'}")
-    print(f"{'-' * 8} {'-' * 16} {'-' * 10} {'-' * 8} {'-' * 10}")
-    for avatar in avatar_list:
-        avatar_id = avatar.get("id", "")
-        avatar_name = avatar.get("name", "")
-        avatar_source = avatar.get("source", "")
-        avatar_gender = avatar.get("gender", "")
-        aspect_ratio = avatar.get("aspectRatio", "")
-        print(f"{avatar_id:<8} {avatar_name:<16} {avatar_source:<10} {avatar_gender:<8} {aspect_ratio}")
-
-    print(f"\n{len(avatar_list)} avatar(s) available")
-    print("💡 Use --avatar-id <ID> to pick an avatar for generation")
-
-
-# ---------------------------------------------------------------------------
-# 模式 2: 生成视频（默认模式）
-# ---------------------------------------------------------------------------
-
-def fetch_avatar_info(avatar_id: int) -> dict:
-    """
-    获取指定 avatar 的详细信息（默认 source / aspectRatio 等）。
-    先查普通分页，未命中再查 mine: true（与列表接口行为一致）。
-    """
-    payloads = [
-        {"current": 1, "pageSize": 100},
-        {"current": 1, "pageSize": 100, "mine": True},
-    ]
-    for payload in payloads:
-        result = api_request("/digital-human/avatar/page", payload)
-
-        if result.get("code") != 0:
-            msg = result.get("msg", "unknown error")
-            print(f"❌ avatar lookup failed: {msg}", file=sys.stderr)
-            sys.exit(1)
-
-        avatar_list = (result.get("data") or {}).get("list", [])
-        for avatar in avatar_list:
-            if str(avatar.get("id")) == str(avatar_id):
-                return avatar
-
-    print(f"❌ avatar_id={avatar_id} not found; use --list-avatars [--mine] to see what is available",
-          file=sys.stderr)
-    sys.exit(1)
-
-
-def _resolve_source(explicit_source: str | None, avatar_info: dict) -> str:
-    """
-    接口要求 source 为 jimeng / hifly；优先使用命令行，其次形象上的来源。
-    """
-    src = (explicit_source or "").strip() or (avatar_info.get("source") or "").strip()
-    if src not in VALID_SOURCES:
-        print(
-            "❌ 无法确定有效的 source（jimeng/hifly）。请用 --source 指定，或确认该形象带有来源字段。",
-            file=sys.stderr,
-        )
-        sys.exit(1)
-    return src
-
-
-def build_generate_payload(
-    avatar_id: int,
-    avatar_info: dict,
-    source: str | None,
-    text: str | None,
-    voice_id: str | None,
-    voice_name: str | None,
-    aspect_ratio: str | None,
-    prompt: str | None,
-    audio_url: str | None,
-) -> dict:
-    """
-    组装 /digital-human/video/generate 请求体（与后端 DigitalHumanVideoGenerateDTO 一致）。
-    有 audioUrl 时为音频驱动，可不传 text / voiceId / voiceName。
-    """
-    resolved_source = _resolve_source(source, avatar_info)
-    payload = {
-        "avatarId": int(avatar_id),
-        "source": resolved_source,
-    }
-
-    audio_url_stripped = (audio_url or "").strip()
-    if audio_url_stripped:
-        payload["audioUrl"] = audio_url_stripped
-        # 音频驱动下 text / voice 均可选；若仍传入 text 会一并提交供记录
-        if text and str(text).strip():
-            payload["text"] = str(text).strip()
-    else:
-        payload["text"] = text
-        payload["voiceId"] = voice_id
-        if voice_name:
-            payload["voiceName"] = voice_name
-
-    if aspect_ratio:
-        payload["aspectRatio"] = aspect_ratio
-    elif avatar_info.get("aspectRatio"):
-        payload["aspectRatio"] = avatar_info["aspectRatio"]
-
-    if prompt:
-        payload["prompt"] = prompt
-
-    return payload
-
-
-def submit_video_task(
-    avatar_id: int,
-    source: str | None,
-    text: str | None,
-    voice_id: str | None,
-    voice_name: str | None,
-    aspect_ratio: str | None,
-    prompt: str | None,
-    avatar_info: dict,
-    audio_url: str | None,
-) -> int:
-    """
-    提交数字人视频生成任务，返回 generationId
-    """
-    payload = build_generate_payload(
-        avatar_id=avatar_id,
-        avatar_info=avatar_info,
-        source=source,
-        text=text,
-        voice_id=voice_id,
-        voice_name=voice_name,
-        aspect_ratio=aspect_ratio,
-        prompt=prompt,
-        audio_url=audio_url,
-    )
-
-    print("🎬 Submitting digital-human video generation task...")
-    print(f"   avatar: {avatar_info.get('name', '')} (ID: {avatar_id})")
-    print(f"   source: {payload['source']}")
-    if payload.get("audioUrl"):
-        preview = payload["audioUrl"]
-        if len(preview) > 60:
-            preview = preview[:57] + "..."
-        print(f"   mode: audio-driven | audioUrl: {preview}")
-        if payload.get("text"):
-            tx = payload["text"]
-            print(f"   text (record-only): {tx[:50]}{'...' if len(tx) > 50 else ''}")
-    else:
-        t = text or ""
-        print(f"   mode: TTS | text: {t[:50]}{'...' if len(t) > 50 else ''}")
-        print(f"   voice: {voice_id}")
-    ar = payload.get("aspectRatio")
-    if ar:
-        print(f"   aspect ratio: {ar}")
-    if prompt:
-        print(f"   action prompt: {prompt}")
-
-    result = api_request("/digital-human/video/generate", payload)
-
-    if result.get("code") != 0:
-        msg = result.get("msg", "unknown error")
-        print(f"❌ API returned an error: {msg}", file=sys.stderr)
-        sys.exit(1)
-
-    result_data = result.get("data") or {}
-    generation_id = result_data.get("generationId")
-    if not generation_id:
-        print(f"❌ could not parse generationId; API returned: {json.dumps(result, ensure_ascii=False)}",
-              file=sys.stderr)
-        sys.exit(1)
-
-    print(f"✅ task submitted, generationId: {generation_id}")
-    return generation_id
-
-
-def poll_video_status(
-    generation_id: int,
-    poll_interval: int = 5,
-    max_wait: int = 600,
-) -> dict:
-    """
-    轮询数字人视频生成状态，返回最终结果
-    """
-    print(f"\n⏳ Waiting for video generation (up to {max_wait} s)...")
-
-    start_time = time.time()
-
-    while True:
-        elapsed = time.time() - start_time
-        if elapsed > max_wait:
-            print(f"\n❌ wait timed out (waited {elapsed:.0f} s)", file=sys.stderr)
-            print(f"   generationId: {generation_id}", file=sys.stderr)
-            print(f"   You can run --check-status --generation-id {generation_id} later",
-                  file=sys.stderr)
-            sys.exit(1)
-
-        result = api_request("/digital-human/video/status", {"generationId": generation_id})
-
-        if result.get("code") != 0:
-            msg = result.get("msg", "unknown error")
-            print(f"\n❌ status query failed: {msg}", file=sys.stderr)
-            sys.exit(1)
-
-        result_data = result.get("data") or {}
-        status = result_data.get("status", "unknown")
-
-        status_icons = {
-            STATUS_GENERATING: "🔄",
-            STATUS_COMPLETED: "✅",
-            STATUS_FAILED: "❌",
-        }
-        icon = status_icons.get(status, "⏳")
-        print(f"\r   {icon} status: {status} | elapsed: {elapsed:.0f}s", end="", flush=True)
-
-        if status == STATUS_COMPLETED:
-            print()  # 换行
-            return result_data
-
-        if status == STATUS_FAILED:
-            print()
-            error_msg = (
-                result_data.get("errorMessage")
-                or result_data.get("error")
-                or result_data.get("message")
-                or "unknown error"
-            )
-            print(f"❌ video generation failed: {error_msg}", file=sys.stderr)
-            sys.exit(1)
-
-        time.sleep(poll_interval)
-
-
-def generate_video(args: argparse.Namespace) -> None:
-    """
-    生成数字人视频的完整流程
-    """
-    # Step 1: fetch avatar info
-    print("🔍 Fetching avatar info...")
-    avatar_info = fetch_avatar_info(args.avatar_id)
-    print(f"   found avatar: {avatar_info.get('name', '')} (source: {avatar_info.get('source', '')})")
-
-    # Step 2: 提交视频生成任务
-    generation_id = submit_video_task(
-        avatar_id=args.avatar_id,
-        source=args.source,
-        text=args.text,
-        voice_id=args.voice_id,
-        voice_name=args.voice_name,
-        aspect_ratio=args.aspect_ratio,
-        prompt=args.prompt,
-        avatar_info=avatar_info,
-        audio_url=args.audio_url,
-    )
-
-    # Step 3: 轮询等待完成
-    result_data = poll_video_status(
-        generation_id=generation_id,
-        poll_interval=args.poll_interval,
-        max_wait=args.max_wait,
-    )
-
-    print("🎉 Video generated successfully!")
-
-    # Step 4: emit URL
-    video_url = result_data.get("videoUrl")
-    if video_url:
-        print("\n🔗 Video URL (renders directly, no download needed):")
-        print(video_url)
-    else:
-        print("⚠️  video generated but no URL was returned")
-        print(f"   generationId: {generation_id}")
-        print(f"   you can run --check-status --generation-id {generation_id} later")
-    print("\n🎉 Digital-human video generation done!")
-
-
-# ---------------------------------------------------------------------------
-# 模式 3: 查询状态
-# ---------------------------------------------------------------------------
-
-def check_status(generation_id: int) -> None:
-    """
-    查询数字人视频生成任务状态
-    """
-    print(f"🔍 Querying task status (generationId: {generation_id})...")
-
-    result = api_request("/digital-human/video/status", {"generationId": generation_id})
-
-    if result.get("code") != 0:
-        msg = result.get("msg", "unknown error")
-        print(f"❌ API returned an error: {msg}", file=sys.stderr)
-        sys.exit(1)
-
-    result_data = result.get("data") or {}
-    status = result_data.get("status", "unknown")
-    video_url = result_data.get("videoUrl", "")
-
-    status_icons = {
-        STATUS_GENERATING: "🔄",
-        STATUS_COMPLETED: "✅",
-        STATUS_FAILED: "❌",
-    }
-    icon = status_icons.get(status, "❓")
-    print(f"\n{icon} task status: {status}")
-
-    if video_url:
-        print(f"   video URL: {video_url}")
-
-    if status == STATUS_FAILED:
-        error_msg = (
-            result_data.get("errorMessage")
-            or result_data.get("error")
-            or result_data.get("message")
-            or "unknown error"
-        )
-        print(f"   failure reason: {error_msg}")
-
-
-# ---------------------------------------------------------------------------
-# main
-# ---------------------------------------------------------------------------
-
-def main() -> None:
-    parser = argparse.ArgumentParser(
-        description="Digital-human talking-head video tool — list avatars / generate video / check status",
-        formatter_class=argparse.RawDescriptionHelpFormatter,
-        epilog="""
-Examples:
-  # List all avatars
-  python gen_digital_human_video.py --list-avatars
-  python gen_digital_human_video.py --list-avatars --source hifly --gender female
-
-  # Default mode: only the URL is printed so it can be displayed inline (recommended)
-  python gen_digital_human_video.py --avatar-id 7 --text "Hello and welcome to the live stream"
-  python gen_digital_human_video.py --avatar-id 7 -t "Hi there" --voice-id "female-shaonv" --aspect-ratio 9:16
-
-  # Audio-driven (text / voice-id / voice-name optional)
-  python gen_digital_human_video.py --avatar-id 39 --source hifly --audio-url "https://example.com/a.mp3" --aspect-ratio 16:9
-
-  # Check task status
-  python gen_digital_human_video.py --check-status --generation-id 593
-        """,
-    )
-
-    parser.add_argument(
-        "--priv-token",
-        metavar="TOKEN",
-        default=None,
-        help="PrivToken (priority: this flag > PRIV_TOKEN env var > interactive prompt)",
-    )
-
-    # Mode switches
-    parser.add_argument(
-        "--list-avatars",
-        action="store_true",
-        help="List available digital-human avatars and exit",
-    )
-    parser.add_argument(
-        "--check-status",
-        action="store_true",
-        help="Check the status of a digital-human video generation task",
-    )
-
-    # Listing filters / generation-time source override (defaults to the avatar's own source)
-    parser.add_argument("--source", choices=["jimeng", "hifly"],
-                        help="jimeng / hifly; filter in list mode, override the avatar's default source in generate mode")
-    parser.add_argument("--gender", choices=["male", "female"],
-                        help="Filter avatars by gender: male / female")
-    parser.add_argument("--mine", action="store_true",
-                        help="Only list user-custom avatars")
-    parser.add_argument("--name", help="Fuzzy-search avatars by name")
-
-    # Generation parameters
-    parser.add_argument("--avatar-id", type=int, help="Digital-human avatar id (required in generate mode)")
-    parser.add_argument("--text", "-t", help="Narration text (required in TTS mode; optional with --audio-url)")
-    parser.add_argument(
-        "--audio-url",
-        help="Pre-synthesized audio URL (audio-driven; with this set, --text / --voice-id / --voice-name are optional)",
-    )
-    parser.add_argument(
-        "--voice-id",
-        default=None,
-        help="Voice id (required for TTS; default male-qn-qingse; optional with --audio-url)",
-    )
-    parser.add_argument("--voice-name", default="", help="Voice display name (optional for TTS)")
-    parser.add_argument("--aspect-ratio", choices=VALID_ASPECT_RATIOS,
-                        help=f"Aspect ratio: {', '.join(VALID_ASPECT_RATIOS)}")
-    parser.add_argument("--prompt", help="Action prompt (e.g. 'more hand gestures')")
-    parser.add_argument("--poll-interval", type=int, default=5,
-                        help="Status polling interval (seconds, default: 5)")
-    parser.add_argument("--max-wait", type=int, default=600,
-                        help="Maximum wait time (seconds, default: 600)")
-
-    # Status-query parameters
-    parser.add_argument("--generation-id", type=int, help="Generation task id (required in check-status mode)")
-
-    args = parser.parse_args()
-
-    # 初始化 token：--priv-token 参数 > PRIV_TOKEN 环境变量 > 手动输入
-    global PRIVATE_TOKEN
-    if args.priv_token is not None:
-        PRIVATE_TOKEN = args.priv_token
-    else:
-        PRIVATE_TOKEN = resolve_token()
-
-    # 模式 1: 列出形象
-    if args.list_avatars:
-        list_avatars(
-            source=args.source,
-            gender=args.gender,
-            mine=args.mine,
-            name=args.name,
-        )
-        return
-
-    # 模式 3: 查询状态
-    if args.check_status:
-        if not args.generation_id:
-            print("❌ check-status mode requires --generation-id", file=sys.stderr)
-            parser.print_help()
-            sys.exit(1)
-        check_status(generation_id=args.generation_id)
-        return
-
-    # 模式 2: 生成视频（默认）
-    if not args.avatar_id:
-        print("❌ provide --avatar-id to pick the avatar", file=sys.stderr)
-        print("   use --list-avatars to see available avatars", file=sys.stderr)
-        parser.print_help()
-        sys.exit(1)
-
-    audio_url_stripped = (args.audio_url or "").strip()
-    if audio_url_stripped:
-        args.audio_url = audio_url_stripped
-    else:
-        args.audio_url = ""
-
-    if not args.audio_url:
-        if not args.text:
-            print("❌ in TTS mode provide --text, or use --audio-url for audio-driven mode", file=sys.stderr)
-            parser.print_help()
-            sys.exit(1)
-        if not args.text.strip():
-            print("❌ narration text cannot be empty", file=sys.stderr)
-            sys.exit(1)
-        if not args.voice_id:
-            args.voice_id = "male-qn-qingse"
-    else:
-        args.text = args.text or ""
-
-    generate_video(args)
-
-
-if __name__ == "__main__":
-    main()