|
|
@@ -0,0 +1,308 @@
|
|
|
+#!/usr/bin/env python3
|
|
|
+# -*- coding: utf-8 -*-
|
|
|
+"""
|
|
|
+百炼 wan3（万相3.0）视频生成调用脚本
|
|
|
+====================================
|
|
|
+仅供 beauty-talk-science-video-script 技能在「用户确认脚本后」调用。
|
|
|
+
|
|
|
+流程：
|
|
|
+ 1. 解析参考素材（人物完整形象照 / 门店内景照 / 插画参考图 / 音色文件）。
|
|
|
+ - 公网 URL（--media 'https://...:类型'）：默认直接透传，wan3 原生支持公网 URL。
|
|
|
+ 若 config 的 reupload_external_urls=true（wan3 拒收站外 URL 时），先下载再编码 base64 直传。
|
|
|
+ - 本地文件（生成的人物形象照 / 门店内景 / 插画）：读取后编码为
|
|
|
+ data:{mime};base64,{b64} 直接放入 media[].url（wan3 原生支持 base64 图像，免去文件上传端点）。
|
|
|
+ 2. 组装 video-synthesis 异步请求（model / prompt / media / parameters），提交任务拿 task_id。
|
|
|
+ 3. 轮询任务状态，直到 SUCCEEDED。
|
|
|
+ 4. 下载生成的视频到本地 output_dir。
|
|
|
+
|
|
|
+端点（与用户所给 LLM base_url 不同，视频服务走同 workspace、同地域、但**不带 llm- 前缀**的主机）：
|
|
|
+ POST https://{WorkspaceId}.{region}.maas.aliyuncs.com/api/v1/services/aigc/video-generation/video-synthesis
|
|
|
+ GET https://{WorkspaceId}.{region}.maas.aliyuncs.com/api/v1/tasks/{task_id}
|
|
|
+用户所给 "https://llm-xxxx.cn-beijing.maas.aliyuncs.com/compatible-mode/v1" 是 **LLM 对话网关**
|
|
|
+（OpenAI 兼容模式），视频生成不使用它；脚本按 workspace_id + region 自动拼视频主机，
|
|
|
+也可在 config.api_base 显式覆盖。
|
|
|
+
|
|
|
+凭证：
|
|
|
+ - DASHSCOPE_API_KEY 必须存在于环境变量（不写死在代码/配置里）。
|
|
|
+ - region / workspace_id / model / resolution / ratio / output_dir 写在同目录 config.json。
|
|
|
+ - 模型 / Endpoint / Key 必须同地域，跨地域调用会失败。
|
|
|
+
|
|
|
+注意：
|
|
|
+ - 本脚本只做「整段重新生成」，不使用 wan3 的视频编辑 / 延长模式（对应技能第 8 条：不支持二次创作）。
|
|
|
+ - 单次生成时长上限 30 秒（wan3.0 限制）。超过 30 秒的脚本请在技能侧先拆分/精简，本脚本强制 duration≤30。
|
|
|
+"""
|
|
|
+
|
|
|
+import os
|
|
|
+import sys
|
|
|
+import json
|
|
|
+import time
|
|
|
+import base64
|
|
|
+import argparse
|
|
|
+import mimetypes
|
|
|
+
|
|
|
+import requests
|
|
|
+
|
|
|
+# --------------------------------------------------------------------------- #
|
|
|
+# 配置与端点
|
|
|
+# --------------------------------------------------------------------------- #
|
|
|
+
|
|
|
+def load_config(path):
|
|
|
+ with open(path, "r", encoding="utf-8") as f:
|
|
|
+ return json.load(f)
|
|
|
+
|
|
|
+
|
|
|
+def api_base(cfg):
|
|
|
+ """视频服务主机：优先用 config.api_base；否则按 workspace_id + region 拼（不带 llm- 前缀）。"""
|
|
|
+ if cfg.get("api_base"):
|
|
|
+ return cfg["api_base"].rstrip("/")
|
|
|
+ ws = cfg["workspace_id"]
|
|
|
+ region = cfg["region"]
|
|
|
+ return f"https://{ws}.{region}.maas.aliyuncs.com"
|
|
|
+
|
|
|
+
|
|
|
+def auth_headers(key):
|
|
|
+ return {
|
|
|
+ "Authorization": f"Bearer {key}",
|
|
|
+ "Content-Type": "application/json",
|
|
|
+ }
|
|
|
+
|
|
|
+
|
|
|
+def resolve_key(here):
|
|
|
+ """解析百炼 / DashScope API Key。
|
|
|
+ 优先级：环境变量 DASHSCOPE_API_KEY > 脚本同目录 .env（DASHSCOPE_API_KEY=...）。
|
|
|
+ 注意：.env 含密钥，禁止随技能分发；重新打包技能前请删除本地 .env。
|
|
|
+ """
|
|
|
+ env = os.environ.get("DASHSCOPE_API_KEY")
|
|
|
+ if env and env.strip():
|
|
|
+ return env.strip()
|
|
|
+ env_path = os.path.join(here, ".env")
|
|
|
+ if os.path.exists(env_path):
|
|
|
+ try:
|
|
|
+ with open(env_path, "r", encoding="utf-8") as f:
|
|
|
+ for line in f:
|
|
|
+ line = line.strip()
|
|
|
+ if not line or line.startswith("#"):
|
|
|
+ continue
|
|
|
+ if line.startswith("DASHSCOPE_API_KEY="):
|
|
|
+ v = line.split("=", 1)[1].strip().strip('"').strip("'")
|
|
|
+ if v:
|
|
|
+ return v
|
|
|
+ except Exception:
|
|
|
+ pass
|
|
|
+ sys.exit(
|
|
|
+ "未找到 DASHSCOPE_API_KEY：\n"
|
|
|
+ " (1) 运行前 export DASHSCOPE_API_KEY=sk-真实key（当前 shell 生效）；或\n"
|
|
|
+ " (2) 在脚本同目录创建 .env，写入一行 DASHSCOPE_API_KEY=sk-真实key（详见 .env.example）。\n"
|
|
|
+ "注意：北京地域创建、sk- 开头；密钥严禁提交或分发。"
|
|
|
+ )
|
|
|
+
|
|
|
+
|
|
|
+# --------------------------------------------------------------------------- #
|
|
|
+# 素材 -> URL（公网URL直传 / 本地文件 base64 直传）
|
|
|
+# --------------------------------------------------------------------------- #
|
|
|
+
|
|
|
+def local_file_to_data_uri(filepath):
|
|
|
+ """读取本地文件，编码为 data:{mime};base64,{b64}，wan3 原生支持图像 base64 直传。"""
|
|
|
+ mime = mimetypes.guess_type(filepath)[0] or "application/octet-stream"
|
|
|
+ with open(filepath, "rb") as f:
|
|
|
+ b64 = base64.b64encode(f.read()).decode("ascii")
|
|
|
+ return f"data:{mime};base64,{b64}"
|
|
|
+
|
|
|
+
|
|
|
+def download_to_temp(url, suffix=".tmp"):
|
|
|
+ """把公网 URL 下载到临时文件，返回本地路径（用于 reupload_external_urls）。"""
|
|
|
+ import tempfile
|
|
|
+ fd, tmppath = tempfile.mkstemp(suffix=suffix)
|
|
|
+ os.close(fd)
|
|
|
+ try:
|
|
|
+ resp = requests.get(url, timeout=300, stream=True)
|
|
|
+ resp.raise_for_status()
|
|
|
+ with open(tmppath, "wb") as f:
|
|
|
+ for chunk in resp.iter_content(8192):
|
|
|
+ f.write(chunk)
|
|
|
+ except Exception:
|
|
|
+ if os.path.exists(tmppath):
|
|
|
+ os.remove(tmppath)
|
|
|
+ raise
|
|
|
+ return tmppath
|
|
|
+
|
|
|
+
|
|
|
+# --------------------------------------------------------------------------- #
|
|
|
+# 任务创建与轮询
|
|
|
+# --------------------------------------------------------------------------- #
|
|
|
+
|
|
|
+def create_task(base, key, payload):
|
|
|
+ url = base + "/api/v1/services/aigc/video-generation/video-synthesis"
|
|
|
+ headers = auth_headers(key)
|
|
|
+ headers["X-DashScope-Async"] = "enable"
|
|
|
+ resp = requests.post(url, headers=headers, json=payload, timeout=120)
|
|
|
+ if resp.status_code not in (200, 202):
|
|
|
+ raise RuntimeError(f"建任务失败：HTTP {resp.status_code} {resp.text[:500]}")
|
|
|
+ return resp.json()
|
|
|
+
|
|
|
+
|
|
|
+def poll_task(base, key, task_id, interval, timeout):
|
|
|
+ url = base + f"/api/v1/tasks/{task_id}"
|
|
|
+ headers = {"Authorization": f"Bearer {key}"}
|
|
|
+ waited = 0
|
|
|
+ while waited < timeout:
|
|
|
+ try:
|
|
|
+ resp = requests.get(url, headers=headers, timeout=60)
|
|
|
+ if resp.status_code == 200:
|
|
|
+ data = resp.json()
|
|
|
+ status = (
|
|
|
+ (data.get("output", {}) or {}).get("task_status")
|
|
|
+ or data.get("task_status")
|
|
|
+ )
|
|
|
+ if status == "SUCCEEDED":
|
|
|
+ return data
|
|
|
+ if status in ("FAILED", "UNKNOWN"):
|
|
|
+ raise RuntimeError(f"任务失败：{json.dumps(data, ensure_ascii=False)[:800]}")
|
|
|
+ except requests.RequestException:
|
|
|
+ pass
|
|
|
+ time.sleep(interval)
|
|
|
+ waited += interval
|
|
|
+ raise RuntimeError("轮询超时（视频生成可能仍在后台，请用 task_id 手动查询）")
|
|
|
+
|
|
|
+
|
|
|
+def extract_video_url(data):
|
|
|
+ out = data.get("output", {}) or {}
|
|
|
+ for k in ("video_url", "videoUrl", "url"):
|
|
|
+ if out.get(k):
|
|
|
+ return out[k]
|
|
|
+ for r in out.get("results") or []:
|
|
|
+ if r.get("url"):
|
|
|
+ return r["url"]
|
|
|
+ raise RuntimeError("任务成功但未找到视频 URL")
|
|
|
+
|
|
|
+
|
|
|
+def download(url, out_path):
|
|
|
+ resp = requests.get(url, timeout=300, stream=True)
|
|
|
+ resp.raise_for_status()
|
|
|
+ with open(out_path, "wb") as f:
|
|
|
+ for chunk in resp.iter_content(8192):
|
|
|
+ f.write(chunk)
|
|
|
+ return out_path
|
|
|
+
|
|
|
+
|
|
|
+# --------------------------------------------------------------------------- #
|
|
|
+# 入口
|
|
|
+# --------------------------------------------------------------------------- #
|
|
|
+
|
|
|
+def main():
|
|
|
+ here = os.path.dirname(os.path.abspath(__file__))
|
|
|
+ ap = argparse.ArgumentParser(description="百炼 wan3 视频生成（单段，≤30s）")
|
|
|
+ ap.add_argument("--config", default=os.path.join(here, "config.json"))
|
|
|
+ ap.add_argument("--prompt", required=True, help="单段视频提示词（含口播对白与 Image 1 / Audio 1 引用）")
|
|
|
+ ap.add_argument(
|
|
|
+ "--media",
|
|
|
+ nargs="*",
|
|
|
+ default=[],
|
|
|
+ help="参考素材，格式 '路径或URL:类型'。类型：reference_image / reference_audio。"
|
|
|
+ "如 /abs/char.jpg:reference_image 或 https://cdn.x/char.jpg:reference_image；"
|
|
|
+ "公网 URL 默认直接透传；本地文件编码 base64 直传；详见 config.reupload_external_urls。",
|
|
|
+ )
|
|
|
+ ap.add_argument("--duration", type=int, default=30, help="时长（秒），≤30")
|
|
|
+ ap.add_argument("--ratio", default=None, help="画幅，如 9:16")
|
|
|
+ ap.add_argument("--resolution", default=None, help="分辨率，如 720P")
|
|
|
+ ap.add_argument("--model", default=None)
|
|
|
+ ap.add_argument("--output", default=None, help="输出视频本地路径")
|
|
|
+ ap.add_argument("--poll-interval", type=int, default=10)
|
|
|
+ ap.add_argument("--timeout", type=int, default=1200)
|
|
|
+ args = ap.parse_args()
|
|
|
+
|
|
|
+ if args.duration > 30:
|
|
|
+ sys.exit("wan3 单次生成上限 30 秒，duration 必须 ≤30。请先在技能侧精简口播。")
|
|
|
+ if args.duration < 2:
|
|
|
+ sys.exit("duration 必须 ≥2 秒。")
|
|
|
+
|
|
|
+ key = resolve_key(here)
|
|
|
+
|
|
|
+ cfg = load_config(args.config)
|
|
|
+ base = api_base(cfg)
|
|
|
+ model = args.model or cfg.get("model", "wan3.0-video")
|
|
|
+ resolution = args.resolution or cfg.get("resolution", "720P")
|
|
|
+ ratio = args.ratio or cfg.get("ratio", "9:16")
|
|
|
+ out_dir = os.path.join(here, cfg.get("output_dir", "outputs"))
|
|
|
+ os.makedirs(out_dir, exist_ok=True)
|
|
|
+
|
|
|
+ # 1) 解析素材 -> URL（公网URL直传 / 本地文件base64直传）
|
|
|
+ reupload = bool(cfg.get("reupload_external_urls"))
|
|
|
+ media = []
|
|
|
+ for m in args.media:
|
|
|
+ raw = m.strip()
|
|
|
+ if "://" in raw:
|
|
|
+ # 公网 URL 形式：'https://...:reference_image|reference_audio'（类型可选）
|
|
|
+ head, sep, tail = raw.rpartition(":")
|
|
|
+ if sep and tail in ("reference_image", "reference_audio"):
|
|
|
+ raw, mtype = head, tail
|
|
|
+ else:
|
|
|
+ mtype = "reference_image"
|
|
|
+ else:
|
|
|
+ raw, _, mtype = raw.partition(":")
|
|
|
+ mtype = (mtype or "reference_image").strip()
|
|
|
+ raw = raw.strip()
|
|
|
+ if "://" in raw:
|
|
|
+ if reupload:
|
|
|
+ print(f"[下载并base64] {raw} ({mtype}) ...", file=sys.stderr)
|
|
|
+ tmp = download_to_temp(raw)
|
|
|
+ u = local_file_to_data_uri(tmp)
|
|
|
+ try:
|
|
|
+ os.remove(tmp)
|
|
|
+ except OSError:
|
|
|
+ pass
|
|
|
+ else:
|
|
|
+ print(f"[透传URL] {raw} ({mtype})", file=sys.stderr)
|
|
|
+ u = raw
|
|
|
+ media.append({"type": mtype, "url": u})
|
|
|
+ continue
|
|
|
+ fpath = os.path.expanduser(raw)
|
|
|
+ if not os.path.exists(fpath):
|
|
|
+ sys.exit(f"素材不存在：{fpath}")
|
|
|
+ print(f"[base64直传] {fpath} ({mtype}) ...", file=sys.stderr)
|
|
|
+ u = local_file_to_data_uri(fpath)
|
|
|
+ media.append({"type": mtype, "url": u})
|
|
|
+
|
|
|
+ # 2) 组装请求
|
|
|
+ payload = {
|
|
|
+ "model": model,
|
|
|
+ "input": {"prompt": args.prompt},
|
|
|
+ "parameters": {
|
|
|
+ "resolution": resolution,
|
|
|
+ "ratio": ratio,
|
|
|
+ "duration": args.duration,
|
|
|
+ "audio": True,
|
|
|
+ "prompt_extend": False,
|
|
|
+ "watermark": False,
|
|
|
+ },
|
|
|
+ }
|
|
|
+ if media:
|
|
|
+ payload["input"]["media"] = media
|
|
|
+
|
|
|
+ print("[提交] 创建视频生成任务 ...", file=sys.stderr)
|
|
|
+ created = create_task(base, key, payload)
|
|
|
+ task_id = (created.get("output", {}) or {}).get("task_id") or created.get("task_id")
|
|
|
+ if not task_id:
|
|
|
+ sys.exit(f"未返回 task_id：{json.dumps(created, ensure_ascii=False)[:500]}")
|
|
|
+ print(f"[轮询] task_id={task_id}", file=sys.stderr)
|
|
|
+
|
|
|
+ # 3) 轮询
|
|
|
+ result = poll_task(base, key, task_id, args.poll_interval, args.timeout)
|
|
|
+
|
|
|
+ # 4) 下载
|
|
|
+ vurl = extract_video_url(result)
|
|
|
+ out_path = args.output or os.path.join(out_dir, f"video_{int(time.time())}.mp4")
|
|
|
+ out_path = os.path.abspath(out_path)
|
|
|
+ print(f"[下载] {vurl} -> {out_path}", file=sys.stderr)
|
|
|
+ download(vurl, out_path)
|
|
|
+
|
|
|
+ # 仅供技能解析的结构化结果（走 stdout）
|
|
|
+ print("VIDEO_PATH=" + out_path)
|
|
|
+ print("RESULT_JSON=" + json.dumps(
|
|
|
+ {"task_id": task_id, "video_path": out_path, "video_url": vurl},
|
|
|
+ ensure_ascii=False,
|
|
|
+ ))
|
|
|
+
|
|
|
+
|
|
|
+if __name__ == "__main__":
|
|
|
+ main()
|
