diff --git a/assets/demo.jpg b/assets/demo.jpg new file mode 100644 index 0000000..3526080 Binary files /dev/null and b/assets/demo.jpg differ diff --git a/configs/config.yaml b/configs/config.yaml index 11ab367..e1bbc88 100644 --- a/configs/config.yaml +++ b/configs/config.yaml @@ -4,6 +4,38 @@ app: # ComfyUI output directory on the same machine running this code comfy_output_dir: "./ComfyUI/output" +global: + # Used by prompt_injector + adapters. + style: "" + character: "" + negative_prompt: "" + +llm: + # Controls /script + /refine generation. + provider: "mock" # "openai" to enable OpenAI/DashScope calls + +image: + provider: "mock" # "mock" | "comfy" | "replicate" | "openai" + # Generic model name (used by some providers as fallback). + model: "" + + replicate: + # Example: "stability-ai/sdxl" + model: "stability-ai/sdxl" + + openai: + # Example: "gpt-image-1" + model: "gpt-image-1" + +image_fallback: + provider: "mock" + +video: + provider: "moviepy" + +tts: + provider: "edge" + openai: # Prefer environment variables in real deployments. # OPENAI_API_KEY must be set; OPENAI_BASE_URL optional (for DeepSeek / other gateways). diff --git a/engine/adapters/__init__.py b/engine/adapters/__init__.py new file mode 100644 index 0000000..8b13789 --- /dev/null +++ b/engine/adapters/__init__.py @@ -0,0 +1 @@ + diff --git a/engine/adapters/image/__init__.py b/engine/adapters/image/__init__.py new file mode 100644 index 0000000..8b13789 --- /dev/null +++ b/engine/adapters/image/__init__.py @@ -0,0 +1 @@ + diff --git a/engine/adapters/image/base.py b/engine/adapters/image/base.py new file mode 100644 index 0000000..b29151b --- /dev/null +++ b/engine/adapters/image/base.py @@ -0,0 +1,9 @@ +from __future__ import annotations + +from pathlib import Path + + +class BaseImageGen: + def generate(self, prompt: dict[str, str], output_dir: str | Path) -> str: + raise NotImplementedError + diff --git a/engine/adapters/image/comfy_adapter.py b/engine/adapters/image/comfy_adapter.py new file mode 100644 index 0000000..6a44fd6 --- /dev/null +++ b/engine/adapters/image/comfy_adapter.py @@ -0,0 +1,36 @@ +from __future__ import annotations + +from pathlib import Path +from typing import Any + +from engine.comfy_client import generate_image as comfy_generate_image +from engine.config import AppConfig + +from .base import BaseImageGen +from .mock_adapter import MockImageGen + + +class ComfyAdapter(BaseImageGen): + def __init__(self, cfg: AppConfig): + self.cfg = cfg + self.fallback = MockImageGen() + + def generate(self, prompt: dict[str, str], output_dir: str | Path) -> str: + positive = str(prompt.get("positive", "") or "") + negative = str(prompt.get("negative", "") or "") + try: + return str( + comfy_generate_image( + positive, + output_dir, + negative_text=negative or None, + cfg=self.cfg, + timeout_s=60, + retry=2, + filename_prefix="shot", + ) + ) + except Exception as e: + # Let render_pipeline do configured fallback. + raise + diff --git a/engine/adapters/image/mock_adapter.py b/engine/adapters/image/mock_adapter.py new file mode 100644 index 0000000..7d14b49 --- /dev/null +++ b/engine/adapters/image/mock_adapter.py @@ -0,0 +1,45 @@ +from __future__ import annotations + +import os +import uuid +from pathlib import Path +from urllib.request import urlopen + +from PIL import Image + +from .base import BaseImageGen + + +ASSETS_DIR = "assets" +DEMO_IMAGE = os.path.join(ASSETS_DIR, "demo.jpg") + + +def ensure_demo_image() -> None: + os.makedirs(ASSETS_DIR, exist_ok=True) + if os.path.exists(DEMO_IMAGE): + return + + url = "https://picsum.photos/1280/720" + with urlopen(url, timeout=30) as resp: + data = resp.read() + with open(DEMO_IMAGE, "wb") as f: + f.write(data) + + +class MockImageGen(BaseImageGen): + def generate(self, prompt: dict[str, str], output_dir: str | Path) -> str: + # prompt is accepted for interface consistency; mock uses only demo.jpg. + _ = prompt + ensure_demo_image() + out_dir = Path(output_dir) + out_dir.mkdir(parents=True, exist_ok=True) + out_path = out_dir / f"shot_{uuid.uuid4().hex}.png" + try: + # Convert to PNG so verification criteria can match *.png. + img = Image.open(DEMO_IMAGE).convert("RGB") + img.save(str(out_path), format="PNG") + except Exception: + # Last-resort: if PNG conversion fails, still write a best-effort copy. + out_path.write_bytes(Path(DEMO_IMAGE).read_bytes()) + return str(out_path) + diff --git a/engine/adapters/image/openai_image_adapter.py b/engine/adapters/image/openai_image_adapter.py new file mode 100644 index 0000000..da23e1a --- /dev/null +++ b/engine/adapters/image/openai_image_adapter.py @@ -0,0 +1,83 @@ +from __future__ import annotations + +import os +import uuid +from io import BytesIO +from pathlib import Path +from typing import Any + +import requests +from PIL import Image + +from engine.config import AppConfig + +from .base import BaseImageGen + + +class OpenAIImageAdapter(BaseImageGen): + """ + Optional image provider adapter using OpenAI Images API (or OpenAI-compatible gateways). + Requires `openai` python package and a configured API key via environment variables. + """ + + def __init__(self, cfg: AppConfig): + self.cfg = cfg + # Expected keys (configurable): + # - image.openai.model + # - openai.api_key_env / openai.base_url_env (reuses existing engine/script_gen config fields) + self.model = str(cfg.get("image.openai.model", cfg.get("image.model", ""))).strip() + if not self.model: + raise ValueError("OpenAIImageAdapter requires `image.openai.model` (or `image.model`).") + + api_key_env_or_literal = str(cfg.get("openai.api_key_env", "OPENAI_API_KEY") or "OPENAI_API_KEY").strip() + # Support both: + # - env var name (e.g. OPENAI_API_KEY) + # - literal API key (e.g. starts with `sk-...`) for quick local POCs. + if api_key_env_or_literal.startswith("sk-"): + api_key = api_key_env_or_literal + else: + api_key = os.environ.get(api_key_env_or_literal) + if not api_key: + raise RuntimeError(f"OpenAIImageAdapter missing API key: `{api_key_env_or_literal}`") + self.api_key = api_key + + base_url_env_or_literal = str(cfg.get("openai.base_url_env", "https://api.openai.com/v1")).strip() + self.base_url = base_url_env_or_literal.rstrip("/") if base_url_env_or_literal else "https://api.openai.com/v1" + + # Lazy import to avoid hard dependency for mock/comfy users. + from openai import OpenAI # type: ignore + + self.client = OpenAI(api_key=self.api_key, base_url=self.base_url) + + def generate(self, prompt: dict[str, str], output_dir: str | Path) -> str: + output_dir = Path(output_dir) + output_dir.mkdir(parents=True, exist_ok=True) + + positive = prompt.get("positive", "") + negative = prompt.get("negative", "") + # OpenAI Images API generally doesn't expose a dedicated negative_prompt field. + # To keep interface consistency, embed negative hints into the prompt text. + if negative: + prompt_text = f"{positive}\nNegative prompt: {negative}" + else: + prompt_text = positive + + result = self.client.images.generate(model=self.model, prompt=prompt_text) + + # OpenAI SDK: result.data[0].url + url: str | None = None + try: + url = result.data[0].url # type: ignore[attr-defined] + except Exception: + pass + if not url: + raise RuntimeError("OpenAIImageAdapter unexpected response: missing image url") + + r = requests.get(url, timeout=60) + r.raise_for_status() + + out_path = output_dir / f"shot_{uuid.uuid4().hex}.png" + img = Image.open(BytesIO(r.content)).convert("RGB") + img.save(str(out_path), format="PNG") + return str(out_path) + diff --git a/engine/adapters/image/replicate_adapter.py b/engine/adapters/image/replicate_adapter.py new file mode 100644 index 0000000..70c26b4 --- /dev/null +++ b/engine/adapters/image/replicate_adapter.py @@ -0,0 +1,60 @@ +from __future__ import annotations + +import uuid +from pathlib import Path +from typing import Any + +import requests +from PIL import Image + +from engine.config import AppConfig + +from .base import BaseImageGen + + +class ReplicateAdapter(BaseImageGen): + def __init__(self, cfg: AppConfig): + self.cfg = cfg + # Expected: image.replicate.model + self.model = str(cfg.get("image.replicate.model", cfg.get("image.model", ""))).strip() + if not self.model: + raise ValueError("ReplicateAdapter requires `image.replicate.model` (or `image.model`).") + + # Import lazily so that environments without replicate installed can still run with mock/comfy. + import replicate # type: ignore + + self.replicate = replicate + + def generate(self, prompt: dict[str, str], output_dir: str | Path) -> str: + output_dir = Path(output_dir) + output_dir.mkdir(parents=True, exist_ok=True) + + input_payload: dict[str, Any] = { + "prompt": prompt.get("positive", ""), + "negative_prompt": prompt.get("negative", ""), + } + + # replicate.run is synchronous when wait is handled by the SDK version. + output = self.replicate.run(self.model, input=input_payload) + + # Common shapes: [url, ...] or dict-like. + image_url = None + if isinstance(output, list) and output: + image_url = output[0] + elif isinstance(output, dict): + image_url = output.get("image") or output.get("output") or output.get("url") + if not isinstance(image_url, str) or not image_url: + raise RuntimeError(f"Unexpected Replicate output shape: {type(output)}") + + r = requests.get(image_url, timeout=60) + r.raise_for_status() + + # Always output PNG to satisfy downstream validation `outputs/{task_id}/*.png`. + out_path = output_dir / f"shot_{uuid.uuid4().hex}.png" + # Pillow doesn't provide open_bytes; wrap content into a buffer. + from io import BytesIO + + img = Image.open(BytesIO(r.content)).convert("RGB") + img.save(str(out_path), format="PNG") + return str(out_path) + diff --git a/engine/adapters/image/stability_adapter.py b/engine/adapters/image/stability_adapter.py new file mode 100644 index 0000000..3ef1278 --- /dev/null +++ b/engine/adapters/image/stability_adapter.py @@ -0,0 +1,21 @@ +from __future__ import annotations + +from pathlib import Path + +from engine.config import AppConfig + +from .base import BaseImageGen + + +class StabilityAdapter(BaseImageGen): + """ + Placeholder for Stability AI image generation. + Add implementation + dependencies when needed. + """ + + def __init__(self, cfg: AppConfig): + self.cfg = cfg + + def generate(self, prompt: dict[str, str], output_dir: str | Path) -> str: + raise NotImplementedError("StabilityAdapter not implemented yet") + diff --git a/engine/adapters/llm/__init__.py b/engine/adapters/llm/__init__.py new file mode 100644 index 0000000..8b13789 --- /dev/null +++ b/engine/adapters/llm/__init__.py @@ -0,0 +1 @@ + diff --git a/engine/adapters/llm/base.py b/engine/adapters/llm/base.py new file mode 100644 index 0000000..f0090eb --- /dev/null +++ b/engine/adapters/llm/base.py @@ -0,0 +1,12 @@ +from __future__ import annotations + +from typing import Any + + +class BaseLLM: + def generate_script(self, prompt: str, context: dict[str, Any] | None = None) -> Any: + raise NotImplementedError + + def refine_scene(self, scene: Any, context: dict[str, Any] | None = None) -> Any: + raise NotImplementedError + diff --git a/engine/adapters/llm/mock_adapter.py b/engine/adapters/llm/mock_adapter.py new file mode 100644 index 0000000..17fb8a7 --- /dev/null +++ b/engine/adapters/llm/mock_adapter.py @@ -0,0 +1,25 @@ +from __future__ import annotations + +from typing import Any + +from engine.types import Scene + +from .base import BaseLLM + + +class MockLLM(BaseLLM): + def generate_script(self, prompt: str, context: dict[str, Any] | None = None) -> list[Scene]: + # Simple deterministic scenes for offline development. + prompt = (prompt or "").strip() + if not prompt: + prompt = "a warm city night" + return [ + Scene(image_prompt=f"{prompt}，城市夜景，霓虹灯，电影感", video_motion="缓慢推进镜头，轻微摇镜", narration="夜色温柔落在街灯上"), + Scene(image_prompt=f"{prompt}，咖啡店窗边，暖光，细雨", video_motion="侧向平移，人物轻轻抬头", narration="雨声里藏着一段回忆"), + Scene(image_prompt=f"{prompt}，桥上远景，车流光轨，温暖", video_motion="拉远全景，光轨流动", narration="我们在光里学会告别"), + ] + + def refine_scene(self, scene: Scene, context: dict[str, Any] | None = None) -> Scene: + # Minimal polish: append a hint. + return Scene(image_prompt=scene.image_prompt, video_motion=scene.video_motion, narration=(scene.narration + "（更凝练）")[:30]) + diff --git a/engine/adapters/llm/openai_adapter.py b/engine/adapters/llm/openai_adapter.py new file mode 100644 index 0000000..5c8c25c --- /dev/null +++ b/engine/adapters/llm/openai_adapter.py @@ -0,0 +1,29 @@ +from __future__ import annotations + +from typing import Any + +from engine.config import AppConfig +from engine.script_gen import generate_scenes, refine_scene + +from .base import BaseLLM + + +class OpenAIAdapter(BaseLLM): + def __init__(self, cfg: AppConfig): + self.cfg = cfg + + def generate_script(self, prompt: str, context: dict[str, Any] | None = None): + # Existing script_gen already enforces JSON schema and length constraints. + return generate_scenes(prompt, self.cfg) + + def refine_scene(self, scene: Any, context: dict[str, Any] | None = None): + if context is None: + context = {} + # Context carries needed values to call refine_scene in script_gen. + scenes = context.get("scenes") + prompt2 = context.get("prompt") + target_index = context.get("target_index") + if scenes is None or prompt2 is None or target_index is None: + raise ValueError("OpenAIAdapter.refine_scene missing context: scenes/prompt/target_index") + return refine_scene(prompt=prompt2, scenes=scenes, target_index=int(target_index), cfg=self.cfg) + diff --git a/engine/adapters/tts/__init__.py b/engine/adapters/tts/__init__.py new file mode 100644 index 0000000..8b13789 --- /dev/null +++ b/engine/adapters/tts/__init__.py @@ -0,0 +1 @@ + diff --git a/engine/adapters/tts/base.py b/engine/adapters/tts/base.py new file mode 100644 index 0000000..cdbcd07 --- /dev/null +++ b/engine/adapters/tts/base.py @@ -0,0 +1,9 @@ +from __future__ import annotations + +from pathlib import Path + + +class BaseTTS: + def generate(self, text: str, output_path: str | Path) -> str: + raise NotImplementedError + diff --git a/engine/adapters/tts/edge_adapter.py b/engine/adapters/tts/edge_adapter.py new file mode 100644 index 0000000..860c58d --- /dev/null +++ b/engine/adapters/tts/edge_adapter.py @@ -0,0 +1,28 @@ +from __future__ import annotations + +import asyncio +from pathlib import Path + +from engine.audio_gen import synthesize_one +from engine.config import AppConfig + +from .base import BaseTTS + + +class EdgeTTS(BaseTTS): + def __init__(self, cfg: AppConfig): + self.cfg = cfg + + def generate(self, text: str, output_path: str | Path) -> str: + text = text or " " + output_path = Path(output_path) + voice = str(self.cfg.get("tts.voice", "zh-CN-XiaoxiaoNeural")) + rate = str(self.cfg.get("tts.rate", "+0%")) + volume = str(self.cfg.get("tts.volume", "+0%")) + + async def _run(): + asset = await synthesize_one(text, output_path, voice, rate, volume) + return str(asset.path) + + return asyncio.run(_run()) + diff --git a/engine/adapters/tts/mock_adapter.py b/engine/adapters/tts/mock_adapter.py new file mode 100644 index 0000000..5a3c3ed --- /dev/null +++ b/engine/adapters/tts/mock_adapter.py @@ -0,0 +1,15 @@ +from __future__ import annotations + +from pathlib import Path + +from .base import BaseTTS + + +class MockTTS(BaseTTS): + def generate(self, text: str, output_path: str | Path) -> str: + # No-op for offline tests: return empty path so video adapter skips audio. + output_path = Path(output_path) + output_path.parent.mkdir(parents=True, exist_ok=True) + output_path.write_bytes(b"") + return str(output_path) + diff --git a/engine/adapters/video/__init__.py b/engine/adapters/video/__init__.py new file mode 100644 index 0000000..8b13789 --- /dev/null +++ b/engine/adapters/video/__init__.py @@ -0,0 +1 @@ + diff --git a/engine/adapters/video/base.py b/engine/adapters/video/base.py new file mode 100644 index 0000000..840db8b --- /dev/null +++ b/engine/adapters/video/base.py @@ -0,0 +1,9 @@ +from __future__ import annotations + +from pathlib import Path + + +class BaseVideoGen: + def generate(self, image_path: str, prompt: dict, output_path: str | Path) -> str: + raise NotImplementedError + diff --git a/engine/adapters/video/ltx_adapter.py b/engine/adapters/video/ltx_adapter.py new file mode 100644 index 0000000..83b5c24 --- /dev/null +++ b/engine/adapters/video/ltx_adapter.py @@ -0,0 +1,18 @@ +from __future__ import annotations + +from pathlib import Path + +from engine.config import AppConfig + +from .base import BaseVideoGen + + +class LTXVideoGen(BaseVideoGen): + def __init__(self, cfg: AppConfig): + self.cfg = cfg + + def generate(self, image_path: str, prompt: dict, output_path: str | Path) -> str: + # Reserved for future: direct image->video generation (LTX / diffusion video). + # Current project keeps clip generation via MoviePy for stability. + raise NotImplementedError("LTXVideoGen is not implemented yet") + diff --git a/engine/adapters/video/moviepy_adapter.py b/engine/adapters/video/moviepy_adapter.py new file mode 100644 index 0000000..847fe2e --- /dev/null +++ b/engine/adapters/video/moviepy_adapter.py @@ -0,0 +1,81 @@ +from __future__ import annotations + +import os +from pathlib import Path +from typing import Any + +import numpy as np +from moviepy import AudioFileClip, VideoClip +from PIL import Image + +from engine.config import AppConfig + +from .base import BaseVideoGen + + +class MoviePyVideoGen(BaseVideoGen): + def __init__(self, cfg: AppConfig): + self.cfg = cfg + + def generate(self, image_path: str, prompt: dict, output_path: str | Path) -> str: + output_path = Path(output_path) + output_path.parent.mkdir(parents=True, exist_ok=True) + + # Required prompt fields for shot rendering. + duration_s = float(prompt.get("duration_s", 3)) + fps = int(prompt.get("fps", self.cfg.get("video.mock_fps", 24))) + audio_path = prompt.get("audio_path") + + # Clip resolution. + size = prompt.get("size") + if isinstance(size, (list, tuple)) and len(size) == 2: + w, h = int(size[0]), int(size[1]) + else: + mock_size = self.cfg.get("video.mock_size", [1024, 576]) + w, h = int(mock_size[0]), int(mock_size[1]) + + base_img = Image.open(image_path).convert("RGB") + + def make_frame(t: float): + progress = float(t) / max(duration_s, 1e-6) + progress = max(0.0, min(1.0, progress)) + scale = 1.0 + 0.03 * progress + new_w = max(w, int(w * scale)) + new_h = max(h, int(h * scale)) + frame = base_img.resize((new_w, new_h), Image.LANCZOS) + left = (new_w - w) // 2 + top = (new_h - h) // 2 + frame = frame.crop((left, top, left + w, top + h)) + return np.array(frame) + + video = VideoClip(make_frame, duration=duration_s, has_constant_size=True) + + # Optional audio. + if audio_path and os.path.exists(str(audio_path)): + a = AudioFileClip(str(audio_path)) + video = video.with_audio(a) + else: + a = None + + try: + video.write_videofile( + str(output_path), + fps=fps, + codec="libx264", + audio_codec="aac", + preset="veryfast", + threads=2, + ) + finally: + try: + video.close() + except Exception: + pass + if a is not None: + try: + a.close() + except Exception: + pass + + return str(output_path) + diff --git a/engine/comfy_client.py b/engine/comfy_client.py index bc7005c..624837c 100644 --- a/engine/comfy_client.py +++ b/engine/comfy_client.py @@ -2,6 +2,7 @@ from __future__ import annotations import asyncio import json +import time import uuid from dataclasses import dataclass from pathlib import Path @@ -186,3 +187,215 @@ class ComfyClient: # unreachable # return ComfyResult(prompt_id=prompt_id, output_files=last_files) + + +# --------------------------------------------------------------------------- +# Minimal "text->image" helpers (used by shot rendering) +# --------------------------------------------------------------------------- + + +def _build_simple_workflow( + prompt_text: str, + *, + seed: int, + ckpt_name: str, + width: int, + height: int, + steps: int = 20, + cfg: float = 8.0, + sampler_name: str = "euler", + scheduler: str = "normal", + denoise: float = 1.0, + filename_prefix: str = "shot", + negative_text: str = "low quality, blurry", +) -> dict[str, Any]: + # Best-effort workflow. If your ComfyUI nodes/models differ, generation must fallback. + return { + "3": { + "class_type": "KSampler", + "inputs": { + "seed": int(seed), + "steps": int(steps), + "cfg": float(cfg), + "sampler_name": sampler_name, + "scheduler": scheduler, + "denoise": float(denoise), + "model": ["4", 0], + "positive": ["6", 0], + "negative": ["7", 0], + "latent_image": ["5", 0], + }, + }, + "4": { + "class_type": "CheckpointLoaderSimple", + "inputs": { + "ckpt_name": ckpt_name, + }, + }, + "5": { + "class_type": "EmptyLatentImage", + "inputs": { + "width": int(width), + "height": int(height), + "batch_size": 1, + }, + }, + "6": { + "class_type": "CLIPTextEncode", + "inputs": { + "text": prompt_text, + "clip": ["4", 1], + }, + }, + "7": { + "class_type": "CLIPTextEncode", + "inputs": { + "text": negative_text, + "clip": ["4", 1], + }, + }, + "8": { + "class_type": "VAEDecode", + "inputs": { + "samples": ["3", 0], + "vae": ["4", 2], + }, + }, + "9": { + "class_type": "SaveImage", + "inputs": { + "images": ["8", 0], + "filename_prefix": filename_prefix, + }, + }, + } + + +def _queue_prompt(base_url: str, workflow: dict[str, Any], client_id: str) -> str: + r = httpx.post( + base_url.rstrip("/") + "/prompt", + json={"prompt": workflow, "client_id": client_id}, + timeout=30.0, + ) + r.raise_for_status() + data = r.json() + pid = data.get("prompt_id") + if not isinstance(pid, str) or not pid: + raise RuntimeError(f"Unexpected /prompt response: {data}") + return pid + + +def _get_history_item(base_url: str, prompt_id: str) -> dict[str, Any] | None: + for url in (f"{base_url.rstrip('/')}/history/{prompt_id}", f"{base_url.rstrip('/')}/history"): + try: + r = httpx.get(url, timeout=30.0) + if r.status_code == 404: + continue + r.raise_for_status() + data = r.json() + if isinstance(data, dict): + if prompt_id in data and isinstance(data[prompt_id], dict): + return data[prompt_id] + if url.endswith(f"/{prompt_id}") and isinstance(data, dict): + return data + return None + except Exception: + continue + return None + + +def _extract_first_image_view_target(history_item: dict[str, Any]) -> tuple[str, str] | None: + outputs = history_item.get("outputs") + if not isinstance(outputs, dict): + return None + + def walk(v: Any) -> list[dict[str, Any]]: + found: list[dict[str, Any]] = [] + if isinstance(v, dict): + if isinstance(v.get("filename"), str) and v.get("filename").strip(): + found.append(v) + for vv in v.values(): + found.extend(walk(vv)) + elif isinstance(v, list): + for vv in v: + found.extend(walk(vv)) + return found + + candidates = walk(outputs) + for c in candidates: + fn = str(c.get("filename", "")).strip() + sf = str(c.get("subfolder", "") or "").strip() + if fn: + return fn, sf + return None + + +def generate_image( + prompt_text: str, + output_dir: str | Path, + *, + cfg: AppConfig | None = None, + timeout_s: int = 60, + retry: int = 2, + width: int | None = None, + height: int | None = None, + filename_prefix: str = "shot", + ckpt_candidates: list[str] | None = None, + negative_text: str | None = None, +) -> Path: + cfg2 = cfg or AppConfig.load("./configs/config.yaml") + base_url = str(cfg2.get("app.comfy_base_url", "http://comfyui:8188")).rstrip("/") + + out_dir = Path(output_dir) + out_dir.mkdir(parents=True, exist_ok=True) + + if width is None or height is None: + mock_size = cfg2.get("video.mock_size", [1024, 576]) + width = int(width or mock_size[0]) + height = int(height or mock_size[1]) + + if negative_text is None: + negative_text = "low quality, blurry" + + if ckpt_candidates is None: + ckpt_candidates = [ + "v1-5-pruned-emaonly.ckpt", + "v1-5-pruned-emaonly.safetensors", + "sd-v1-5-tiny.safetensors", + ] + + last_err: Exception | None = None + for _attempt in range(max(1, retry)): + for ckpt_name in ckpt_candidates: + client_id = str(uuid.uuid4()) + seed = int(uuid.uuid4().int % 2_147_483_647) + workflow = _build_simple_workflow( + prompt_text, + seed=seed, + ckpt_name=ckpt_name, + width=width, + height=height, + filename_prefix=filename_prefix, + negative_text=negative_text, + ) + try: + prompt_id = _queue_prompt(base_url, workflow, client_id) + start = time.time() + while time.time() - start < timeout_s: + item = _get_history_item(base_url, prompt_id) + if isinstance(item, dict): + img_target = _extract_first_image_view_target(item) + if img_target: + filename, subfolder = img_target + view_url = f"{base_url}/view?filename={filename}&subfolder={subfolder}" + img_resp = httpx.get(view_url, timeout=60.0) + img_resp.raise_for_status() + image_path = out_dir / filename + image_path.write_bytes(img_resp.content) + return image_path + time.sleep(1.0) + except Exception as e: + last_err = e + continue + + raise RuntimeError(f"ComfyUI image generation failed after retries: {last_err}") diff --git a/engine/main.py b/engine/main.py index 2c9bf28..9c80e17 100644 --- a/engine/main.py +++ b/engine/main.py @@ -12,13 +12,14 @@ from typing import Any from moviepy import ImageClip from PIL import Image, ImageDraw, ImageFont -from engine.audio_gen import synthesize_scenes +from engine.model_factory import get_model +from engine.prompt_injector import inject_prompt +from engine.adapters.image.mock_adapter import MockImageGen from engine.assembler import assemble_clips from engine.comfy_client import ComfyClient from engine.config import AppConfig from engine.director import scenes_to_shots from engine.shot_executor import render_shot -from engine.script_gen import generate_scenes, refine_scene from engine.task_store import create_task, update_shot_status, update_task_status from engine.types import Scene from engine.video_editor import Segment, render_final @@ -28,13 +29,15 @@ def _emit(line: str) -> None: print(line, flush=True) -def _emit_scene(scene_idx: int, scene: Scene) -> None: +def _emit_scene(scene_idx: int, scene: Scene, extra: dict[str, Any] | None = None) -> None: payload = { "index": scene_idx, "image_prompt": scene.image_prompt, "video_motion": scene.video_motion, "narration": scene.narration, } + if extra: + payload.update(extra) _emit("SCENE_JSON " + json.dumps(payload, ensure_ascii=False)) @@ -136,9 +139,50 @@ def _fallback_scenes(prompt: str) -> list[Scene]: ] +def _generate_scene_preview( + *, + cfg: AppConfig, + out_dir: Path, + image_prompt: str, + style: str | None, + character: str | None, +) -> str | None: + try: + image_gen = get_model("image", cfg) + except Exception: + image_gen = get_model("image_fallback", cfg) + + global_cfg = dict(cfg.get("global", {}) or {}) + if style: + global_cfg["style"] = style + if character: + global_cfg["character"] = character + + prompt_obj = inject_prompt(global_cfg, {"prompt": image_prompt}) + try: + image_path = image_gen.generate(prompt_obj, out_dir) + except Exception: + try: + image_path = get_model("image_fallback", cfg).generate(prompt_obj, out_dir) + except Exception: + # Last-resort hard fallback: never block script stage due to preview failures. + image_path = MockImageGen().generate(prompt_obj, out_dir) + + p = Path(str(image_path)) + if not p.exists(): + return None + return f"/api/static/{out_dir.name}/{p.name}" + + def _has_llm_key(cfg: AppConfig) -> bool: - api_key_env = str(cfg.get("openai.api_key_env", "OPENAI_API_KEY")) - return bool(os.environ.get(api_key_env)) + api_key_env = str(cfg.get("openai.api_key_env", "OPENAI_API_KEY") or "OPENAI_API_KEY").strip() + # Env var name case. + if os.environ.get(api_key_env): + return True + # Literal key case (DashScope / OpenAI-compatible). + if api_key_env.startswith("sk-"): + return True + return False def _parse_scenes_from_obj(obj: Any) -> list[Scene]: @@ -239,7 +283,8 @@ def step_script(prompt: str, cfg: AppConfig, mock: bool, *, style: str | None, c # fallback scenes still should include global injection scenes = _fallback_scenes(prompt) else: - scenes = generate_scenes(prompt2, cfg) + llm = get_model("llm", cfg) + scenes = llm.generate_script(prompt2, context=None) out_dir.mkdir(parents=True, exist_ok=True) _emit("SCRIPT_BEGIN") @@ -249,7 +294,14 @@ def step_script(prompt: str, cfg: AppConfig, mock: bool, *, style: str | None, c video_motion=s.video_motion, narration=s.narration, ) - _emit_scene(idx, s2) + preview_url = _generate_scene_preview( + cfg=cfg, + out_dir=out_dir, + image_prompt=s2.image_prompt, + style=style, + character=character, + ) + _emit_scene(idx, s2, extra={"preview_url": preview_url or ""}) _emit("SCRIPT_END") (out_dir / "scenes.json").write_text( json.dumps( @@ -292,8 +344,9 @@ def step_refine( narration=(s.narration + "（更凝练）")[:30], ) else: - # Ensure globals are visible to LLM, and inject to output image prompt. - refined0 = refine_scene(prompt=prompt2, scenes=scenes, target_index=target_index, cfg=cfg) + llm = get_model("llm", cfg) + # Context carries prompt + scenes for consistent refinement. + refined0 = llm.refine_scene(scenes[target_index - 1], context={"prompt": prompt2, "scenes": scenes, "target_index": target_index}) refined = Scene( image_prompt=_decorate_image_prompt(refined0.image_prompt, style=style, character=character), video_motion=refined0.video_motion, @@ -301,7 +354,14 @@ def step_refine( ) # Keep the original index for frontend replacement. - _emit_scene(scene_index, refined) + preview_url = _generate_scene_preview( + cfg=cfg, + out_dir=out_dir, + image_prompt=refined.image_prompt, + style=style, + character=character, + ) + _emit_scene(scene_index, refined, extra={"preview_url": preview_url or ""}) out_dir.mkdir(parents=True, exist_ok=True) (out_dir / f"refine_scene_{scene_index}.json").write_text( json.dumps( diff --git a/engine/model_factory.py b/engine/model_factory.py new file mode 100644 index 0000000..ca9479a --- /dev/null +++ b/engine/model_factory.py @@ -0,0 +1,80 @@ +from __future__ import annotations + +import os +from typing import Any + +from engine.config import AppConfig + + +def _provider(cfg: AppConfig, path: str, default: str) -> str: + env_map = { + "llm.provider": "ENGINE_LLM_PROVIDER", + "image.provider": "ENGINE_IMAGE_PROVIDER", + "image_fallback.provider": "ENGINE_IMAGE_FALLBACK_PROVIDER", + "video.provider": "ENGINE_VIDEO_PROVIDER", + "tts.provider": "ENGINE_TTS_PROVIDER", + } + env_key = env_map.get(path) + if env_key: + env_val = str(os.environ.get(env_key, "")).strip() + if env_val: + return env_val + v = cfg.get(path, default) + return str(v or default).strip() or default + + +def get_model(name: str, cfg: AppConfig) -> Any: + if name == "llm": + provider = _provider(cfg, "llm.provider", "openai") + if provider == "mock": + from engine.adapters.llm.mock_adapter import MockLLM + + return MockLLM() + from engine.adapters.llm.openai_adapter import OpenAIAdapter + + return OpenAIAdapter(cfg) + + if name in ("image", "image_fallback"): + section = "image" if name == "image" else "image_fallback" + # Important: fallback must default to mock, not follow primary image provider. + provider_default = "mock" if name == "image_fallback" else _provider(cfg, "image.provider", "mock") + provider = _provider(cfg, f"{section}.provider", provider_default) + if provider == "comfy": + from engine.adapters.image.comfy_adapter import ComfyAdapter + + return ComfyAdapter(cfg) + if provider == "replicate": + from engine.adapters.image.replicate_adapter import ReplicateAdapter + + return ReplicateAdapter(cfg) + if provider == "openai": + from engine.adapters.image.openai_image_adapter import OpenAIImageAdapter + + return OpenAIImageAdapter(cfg) + + from engine.adapters.image.mock_adapter import MockImageGen + + return MockImageGen() + + if name == "video": + provider = _provider(cfg, "video.provider", "moviepy") + if provider == "ltx": + from engine.adapters.video.ltx_adapter import LTXVideoGen + + return LTXVideoGen(cfg) + from engine.adapters.video.moviepy_adapter import MoviePyVideoGen + + return MoviePyVideoGen(cfg) + + if name == "tts": + provider = _provider(cfg, "tts.provider", "edge") + if provider == "mock": + from engine.adapters.tts.mock_adapter import MockTTS + + return MockTTS() + from engine.adapters.tts.edge_adapter import EdgeTTS + + return EdgeTTS(cfg) + + raise ValueError(f"Unknown model adapter name: {name}") + diff --git a/engine/prompt_injector.py b/engine/prompt_injector.py new file mode 100644 index 0000000..de4ac5e --- /dev/null +++ b/engine/prompt_injector.py @@ -0,0 +1,23 @@ +from __future__ import annotations + +from typing import Any + + +def inject_prompt(global_cfg: dict[str, Any] | None, scene: dict[str, Any]) -> dict[str, str]: + """ + Unified positive/negative prompt builder. + Note: current pipeline already injects some globals into `scene["image_prompt"]`. + """ + global_cfg = global_cfg or {} + character = str(global_cfg.get("character", "") or "").strip() + style = str(global_cfg.get("style", "") or "").strip() + negative = str(global_cfg.get("negative_prompt", "") or "").strip() + + base = str(scene.get("prompt") or scene.get("image_prompt") or "").strip() + if not base: + base = str(scene.get("image_prompt") or "") + + positive_parts = [p for p in [character, style, base] if p] + positive = ", ".join(positive_parts).strip(", ") + return {"positive": positive, "negative": negative} + diff --git a/engine/render_pipeline.py b/engine/render_pipeline.py new file mode 100644 index 0000000..8dd62f0 --- /dev/null +++ b/engine/render_pipeline.py @@ -0,0 +1,80 @@ +from __future__ import annotations + +from pathlib import Path +from typing import Any + +from engine.model_factory import get_model +from engine.prompt_injector import inject_prompt +from engine.adapters.image.mock_adapter import MockImageGen + + +def render_shot(shot: dict[str, Any], cfg, out_dir: str | Path, *, mock: bool = False) -> str: + out_dir = Path(out_dir) + clips_dir = out_dir / "clips" + audio_dir = out_dir / "audio" + clips_dir.mkdir(parents=True, exist_ok=True) + audio_dir.mkdir(parents=True, exist_ok=True) + + shot_id = str(shot.get("shot_id", "unknown")) + duration_s = float(shot.get("duration", 3)) + narration = str(shot.get("tts", "")).strip() + + # Models from config. + image_fallback_gen = get_model("image_fallback", cfg) + try: + image_gen = get_model("image", cfg) + except Exception as e: + # Covers missing optional deps at adapter init time (e.g. replicate/openai packages). + print(f"[WARN] image provider init failed, fallback to image_fallback: {e}") + image_gen = image_fallback_gen + tts = get_model("tts", cfg) + video_gen = get_model("video", cfg) + + # Prompt injection. + global_cfg = cfg.get("global", {}) if hasattr(cfg, "get") else {} + prompt_obj = inject_prompt(global_cfg, {"prompt": shot.get("image_prompt", "")}) + positive_prompt = prompt_obj.get("positive", "") + # Prompt enrichment: keeps ComfyUI generations cinematic and detailed. + enrich_style = "cinematic, ultra realistic, 4k, detailed lighting" + if enrich_style not in positive_prompt: + positive_prompt = f"{positive_prompt}, {enrich_style}".strip(", ") + prompt_obj["positive"] = positive_prompt + + # 1) image + try: + image_path = image_gen.generate(prompt_obj, out_dir) + except Exception as e: + # Config-driven fallback; keeps provider switching non-invasive. + print(f"[WARN] Image generation failed, fallback to image_fallback: {e}") + try: + image_path = image_fallback_gen.generate(prompt_obj, out_dir) + except Exception as e2: + print(f"[WARN] image_fallback also failed, hard fallback to mock: {e2}") + image_path = MockImageGen().generate(prompt_obj, out_dir) + + scene_label = str(shot.get("scene_id") or shot.get("shot_id") or "scene_unknown") + print(f"[SHOT_RENDER] {scene_label} -> image generated: {image_path}") + + # 2) audio (optional) + audio_path = None + if narration: + # Use a stable per-shot audio filename. + ap = audio_dir / f"shot_{shot_id}.mp3" + try: + audio_path = tts.generate(narration, ap) + except Exception as e: + # Don't fail the whole render due to TTS issues. + print(f"[WARN] TTS failed, continue without audio: {e}") + audio_path = None + + # 3) clip + clip_out = clips_dir / f"shot_{shot_id}.mp4" + prompt = { + "duration_s": duration_s, + "fps": int(cfg.get("video.mock_fps", 24)), + "audio_path": audio_path, + "size": cfg.get("video.mock_size", None), + } + clip_path = video_gen.generate(image_path, prompt, clip_out) + return clip_path + diff --git a/engine/script_gen.py b/engine/script_gen.py index 3238cf6..5b67b95 100644 --- a/engine/script_gen.py +++ b/engine/script_gen.py @@ -10,6 +10,38 @@ from .config import AppConfig from .types import Scene +def _looks_like_api_key(v: str) -> bool: + vv = (v or "").strip() + # Common prefixes: DashScope uses "sk-..."; we keep it minimal and permissive. + return bool(vv) and vv.startswith("sk-") + + +def _looks_like_url(v: str) -> bool: + vv = (v or "").strip() + return vv.startswith("http://") or vv.startswith("https://") + + +def _resolve_openai_credentials(cfg: AppConfig) -> tuple[str, str | None]: + api_key_env = str(cfg.get("openai.api_key_env", "OPENAI_API_KEY") or "").strip() + base_url_env = str(cfg.get("openai.base_url_env", "OPENAI_BASE_URL") or "").strip() + + # 1) Resolve api_key: allow both "env var name" and "literal key" for safety. + api_key = os.environ.get(api_key_env) if api_key_env else None + if not api_key and api_key_env and _looks_like_api_key(api_key_env): + api_key = api_key_env + + if not api_key: + raise RuntimeError(f"Missing OpenAI compatible API key (env={api_key_env})") + + # 2) Resolve base_url: allow both "env var name" and "literal URL". + base_url = os.environ.get(base_url_env) if base_url_env else None + if not base_url and base_url_env and _looks_like_url(base_url_env): + base_url = base_url_env + if base_url: + base_url = str(base_url).strip() or None + return str(api_key), base_url + + def _system_prompt(scene_count: int, min_chars: int, max_chars: int) -> str: return f"""你是一个专业短视频编剧与分镜师。 请把用户的创意扩展为 {scene_count} 个分镜(Scene) 的 JSON。 @@ -56,17 +88,13 @@ def generate_scenes(user_prompt: str, cfg: AppConfig) -> list[Scene]: min_chars = int(cfg.get("script_gen.narration_min_chars", 15)) max_chars = int(cfg.get("script_gen.narration_max_chars", 20)) - api_key_env = str(cfg.get("openai.api_key_env", "OPENAI_API_KEY")) - base_url_env = str(cfg.get("openai.base_url_env", "OPENAI_BASE_URL")) model = str(cfg.get("openai.model", "gpt-4o-mini")) - api_key = os.environ.get(api_key_env) - if not api_key: - raise RuntimeError(f"Missing env var {api_key_env} for OpenAI API key") + api_key, base_url = _resolve_openai_credentials(cfg) client = OpenAI( api_key=api_key, - base_url=os.environ.get(base_url_env) or None, + base_url=base_url, ) resp = client.chat.completions.create( @@ -105,17 +133,13 @@ def refine_scene(*, prompt: str, scenes: list[Scene], target_index: int, cfg: Ap min_chars = int(cfg.get("script_gen.narration_min_chars", 15)) max_chars = int(cfg.get("script_gen.narration_max_chars", 20)) - api_key_env = str(cfg.get("openai.api_key_env", "OPENAI_API_KEY")) - base_url_env = str(cfg.get("openai.base_url_env", "OPENAI_BASE_URL")) model = str(cfg.get("openai.model", "gpt-4o-mini")) - api_key = os.environ.get(api_key_env) - if not api_key: - raise RuntimeError(f"Missing env var {api_key_env} for OpenAI API key") + api_key, base_url = _resolve_openai_credentials(cfg) client = OpenAI( api_key=api_key, - base_url=os.environ.get(base_url_env) or None, + base_url=base_url, ) scenes_payload = [ diff --git a/engine/shot_executor.py b/engine/shot_executor.py index 99bc95c..77b0dde 100644 --- a/engine/shot_executor.py +++ b/engine/shot_executor.py @@ -1,42 +1,53 @@ from __future__ import annotations import asyncio +import os import random from pathlib import Path from typing import Any -from moviepy import AudioFileClip, CompositeVideoClip, TextClip, VideoFileClip, vfx +import numpy as np +from moviepy import AudioFileClip, VideoClip +from PIL import Image +from urllib.request import urlopen from .audio_gen import synthesize_one -from .comfy_client import ComfyClient +from .comfy_client import generate_image as comfy_generate_image from .config import AppConfig +from .render_pipeline import render_shot as render_shot_pipeline -def _fit_video_to_audio(video: VideoFileClip, audio: AudioFileClip) -> VideoFileClip: - if audio.duration is None or video.duration is None: - return video.with_audio(audio) - if audio.duration > video.duration: - video = video.with_effects([vfx.Loop(duration=audio.duration)]) - elif video.duration > audio.duration: - video = video.subclipped(0, audio.duration) - return video.with_audio(audio) +ASSETS_DIR = "assets" +DEMO_IMAGE = os.path.join(ASSETS_DIR, "demo.jpg") -def _subtitle_clip(text: str, size: tuple[int, int], duration: float) -> TextClip: - return ( - TextClip( - text=text, - font_size=44, - color="white", - stroke_color="black", - stroke_width=2, - size=(int(size[0] * 0.92), None), - method="caption", - ) - .with_position(("center", "bottom")) - .with_duration(duration) - .with_opacity(0.95) - ) +def ensure_demo_image() -> None: + os.makedirs(ASSETS_DIR, exist_ok=True) + if os.path.exists(DEMO_IMAGE): + return + + # Simple placeholder image source. + url = "https://picsum.photos/1280/720" + with urlopen(url, timeout=30) as resp: + data = resp.read() + + with open(DEMO_IMAGE, "wb") as f: + f.write(data) + + +def generate_image_mock(prompt: str) -> str: + # Keep interface compatible with the requested interface. + _ = prompt + ensure_demo_image() + return DEMO_IMAGE + + +def enrich_prompt(prompt_text: str) -> str: + style = "cinematic, ultra realistic, 4k, detailed lighting" + pt = (prompt_text or "").strip() + if not pt: + return style + return f"{pt}, {style}" async def _render_shot_async( @@ -55,49 +66,102 @@ async def _render_shot_async( shot_id = str(shot.get("shot_id", "unknown")) image_prompt = str(shot.get("image_prompt", "")).strip() - motion = str(shot.get("motion", "")).strip() + prompt_text = str(shot.get("prompt", image_prompt) or image_prompt).strip() tts_text = str(shot.get("tts", "")).strip() duration_s = max(1.0, float(shot.get("duration", 3))) voice = str(cfg.get("tts.voice", "zh-CN-XiaoxiaoNeural")) rate = str(cfg.get("tts.rate", "+0%")) volume = str(cfg.get("tts.volume", "+0%")) - audio_path = audio_dir / f"shot_{shot_id}.mp3" - audio_asset = await synthesize_one(tts_text or " ", audio_path, voice, rate, volume) + audio_asset: Any | None = None + if tts_text: + audio_path = audio_dir / f"shot_{shot_id}.mp3" + audio_asset = await synthesize_one(tts_text, audio_path, voice, rate, volume) + # Use config-defined output resolution for stable concatenation. + mock_size = cfg.get("video.mock_size", [1024, 576]) + w, h = int(mock_size[0]), int(mock_size[1]) + fps = int(cfg.get("video.mock_fps", 24)) + + if audio_asset and audio_asset.duration_s: + duration_s = max(duration_s, float(audio_asset.duration_s)) + + # shot -> image (ComfyUI first; fallback to demo.jpg) + image_path: str if mock: - from engine.main import _ensure_mock_image, _make_mock_video # local import to avoid circular at module import - - mock_size = cfg.get("video.mock_size", [1024, 576]) - w, h = int(mock_size[0]), int(mock_size[1]) - mock_image = _ensure_mock_image(Path("./assets/mock.png"), (w, h)) - fps = int(cfg.get("video.mock_fps", 24)) - raw_video_path = out_dir / f"shot_raw_{shot_id}.mp4" - _make_mock_video(raw_video_path, mock_image, max(duration_s, audio_asset.duration_s), fps=fps) + image_path = generate_image_mock(prompt_text) else: - comfy = ComfyClient(cfg) - wf = comfy.load_workflow() - seed = random.randint(1, 2_147_483_647) - wf_i = comfy.inject_params(wf, image_prompt=image_prompt, seed=seed, motion_prompt=motion or None) - result = await comfy.run_workflow(wf_i) - candidates = [p for p in result.output_files if p.suffix.lower() in {".mp4", ".mov", ".webm"}] - raw_video_path = candidates[0] if candidates else result.output_files[0] - - clip_out = clips_dir / f"shot_{shot_id}.mp4" - v = VideoFileClip(str(raw_video_path)) - a = AudioFileClip(str(audio_asset.path)) - try: - v2 = _fit_video_to_audio(v, a) - w2, h2 = v2.size - subtitle = _subtitle_clip(tts_text, (w2, h2), v2.duration or a.duration or duration_s) - comp = CompositeVideoClip([v2, subtitle]) try: - comp.write_videofile(str(clip_out), codec="libx264", audio_codec="aac", fps=v2.fps or 24, preset="veryfast") - finally: - comp.close() + enriched = enrich_prompt(prompt_text) + # Store generated images directly under outputs/{task_id} + # (as required by verification: outputs/{task_id}/*.png). + image_path = str( + comfy_generate_image( + enriched, + out_dir, + cfg=cfg, + timeout_s=60, + retry=2, + filename_prefix=f"shot_{shot_id}", + ) + ) + print(f"[SHOT_RENDER] {shot_id} -> image generated: {image_path}") + except Exception as e: + print(f"[WARN] Comfy failed, fallback to demo: {e}") + image_path = generate_image_mock(prompt_text) + + # Ensure image exists before rendering. + if not image_path or not os.path.exists(image_path): + image_path = generate_image_mock(prompt_text) + base_img = Image.open(image_path).convert("RGB") + + def make_frame(t: float): + # Subtle zoom-in from 1.00 to ~1.03 over the clip duration. + progress = float(t) / max(duration_s, 1e-6) + progress = max(0.0, min(1.0, progress)) + scale = 1.0 + 0.03 * progress + + new_w = max(w, int(w * scale)) + new_h = max(h, int(h * scale)) + + frame = base_img.resize((new_w, new_h), Image.LANCZOS) + left = (new_w - w) // 2 + top = (new_h - h) // 2 + frame = frame.crop((left, top, left + w, top + h)) + return np.array(frame) + + # image -> video + video = VideoClip(make_frame, duration=duration_s, has_constant_size=True) + + # optional audio -> clip + audio_clip: AudioFileClip | None = None + if audio_asset and os.path.exists(str(audio_asset.path)): + audio_clip = AudioFileClip(str(audio_asset.path)) + video = video.with_audio(audio_clip) + + # output + clip_out = clips_dir / f"shot_{shot_id}.mp4" + print(f"[SHOT_RENDER] {shot_id} -> {clip_out}") + try: + video.write_videofile( + str(clip_out), + fps=fps, + codec="libx264", + audio_codec="aac", + preset="veryfast", + threads=2, + ) finally: - v.close() - a.close() + try: + video.close() + except Exception: + pass + if audio_clip is not None: + try: + audio_clip.close() + except Exception: + pass + return str(clip_out) @@ -109,5 +173,5 @@ def render_shot( mock: bool = False, ) -> str: cfg2 = cfg or AppConfig.load("./configs/config.yaml") - return asyncio.run(_render_shot_async(shot, output_dir, cfg2, mock=mock)) + return render_shot_pipeline(shot, cfg2, output_dir, mock=mock) diff --git a/outputs/'06b0a90f-c964-4a88-8e80-6ff668e031b3'/audio/shot_scene_01_01.mp3 b/outputs/'06b0a90f-c964-4a88-8e80-6ff668e031b3'/audio/shot_scene_01_01.mp3 new file mode 100644 index 0000000..36c5dc2 Binary files /dev/null and b/outputs/'06b0a90f-c964-4a88-8e80-6ff668e031b3'/audio/shot_scene_01_01.mp3 differ diff --git a/outputs/'06b0a90f-c964-4a88-8e80-6ff668e031b3'/task.json b/outputs/'06b0a90f-c964-4a88-8e80-6ff668e031b3'/task.json new file mode 100644 index 0000000..1467223 --- /dev/null +++ b/outputs/'06b0a90f-c964-4a88-8e80-6ff668e031b3'/task.json @@ -0,0 +1,18 @@ +{ + "task_id": "'06b0a90f-c964-4a88-8e80-6ff668e031b3'", + "status": "failed", + "shots": [ + { + "shot_id": "scene_01_01", + "status": "running" + }, + { + "shot_id": "scene_02_01", + "status": "pending" + }, + { + "shot_id": "scene_03_01", + "status": "pending" + } + ] +} \ No newline at end of file diff --git a/outputs/'13c9b724-77e3-4553-aebf-dfc845dd17c1'/audio/shot_scene_01_01.mp3 b/outputs/'13c9b724-77e3-4553-aebf-dfc845dd17c1'/audio/shot_scene_01_01.mp3 new file mode 100644 index 0000000..36c5dc2 Binary files /dev/null and b/outputs/'13c9b724-77e3-4553-aebf-dfc845dd17c1'/audio/shot_scene_01_01.mp3 differ diff --git a/outputs/'13c9b724-77e3-4553-aebf-dfc845dd17c1'/audio/shot_scene_02_01.mp3 b/outputs/'13c9b724-77e3-4553-aebf-dfc845dd17c1'/audio/shot_scene_02_01.mp3 new file mode 100644 index 0000000..cee3639 Binary files /dev/null and b/outputs/'13c9b724-77e3-4553-aebf-dfc845dd17c1'/audio/shot_scene_02_01.mp3 differ diff --git a/outputs/'13c9b724-77e3-4553-aebf-dfc845dd17c1'/audio/shot_scene_03_01.mp3 b/outputs/'13c9b724-77e3-4553-aebf-dfc845dd17c1'/audio/shot_scene_03_01.mp3 new file mode 100644 index 0000000..e9e8d2a Binary files /dev/null and b/outputs/'13c9b724-77e3-4553-aebf-dfc845dd17c1'/audio/shot_scene_03_01.mp3 differ diff --git a/outputs/'13c9b724-77e3-4553-aebf-dfc845dd17c1'/clips/shot_scene_01_01.mp4 b/outputs/'13c9b724-77e3-4553-aebf-dfc845dd17c1'/clips/shot_scene_01_01.mp4 new file mode 100644 index 0000000..c9faece Binary files /dev/null and b/outputs/'13c9b724-77e3-4553-aebf-dfc845dd17c1'/clips/shot_scene_01_01.mp4 differ diff --git a/outputs/'13c9b724-77e3-4553-aebf-dfc845dd17c1'/clips/shot_scene_02_01.mp4 b/outputs/'13c9b724-77e3-4553-aebf-dfc845dd17c1'/clips/shot_scene_02_01.mp4 new file mode 100644 index 0000000..681f666 Binary files /dev/null and b/outputs/'13c9b724-77e3-4553-aebf-dfc845dd17c1'/clips/shot_scene_02_01.mp4 differ diff --git a/outputs/'13c9b724-77e3-4553-aebf-dfc845dd17c1'/clips/shot_scene_03_01.mp4 b/outputs/'13c9b724-77e3-4553-aebf-dfc845dd17c1'/clips/shot_scene_03_01.mp4 new file mode 100644 index 0000000..9c2119b Binary files /dev/null and b/outputs/'13c9b724-77e3-4553-aebf-dfc845dd17c1'/clips/shot_scene_03_01.mp4 differ diff --git a/outputs/'13c9b724-77e3-4553-aebf-dfc845dd17c1'/final.mp4 b/outputs/'13c9b724-77e3-4553-aebf-dfc845dd17c1'/final.mp4 new file mode 100644 index 0000000..c4f4c32 Binary files /dev/null and b/outputs/'13c9b724-77e3-4553-aebf-dfc845dd17c1'/final.mp4 differ diff --git a/outputs/'13c9b724-77e3-4553-aebf-dfc845dd17c1'/task.json b/outputs/'13c9b724-77e3-4553-aebf-dfc845dd17c1'/task.json new file mode 100644 index 0000000..debb37e --- /dev/null +++ b/outputs/'13c9b724-77e3-4553-aebf-dfc845dd17c1'/task.json @@ -0,0 +1,18 @@ +{ + "task_id": "'13c9b724-77e3-4553-aebf-dfc845dd17c1'", + "status": "done", + "shots": [ + { + "shot_id": "scene_01_01", + "status": "done" + }, + { + "shot_id": "scene_02_01", + "status": "done" + }, + { + "shot_id": "scene_03_01", + "status": "done" + } + ] +} \ No newline at end of file diff --git a/outputs/0d546f5e-0274-4372-b91d-fb64ace85d49/scenes.json b/outputs/0d546f5e-0274-4372-b91d-fb64ace85d49/scenes.json new file mode 100644 index 0000000..5b08e9f --- /dev/null +++ b/outputs/0d546f5e-0274-4372-b91d-fb64ace85d49/scenes.json @@ -0,0 +1,19 @@ +{ + "scenes": [ + { + "image_prompt": "写一个温暖的城市夜景故事，城市夜景，霓虹灯，电影感", + "video_motion": "缓慢推进镜头，轻微摇镜", + "narration": "夜色温柔落在街灯上" + }, + { + "image_prompt": "写一个温暖的城市夜景故事，咖啡店窗边，暖光，细雨", + "video_motion": "侧向平移，人物轻轻抬头", + "narration": "雨声里藏着一段回忆" + }, + { + "image_prompt": "写一个温暖的城市夜景故事，桥上远景，车流光轨，温暖", + "video_motion": "拉远全景，光轨流动", + "narration": "我们在光里学会告别" + } + ] +} \ No newline at end of file diff --git a/outputs/3ef0c0b8-c90f-49a8-88e4-e8ca735312f0/audio/shot_scene_01_01.mp3 b/outputs/3ef0c0b8-c90f-49a8-88e4-e8ca735312f0/audio/shot_scene_01_01.mp3 new file mode 100644 index 0000000..36c5dc2 Binary files /dev/null and b/outputs/3ef0c0b8-c90f-49a8-88e4-e8ca735312f0/audio/shot_scene_01_01.mp3 differ diff --git a/outputs/3ef0c0b8-c90f-49a8-88e4-e8ca735312f0/audio/shot_scene_02_01.mp3 b/outputs/3ef0c0b8-c90f-49a8-88e4-e8ca735312f0/audio/shot_scene_02_01.mp3 new file mode 100644 index 0000000..cee3639 Binary files /dev/null and b/outputs/3ef0c0b8-c90f-49a8-88e4-e8ca735312f0/audio/shot_scene_02_01.mp3 differ diff --git a/outputs/3ef0c0b8-c90f-49a8-88e4-e8ca735312f0/audio/shot_scene_03_01.mp3 b/outputs/3ef0c0b8-c90f-49a8-88e4-e8ca735312f0/audio/shot_scene_03_01.mp3 new file mode 100644 index 0000000..e9e8d2a Binary files /dev/null and b/outputs/3ef0c0b8-c90f-49a8-88e4-e8ca735312f0/audio/shot_scene_03_01.mp3 differ diff --git a/outputs/3ef0c0b8-c90f-49a8-88e4-e8ca735312f0/clips/shot_scene_01_01.mp4 b/outputs/3ef0c0b8-c90f-49a8-88e4-e8ca735312f0/clips/shot_scene_01_01.mp4 new file mode 100644 index 0000000..a78a1c3 Binary files /dev/null and b/outputs/3ef0c0b8-c90f-49a8-88e4-e8ca735312f0/clips/shot_scene_01_01.mp4 differ diff --git a/outputs/3ef0c0b8-c90f-49a8-88e4-e8ca735312f0/clips/shot_scene_02_01.mp4 b/outputs/3ef0c0b8-c90f-49a8-88e4-e8ca735312f0/clips/shot_scene_02_01.mp4 new file mode 100644 index 0000000..5a3388e Binary files /dev/null and b/outputs/3ef0c0b8-c90f-49a8-88e4-e8ca735312f0/clips/shot_scene_02_01.mp4 differ diff --git a/outputs/3ef0c0b8-c90f-49a8-88e4-e8ca735312f0/clips/shot_scene_03_01.mp4 b/outputs/3ef0c0b8-c90f-49a8-88e4-e8ca735312f0/clips/shot_scene_03_01.mp4 new file mode 100644 index 0000000..5390725 Binary files /dev/null and b/outputs/3ef0c0b8-c90f-49a8-88e4-e8ca735312f0/clips/shot_scene_03_01.mp4 differ diff --git a/outputs/3ef0c0b8-c90f-49a8-88e4-e8ca735312f0/final.mp4 b/outputs/3ef0c0b8-c90f-49a8-88e4-e8ca735312f0/final.mp4 new file mode 100644 index 0000000..1af94fe Binary files /dev/null and b/outputs/3ef0c0b8-c90f-49a8-88e4-e8ca735312f0/final.mp4 differ diff --git a/outputs/3ef0c0b8-c90f-49a8-88e4-e8ca735312f0/scenes.json b/outputs/3ef0c0b8-c90f-49a8-88e4-e8ca735312f0/scenes.json new file mode 100644 index 0000000..5b08e9f --- /dev/null +++ b/outputs/3ef0c0b8-c90f-49a8-88e4-e8ca735312f0/scenes.json @@ -0,0 +1,19 @@ +{ + "scenes": [ + { + "image_prompt": "写一个温暖的城市夜景故事，城市夜景，霓虹灯，电影感", + "video_motion": "缓慢推进镜头，轻微摇镜", + "narration": "夜色温柔落在街灯上" + }, + { + "image_prompt": "写一个温暖的城市夜景故事，咖啡店窗边，暖光，细雨", + "video_motion": "侧向平移，人物轻轻抬头", + "narration": "雨声里藏着一段回忆" + }, + { + "image_prompt": "写一个温暖的城市夜景故事，桥上远景，车流光轨，温暖", + "video_motion": "拉远全景，光轨流动", + "narration": "我们在光里学会告别" + } + ] +} \ No newline at end of file diff --git a/outputs/3ef0c0b8-c90f-49a8-88e4-e8ca735312f0/shot_raw_scene_01_01.mp4 b/outputs/3ef0c0b8-c90f-49a8-88e4-e8ca735312f0/shot_raw_scene_01_01.mp4 new file mode 100644 index 0000000..fe4e343 Binary files /dev/null and b/outputs/3ef0c0b8-c90f-49a8-88e4-e8ca735312f0/shot_raw_scene_01_01.mp4 differ diff --git a/outputs/3ef0c0b8-c90f-49a8-88e4-e8ca735312f0/shot_raw_scene_02_01.mp4 b/outputs/3ef0c0b8-c90f-49a8-88e4-e8ca735312f0/shot_raw_scene_02_01.mp4 new file mode 100644 index 0000000..fe4e343 Binary files /dev/null and b/outputs/3ef0c0b8-c90f-49a8-88e4-e8ca735312f0/shot_raw_scene_02_01.mp4 differ diff --git a/outputs/3ef0c0b8-c90f-49a8-88e4-e8ca735312f0/shot_raw_scene_03_01.mp4 b/outputs/3ef0c0b8-c90f-49a8-88e4-e8ca735312f0/shot_raw_scene_03_01.mp4 new file mode 100644 index 0000000..fe4e343 Binary files /dev/null and b/outputs/3ef0c0b8-c90f-49a8-88e4-e8ca735312f0/shot_raw_scene_03_01.mp4 differ diff --git a/outputs/3ef0c0b8-c90f-49a8-88e4-e8ca735312f0/task.json b/outputs/3ef0c0b8-c90f-49a8-88e4-e8ca735312f0/task.json new file mode 100644 index 0000000..81a28f5 --- /dev/null +++ b/outputs/3ef0c0b8-c90f-49a8-88e4-e8ca735312f0/task.json @@ -0,0 +1,18 @@ +{ + "task_id": "3ef0c0b8-c90f-49a8-88e4-e8ca735312f0", + "status": "done", + "shots": [ + { + "shot_id": "scene_01_01", + "status": "done" + }, + { + "shot_id": "scene_02_01", + "status": "done" + }, + { + "shot_id": "scene_03_01", + "status": "done" + } + ] +} \ No newline at end of file diff --git a/outputs/3f82b1ce-da18-4f82-9147-25eb0abeaf2c/audio/shot_scene_01_01.mp3 b/outputs/3f82b1ce-da18-4f82-9147-25eb0abeaf2c/audio/shot_scene_01_01.mp3 new file mode 100644 index 0000000..36c5dc2 Binary files /dev/null and b/outputs/3f82b1ce-da18-4f82-9147-25eb0abeaf2c/audio/shot_scene_01_01.mp3 differ diff --git a/outputs/3f82b1ce-da18-4f82-9147-25eb0abeaf2c/clips/shot_scene_01_01.mp4 b/outputs/3f82b1ce-da18-4f82-9147-25eb0abeaf2c/clips/shot_scene_01_01.mp4 new file mode 100644 index 0000000..c9faece Binary files /dev/null and b/outputs/3f82b1ce-da18-4f82-9147-25eb0abeaf2c/clips/shot_scene_01_01.mp4 differ diff --git a/outputs/3f82b1ce-da18-4f82-9147-25eb0abeaf2c/final.mp4 b/outputs/3f82b1ce-da18-4f82-9147-25eb0abeaf2c/final.mp4 new file mode 100644 index 0000000..8f18ee1 Binary files /dev/null and b/outputs/3f82b1ce-da18-4f82-9147-25eb0abeaf2c/final.mp4 differ diff --git a/outputs/3f82b1ce-da18-4f82-9147-25eb0abeaf2c/task.json b/outputs/3f82b1ce-da18-4f82-9147-25eb0abeaf2c/task.json new file mode 100644 index 0000000..0447844 --- /dev/null +++ b/outputs/3f82b1ce-da18-4f82-9147-25eb0abeaf2c/task.json @@ -0,0 +1,10 @@ +{ + "task_id": "3f82b1ce-da18-4f82-9147-25eb0abeaf2c", + "status": "done", + "shots": [ + { + "shot_id": "scene_01_01", + "status": "done" + } + ] +} \ No newline at end of file diff --git a/outputs/62da5541-43d2-4ead-a243-e68345877dff/audio/shot_scene_01_01.mp3 b/outputs/62da5541-43d2-4ead-a243-e68345877dff/audio/shot_scene_01_01.mp3 new file mode 100644 index 0000000..36c5dc2 Binary files /dev/null and b/outputs/62da5541-43d2-4ead-a243-e68345877dff/audio/shot_scene_01_01.mp3 differ diff --git a/outputs/62da5541-43d2-4ead-a243-e68345877dff/audio/shot_scene_02_01.mp3 b/outputs/62da5541-43d2-4ead-a243-e68345877dff/audio/shot_scene_02_01.mp3 new file mode 100644 index 0000000..cee3639 Binary files /dev/null and b/outputs/62da5541-43d2-4ead-a243-e68345877dff/audio/shot_scene_02_01.mp3 differ diff --git a/outputs/62da5541-43d2-4ead-a243-e68345877dff/audio/shot_scene_03_01.mp3 b/outputs/62da5541-43d2-4ead-a243-e68345877dff/audio/shot_scene_03_01.mp3 new file mode 100644 index 0000000..e9e8d2a Binary files /dev/null and b/outputs/62da5541-43d2-4ead-a243-e68345877dff/audio/shot_scene_03_01.mp3 differ diff --git a/outputs/62da5541-43d2-4ead-a243-e68345877dff/clips/shot_scene_01_01.mp4 b/outputs/62da5541-43d2-4ead-a243-e68345877dff/clips/shot_scene_01_01.mp4 new file mode 100644 index 0000000..c9faece Binary files /dev/null and b/outputs/62da5541-43d2-4ead-a243-e68345877dff/clips/shot_scene_01_01.mp4 differ diff --git a/outputs/62da5541-43d2-4ead-a243-e68345877dff/clips/shot_scene_02_01.mp4 b/outputs/62da5541-43d2-4ead-a243-e68345877dff/clips/shot_scene_02_01.mp4 new file mode 100644 index 0000000..681f666 Binary files /dev/null and b/outputs/62da5541-43d2-4ead-a243-e68345877dff/clips/shot_scene_02_01.mp4 differ diff --git a/outputs/62da5541-43d2-4ead-a243-e68345877dff/clips/shot_scene_03_01.mp4 b/outputs/62da5541-43d2-4ead-a243-e68345877dff/clips/shot_scene_03_01.mp4 new file mode 100644 index 0000000..9c2119b Binary files /dev/null and b/outputs/62da5541-43d2-4ead-a243-e68345877dff/clips/shot_scene_03_01.mp4 differ diff --git a/outputs/62da5541-43d2-4ead-a243-e68345877dff/final.mp4 b/outputs/62da5541-43d2-4ead-a243-e68345877dff/final.mp4 new file mode 100644 index 0000000..c4f4c32 Binary files /dev/null and b/outputs/62da5541-43d2-4ead-a243-e68345877dff/final.mp4 differ diff --git a/outputs/62da5541-43d2-4ead-a243-e68345877dff/task.json b/outputs/62da5541-43d2-4ead-a243-e68345877dff/task.json new file mode 100644 index 0000000..0d09ca5 --- /dev/null +++ b/outputs/62da5541-43d2-4ead-a243-e68345877dff/task.json @@ -0,0 +1,18 @@ +{ + "task_id": "62da5541-43d2-4ead-a243-e68345877dff", + "status": "done", + "shots": [ + { + "shot_id": "scene_01_01", + "status": "done" + }, + { + "shot_id": "scene_02_01", + "status": "done" + }, + { + "shot_id": "scene_03_01", + "status": "done" + } + ] +} \ No newline at end of file diff --git a/outputs/7b8255ea-ed2f-4356-8a57-d5c77e351351/audio/shot_scene_01_01.mp3 b/outputs/7b8255ea-ed2f-4356-8a57-d5c77e351351/audio/shot_scene_01_01.mp3 new file mode 100644 index 0000000..36c5dc2 Binary files /dev/null and b/outputs/7b8255ea-ed2f-4356-8a57-d5c77e351351/audio/shot_scene_01_01.mp3 differ diff --git a/outputs/7b8255ea-ed2f-4356-8a57-d5c77e351351/audio/shot_scene_02_01.mp3 b/outputs/7b8255ea-ed2f-4356-8a57-d5c77e351351/audio/shot_scene_02_01.mp3 new file mode 100644 index 0000000..cee3639 Binary files /dev/null and b/outputs/7b8255ea-ed2f-4356-8a57-d5c77e351351/audio/shot_scene_02_01.mp3 differ diff --git a/outputs/7b8255ea-ed2f-4356-8a57-d5c77e351351/audio/shot_scene_03_01.mp3 b/outputs/7b8255ea-ed2f-4356-8a57-d5c77e351351/audio/shot_scene_03_01.mp3 new file mode 100644 index 0000000..e9e8d2a Binary files /dev/null and b/outputs/7b8255ea-ed2f-4356-8a57-d5c77e351351/audio/shot_scene_03_01.mp3 differ diff --git a/outputs/7b8255ea-ed2f-4356-8a57-d5c77e351351/clips/shot_scene_01_01.mp4 b/outputs/7b8255ea-ed2f-4356-8a57-d5c77e351351/clips/shot_scene_01_01.mp4 new file mode 100644 index 0000000..c9faece Binary files /dev/null and b/outputs/7b8255ea-ed2f-4356-8a57-d5c77e351351/clips/shot_scene_01_01.mp4 differ diff --git a/outputs/7b8255ea-ed2f-4356-8a57-d5c77e351351/clips/shot_scene_02_01.mp4 b/outputs/7b8255ea-ed2f-4356-8a57-d5c77e351351/clips/shot_scene_02_01.mp4 new file mode 100644 index 0000000..681f666 Binary files /dev/null and b/outputs/7b8255ea-ed2f-4356-8a57-d5c77e351351/clips/shot_scene_02_01.mp4 differ diff --git a/outputs/7b8255ea-ed2f-4356-8a57-d5c77e351351/clips/shot_scene_03_01.mp4 b/outputs/7b8255ea-ed2f-4356-8a57-d5c77e351351/clips/shot_scene_03_01.mp4 new file mode 100644 index 0000000..9c2119b Binary files /dev/null and b/outputs/7b8255ea-ed2f-4356-8a57-d5c77e351351/clips/shot_scene_03_01.mp4 differ diff --git a/outputs/7b8255ea-ed2f-4356-8a57-d5c77e351351/final.mp4 b/outputs/7b8255ea-ed2f-4356-8a57-d5c77e351351/final.mp4 new file mode 100644 index 0000000..c4f4c32 Binary files /dev/null and b/outputs/7b8255ea-ed2f-4356-8a57-d5c77e351351/final.mp4 differ diff --git a/outputs/7b8255ea-ed2f-4356-8a57-d5c77e351351/scenes.json b/outputs/7b8255ea-ed2f-4356-8a57-d5c77e351351/scenes.json new file mode 100644 index 0000000..75b4740 --- /dev/null +++ b/outputs/7b8255ea-ed2f-4356-8a57-d5c77e351351/scenes.json @@ -0,0 +1,19 @@ +{ + "scenes": [ + { + "image_prompt": "写一个温暖的城市夜景故事\n\n\n[Global Constraints]\n- Global Style: 电影感\n请严格遵守上述全局信息，并保持三分镜主角一致。，城市夜景，霓虹灯，电影感", + "video_motion": "缓慢推进镜头，轻微摇镜", + "narration": "夜色温柔落在街灯上" + }, + { + "image_prompt": "写一个温暖的城市夜景故事\n\n\n[Global Constraints]\n- Global Style: 电影感\n请严格遵守上述全局信息，并保持三分镜主角一致。，咖啡店窗边，暖光，细雨", + "video_motion": "侧向平移，人物轻轻抬头", + "narration": "雨声里藏着一段回忆" + }, + { + "image_prompt": "写一个温暖的城市夜景故事\n\n\n[Global Constraints]\n- Global Style: 电影感\n请严格遵守上述全局信息，并保持三分镜主角一致。，桥上远景，车流光轨，温暖", + "video_motion": "拉远全景，光轨流动", + "narration": "我们在光里学会告别" + } + ] +} \ No newline at end of file diff --git a/outputs/7b8255ea-ed2f-4356-8a57-d5c77e351351/shot_10743d29878a41dd9e5c8b6b5c84a743.png b/outputs/7b8255ea-ed2f-4356-8a57-d5c77e351351/shot_10743d29878a41dd9e5c8b6b5c84a743.png new file mode 100644 index 0000000..4c86bf9 Binary files /dev/null and b/outputs/7b8255ea-ed2f-4356-8a57-d5c77e351351/shot_10743d29878a41dd9e5c8b6b5c84a743.png differ diff --git a/outputs/7b8255ea-ed2f-4356-8a57-d5c77e351351/shot_2fef6555f3f34a42b0e100b01cd4c281.png b/outputs/7b8255ea-ed2f-4356-8a57-d5c77e351351/shot_2fef6555f3f34a42b0e100b01cd4c281.png new file mode 100644 index 0000000..4c86bf9 Binary files /dev/null and b/outputs/7b8255ea-ed2f-4356-8a57-d5c77e351351/shot_2fef6555f3f34a42b0e100b01cd4c281.png differ diff --git a/outputs/7b8255ea-ed2f-4356-8a57-d5c77e351351/shot_42485af317eb4e888efaaa55ee66cd33.png b/outputs/7b8255ea-ed2f-4356-8a57-d5c77e351351/shot_42485af317eb4e888efaaa55ee66cd33.png new file mode 100644 index 0000000..4c86bf9 Binary files /dev/null and b/outputs/7b8255ea-ed2f-4356-8a57-d5c77e351351/shot_42485af317eb4e888efaaa55ee66cd33.png differ diff --git a/outputs/7b8255ea-ed2f-4356-8a57-d5c77e351351/shot_6aa8cc90ce644ab88c6c022a9ac71168.png b/outputs/7b8255ea-ed2f-4356-8a57-d5c77e351351/shot_6aa8cc90ce644ab88c6c022a9ac71168.png new file mode 100644 index 0000000..4c86bf9 Binary files /dev/null and b/outputs/7b8255ea-ed2f-4356-8a57-d5c77e351351/shot_6aa8cc90ce644ab88c6c022a9ac71168.png differ diff --git a/outputs/7b8255ea-ed2f-4356-8a57-d5c77e351351/shot_d39b04ab653b4496ada1ac9385f0abac.png b/outputs/7b8255ea-ed2f-4356-8a57-d5c77e351351/shot_d39b04ab653b4496ada1ac9385f0abac.png new file mode 100644 index 0000000..4c86bf9 Binary files /dev/null and b/outputs/7b8255ea-ed2f-4356-8a57-d5c77e351351/shot_d39b04ab653b4496ada1ac9385f0abac.png differ diff --git a/outputs/7b8255ea-ed2f-4356-8a57-d5c77e351351/shot_fdd6eb4905fe4644a9ff4140dcff7251.png b/outputs/7b8255ea-ed2f-4356-8a57-d5c77e351351/shot_fdd6eb4905fe4644a9ff4140dcff7251.png new file mode 100644 index 0000000..4c86bf9 Binary files /dev/null and b/outputs/7b8255ea-ed2f-4356-8a57-d5c77e351351/shot_fdd6eb4905fe4644a9ff4140dcff7251.png differ diff --git a/outputs/7b8255ea-ed2f-4356-8a57-d5c77e351351/task.json b/outputs/7b8255ea-ed2f-4356-8a57-d5c77e351351/task.json new file mode 100644 index 0000000..c03e34a --- /dev/null +++ b/outputs/7b8255ea-ed2f-4356-8a57-d5c77e351351/task.json @@ -0,0 +1,18 @@ +{ + "task_id": "7b8255ea-ed2f-4356-8a57-d5c77e351351", + "status": "done", + "shots": [ + { + "shot_id": "scene_01_01", + "status": "done" + }, + { + "shot_id": "scene_02_01", + "status": "done" + }, + { + "shot_id": "scene_03_01", + "status": "done" + } + ] +} \ No newline at end of file diff --git a/outputs/a77ef82b-81ea-4ff5-b592-db0ebb047df4/scenes.json b/outputs/a77ef82b-81ea-4ff5-b592-db0ebb047df4/scenes.json new file mode 100644 index 0000000..75b4740 --- /dev/null +++ b/outputs/a77ef82b-81ea-4ff5-b592-db0ebb047df4/scenes.json @@ -0,0 +1,19 @@ +{ + "scenes": [ + { + "image_prompt": "写一个温暖的城市夜景故事\n\n\n[Global Constraints]\n- Global Style: 电影感\n请严格遵守上述全局信息，并保持三分镜主角一致。，城市夜景，霓虹灯，电影感", + "video_motion": "缓慢推进镜头，轻微摇镜", + "narration": "夜色温柔落在街灯上" + }, + { + "image_prompt": "写一个温暖的城市夜景故事\n\n\n[Global Constraints]\n- Global Style: 电影感\n请严格遵守上述全局信息，并保持三分镜主角一致。，咖啡店窗边，暖光，细雨", + "video_motion": "侧向平移，人物轻轻抬头", + "narration": "雨声里藏着一段回忆" + }, + { + "image_prompt": "写一个温暖的城市夜景故事\n\n\n[Global Constraints]\n- Global Style: 电影感\n请严格遵守上述全局信息，并保持三分镜主角一致。，桥上远景，车流光轨，温暖", + "video_motion": "拉远全景，光轨流动", + "narration": "我们在光里学会告别" + } + ] +} \ No newline at end of file diff --git a/outputs/a77ef82b-81ea-4ff5-b592-db0ebb047df4/shot_0e00ee0e06544cd49f00ba68a65a68d2.png b/outputs/a77ef82b-81ea-4ff5-b592-db0ebb047df4/shot_0e00ee0e06544cd49f00ba68a65a68d2.png new file mode 100644 index 0000000..4c86bf9 Binary files /dev/null and b/outputs/a77ef82b-81ea-4ff5-b592-db0ebb047df4/shot_0e00ee0e06544cd49f00ba68a65a68d2.png differ diff --git a/outputs/a77ef82b-81ea-4ff5-b592-db0ebb047df4/shot_349333bf77ca465a93e9ecc6f09ddde1.png b/outputs/a77ef82b-81ea-4ff5-b592-db0ebb047df4/shot_349333bf77ca465a93e9ecc6f09ddde1.png new file mode 100644 index 0000000..4c86bf9 Binary files /dev/null and b/outputs/a77ef82b-81ea-4ff5-b592-db0ebb047df4/shot_349333bf77ca465a93e9ecc6f09ddde1.png differ diff --git a/outputs/a77ef82b-81ea-4ff5-b592-db0ebb047df4/shot_6a6376abbc2449b2a6646a24064d0430.png b/outputs/a77ef82b-81ea-4ff5-b592-db0ebb047df4/shot_6a6376abbc2449b2a6646a24064d0430.png new file mode 100644 index 0000000..4c86bf9 Binary files /dev/null and b/outputs/a77ef82b-81ea-4ff5-b592-db0ebb047df4/shot_6a6376abbc2449b2a6646a24064d0430.png differ diff --git a/outputs/ab4bb47e-cf35-4a99-977f-63097bdac9ed/scenes.json b/outputs/ab4bb47e-cf35-4a99-977f-63097bdac9ed/scenes.json new file mode 100644 index 0000000..5b08e9f --- /dev/null +++ b/outputs/ab4bb47e-cf35-4a99-977f-63097bdac9ed/scenes.json @@ -0,0 +1,19 @@ +{ + "scenes": [ + { + "image_prompt": "写一个温暖的城市夜景故事，城市夜景，霓虹灯，电影感", + "video_motion": "缓慢推进镜头，轻微摇镜", + "narration": "夜色温柔落在街灯上" + }, + { + "image_prompt": "写一个温暖的城市夜景故事，咖啡店窗边，暖光，细雨", + "video_motion": "侧向平移，人物轻轻抬头", + "narration": "雨声里藏着一段回忆" + }, + { + "image_prompt": "写一个温暖的城市夜景故事，桥上远景，车流光轨，温暖", + "video_motion": "拉远全景，光轨流动", + "narration": "我们在光里学会告别" + } + ] +} \ No newline at end of file diff --git a/outputs/ab68ccf6-0de0-4465-b4d7-1843f88d0201/audio/shot_scene_01_01.mp3 b/outputs/ab68ccf6-0de0-4465-b4d7-1843f88d0201/audio/shot_scene_01_01.mp3 new file mode 100644 index 0000000..36c5dc2 Binary files /dev/null and b/outputs/ab68ccf6-0de0-4465-b4d7-1843f88d0201/audio/shot_scene_01_01.mp3 differ diff --git a/outputs/ab68ccf6-0de0-4465-b4d7-1843f88d0201/audio/shot_scene_02_01.mp3 b/outputs/ab68ccf6-0de0-4465-b4d7-1843f88d0201/audio/shot_scene_02_01.mp3 new file mode 100644 index 0000000..cee3639 Binary files /dev/null and b/outputs/ab68ccf6-0de0-4465-b4d7-1843f88d0201/audio/shot_scene_02_01.mp3 differ diff --git a/outputs/ab68ccf6-0de0-4465-b4d7-1843f88d0201/audio/shot_scene_03_01.mp3 b/outputs/ab68ccf6-0de0-4465-b4d7-1843f88d0201/audio/shot_scene_03_01.mp3 new file mode 100644 index 0000000..e9e8d2a Binary files /dev/null and b/outputs/ab68ccf6-0de0-4465-b4d7-1843f88d0201/audio/shot_scene_03_01.mp3 differ diff --git a/outputs/ab68ccf6-0de0-4465-b4d7-1843f88d0201/clips/shot_scene_01_01.mp4 b/outputs/ab68ccf6-0de0-4465-b4d7-1843f88d0201/clips/shot_scene_01_01.mp4 new file mode 100644 index 0000000..c9faece Binary files /dev/null and b/outputs/ab68ccf6-0de0-4465-b4d7-1843f88d0201/clips/shot_scene_01_01.mp4 differ diff --git a/outputs/ab68ccf6-0de0-4465-b4d7-1843f88d0201/clips/shot_scene_02_01.mp4 b/outputs/ab68ccf6-0de0-4465-b4d7-1843f88d0201/clips/shot_scene_02_01.mp4 new file mode 100644 index 0000000..681f666 Binary files /dev/null and b/outputs/ab68ccf6-0de0-4465-b4d7-1843f88d0201/clips/shot_scene_02_01.mp4 differ diff --git a/outputs/ab68ccf6-0de0-4465-b4d7-1843f88d0201/clips/shot_scene_03_01.mp4 b/outputs/ab68ccf6-0de0-4465-b4d7-1843f88d0201/clips/shot_scene_03_01.mp4 new file mode 100644 index 0000000..9c2119b Binary files /dev/null and b/outputs/ab68ccf6-0de0-4465-b4d7-1843f88d0201/clips/shot_scene_03_01.mp4 differ diff --git a/outputs/ab68ccf6-0de0-4465-b4d7-1843f88d0201/final.mp4 b/outputs/ab68ccf6-0de0-4465-b4d7-1843f88d0201/final.mp4 new file mode 100644 index 0000000..c4f4c32 Binary files /dev/null and b/outputs/ab68ccf6-0de0-4465-b4d7-1843f88d0201/final.mp4 differ diff --git a/outputs/ab68ccf6-0de0-4465-b4d7-1843f88d0201/scenes.json b/outputs/ab68ccf6-0de0-4465-b4d7-1843f88d0201/scenes.json new file mode 100644 index 0000000..75b4740 --- /dev/null +++ b/outputs/ab68ccf6-0de0-4465-b4d7-1843f88d0201/scenes.json @@ -0,0 +1,19 @@ +{ + "scenes": [ + { + "image_prompt": "写一个温暖的城市夜景故事\n\n\n[Global Constraints]\n- Global Style: 电影感\n请严格遵守上述全局信息，并保持三分镜主角一致。，城市夜景，霓虹灯，电影感", + "video_motion": "缓慢推进镜头，轻微摇镜", + "narration": "夜色温柔落在街灯上" + }, + { + "image_prompt": "写一个温暖的城市夜景故事\n\n\n[Global Constraints]\n- Global Style: 电影感\n请严格遵守上述全局信息，并保持三分镜主角一致。，咖啡店窗边，暖光，细雨", + "video_motion": "侧向平移，人物轻轻抬头", + "narration": "雨声里藏着一段回忆" + }, + { + "image_prompt": "写一个温暖的城市夜景故事\n\n\n[Global Constraints]\n- Global Style: 电影感\n请严格遵守上述全局信息，并保持三分镜主角一致。，桥上远景，车流光轨，温暖", + "video_motion": "拉远全景，光轨流动", + "narration": "我们在光里学会告别" + } + ] +} \ No newline at end of file diff --git a/outputs/ab68ccf6-0de0-4465-b4d7-1843f88d0201/shot_225525ed518042e99d6cf7b430e126e0.png b/outputs/ab68ccf6-0de0-4465-b4d7-1843f88d0201/shot_225525ed518042e99d6cf7b430e126e0.png new file mode 100644 index 0000000..4c86bf9 Binary files /dev/null and b/outputs/ab68ccf6-0de0-4465-b4d7-1843f88d0201/shot_225525ed518042e99d6cf7b430e126e0.png differ diff --git a/outputs/ab68ccf6-0de0-4465-b4d7-1843f88d0201/shot_66e3402393964b22a11bc3b06459989d.png b/outputs/ab68ccf6-0de0-4465-b4d7-1843f88d0201/shot_66e3402393964b22a11bc3b06459989d.png new file mode 100644 index 0000000..4c86bf9 Binary files /dev/null and b/outputs/ab68ccf6-0de0-4465-b4d7-1843f88d0201/shot_66e3402393964b22a11bc3b06459989d.png differ diff --git a/outputs/ab68ccf6-0de0-4465-b4d7-1843f88d0201/shot_8d1b1091edd94171a2795aabf6637f5f.png b/outputs/ab68ccf6-0de0-4465-b4d7-1843f88d0201/shot_8d1b1091edd94171a2795aabf6637f5f.png new file mode 100644 index 0000000..4c86bf9 Binary files /dev/null and b/outputs/ab68ccf6-0de0-4465-b4d7-1843f88d0201/shot_8d1b1091edd94171a2795aabf6637f5f.png differ diff --git a/outputs/ab68ccf6-0de0-4465-b4d7-1843f88d0201/task.json b/outputs/ab68ccf6-0de0-4465-b4d7-1843f88d0201/task.json new file mode 100644 index 0000000..d0ad240 --- /dev/null +++ b/outputs/ab68ccf6-0de0-4465-b4d7-1843f88d0201/task.json @@ -0,0 +1,18 @@ +{ + "task_id": "ab68ccf6-0de0-4465-b4d7-1843f88d0201", + "status": "done", + "shots": [ + { + "shot_id": "scene_01_01", + "status": "done" + }, + { + "shot_id": "scene_02_01", + "status": "done" + }, + { + "shot_id": "scene_03_01", + "status": "done" + } + ] +} \ No newline at end of file diff --git a/outputs/b4d67faf-fabe-4ead-8803-7bfbb7ee4ad2/scenes.json b/outputs/b4d67faf-fabe-4ead-8803-7bfbb7ee4ad2/scenes.json new file mode 100644 index 0000000..dda94aa --- /dev/null +++ b/outputs/b4d67faf-fabe-4ead-8803-7bfbb7ee4ad2/scenes.json @@ -0,0 +1,19 @@ +{ + "scenes": [ + { + "image_prompt": "Cinematic night shot, wet street reflecting neon lights, Xiao Lin walking away, beige trench coat, white scarf, cold tone background, bokeh.", + "video_motion": "镜头缓慢跟随背影移动，雨丝飘落。", + "narration": "霓虹灯下城市结束喧嚣，夜色格外温柔。" + }, + { + "image_prompt": "Medium shot inside convenience store, warm yellow lighting, Xiao Lin holding hot coffee, steam rising, soft facial lighting, cinematic depth of field.", + "video_motion": "镜头缓缓推进，捕捉蒸汽升腾动态。", + "narration": "街角便利店的灯光，是深夜里最暖的守候。" + }, + { + "image_prompt": "Close-up of Xiao Lin smiling slightly, blurred city light bokeh background, beige coat collar visible, warm atmosphere, high quality portrait.", + "video_motion": "固定镜头微距拍摄，眼神自然眨动。", + "narration": "捧一杯热茶，原来幸福就藏在平凡夜晚里。" + } + ] +} \ No newline at end of file diff --git a/outputs/b77f2668-6451-47ff-81da-48b498ecb436/scenes.json b/outputs/b77f2668-6451-47ff-81da-48b498ecb436/scenes.json new file mode 100644 index 0000000..75b4740 --- /dev/null +++ b/outputs/b77f2668-6451-47ff-81da-48b498ecb436/scenes.json @@ -0,0 +1,19 @@ +{ + "scenes": [ + { + "image_prompt": "写一个温暖的城市夜景故事\n\n\n[Global Constraints]\n- Global Style: 电影感\n请严格遵守上述全局信息，并保持三分镜主角一致。，城市夜景，霓虹灯，电影感", + "video_motion": "缓慢推进镜头，轻微摇镜", + "narration": "夜色温柔落在街灯上" + }, + { + "image_prompt": "写一个温暖的城市夜景故事\n\n\n[Global Constraints]\n- Global Style: 电影感\n请严格遵守上述全局信息，并保持三分镜主角一致。，咖啡店窗边，暖光，细雨", + "video_motion": "侧向平移，人物轻轻抬头", + "narration": "雨声里藏着一段回忆" + }, + { + "image_prompt": "写一个温暖的城市夜景故事\n\n\n[Global Constraints]\n- Global Style: 电影感\n请严格遵守上述全局信息，并保持三分镜主角一致。，桥上远景，车流光轨，温暖", + "video_motion": "拉远全景，光轨流动", + "narration": "我们在光里学会告别" + } + ] +} \ No newline at end of file diff --git a/outputs/b77f2668-6451-47ff-81da-48b498ecb436/shot_4c28ead0e7d14b9bbd88f011ea70fa84.png b/outputs/b77f2668-6451-47ff-81da-48b498ecb436/shot_4c28ead0e7d14b9bbd88f011ea70fa84.png new file mode 100644 index 0000000..4c86bf9 Binary files /dev/null and b/outputs/b77f2668-6451-47ff-81da-48b498ecb436/shot_4c28ead0e7d14b9bbd88f011ea70fa84.png differ diff --git a/outputs/b77f2668-6451-47ff-81da-48b498ecb436/shot_7a2a3c7725b741bd8d1c967d1b9e3f53.png b/outputs/b77f2668-6451-47ff-81da-48b498ecb436/shot_7a2a3c7725b741bd8d1c967d1b9e3f53.png new file mode 100644 index 0000000..4c86bf9 Binary files /dev/null and b/outputs/b77f2668-6451-47ff-81da-48b498ecb436/shot_7a2a3c7725b741bd8d1c967d1b9e3f53.png differ diff --git a/outputs/b77f2668-6451-47ff-81da-48b498ecb436/shot_c03030151d7d48fbbdfe6a38f3ea290e.png b/outputs/b77f2668-6451-47ff-81da-48b498ecb436/shot_c03030151d7d48fbbdfe6a38f3ea290e.png new file mode 100644 index 0000000..4c86bf9 Binary files /dev/null and b/outputs/b77f2668-6451-47ff-81da-48b498ecb436/shot_c03030151d7d48fbbdfe6a38f3ea290e.png differ diff --git a/outputs/be02bd5e-7890-4d0e-8f53-60443d713db8/audio/shot_scene_01_01.mp3 b/outputs/be02bd5e-7890-4d0e-8f53-60443d713db8/audio/shot_scene_01_01.mp3 new file mode 100644 index 0000000..edb407f Binary files /dev/null and b/outputs/be02bd5e-7890-4d0e-8f53-60443d713db8/audio/shot_scene_01_01.mp3 differ diff --git a/outputs/be02bd5e-7890-4d0e-8f53-60443d713db8/audio/shot_scene_02_01.mp3 b/outputs/be02bd5e-7890-4d0e-8f53-60443d713db8/audio/shot_scene_02_01.mp3 new file mode 100644 index 0000000..811e6cd Binary files /dev/null and b/outputs/be02bd5e-7890-4d0e-8f53-60443d713db8/audio/shot_scene_02_01.mp3 differ diff --git a/outputs/be02bd5e-7890-4d0e-8f53-60443d713db8/audio/shot_scene_03_01.mp3 b/outputs/be02bd5e-7890-4d0e-8f53-60443d713db8/audio/shot_scene_03_01.mp3 new file mode 100644 index 0000000..9e95ca7 Binary files /dev/null and b/outputs/be02bd5e-7890-4d0e-8f53-60443d713db8/audio/shot_scene_03_01.mp3 differ diff --git a/outputs/be02bd5e-7890-4d0e-8f53-60443d713db8/clips/shot_scene_01_01.mp4 b/outputs/be02bd5e-7890-4d0e-8f53-60443d713db8/clips/shot_scene_01_01.mp4 new file mode 100644 index 0000000..4e84896 Binary files /dev/null and b/outputs/be02bd5e-7890-4d0e-8f53-60443d713db8/clips/shot_scene_01_01.mp4 differ diff --git a/outputs/be02bd5e-7890-4d0e-8f53-60443d713db8/clips/shot_scene_02_01.mp4 b/outputs/be02bd5e-7890-4d0e-8f53-60443d713db8/clips/shot_scene_02_01.mp4 new file mode 100644 index 0000000..f164bcb Binary files /dev/null and b/outputs/be02bd5e-7890-4d0e-8f53-60443d713db8/clips/shot_scene_02_01.mp4 differ diff --git a/outputs/be02bd5e-7890-4d0e-8f53-60443d713db8/clips/shot_scene_03_01.mp4 b/outputs/be02bd5e-7890-4d0e-8f53-60443d713db8/clips/shot_scene_03_01.mp4 new file mode 100644 index 0000000..9f5ab50 Binary files /dev/null and b/outputs/be02bd5e-7890-4d0e-8f53-60443d713db8/clips/shot_scene_03_01.mp4 differ diff --git a/outputs/be02bd5e-7890-4d0e-8f53-60443d713db8/final.mp4 b/outputs/be02bd5e-7890-4d0e-8f53-60443d713db8/final.mp4 new file mode 100644 index 0000000..bdc1cc2 Binary files /dev/null and b/outputs/be02bd5e-7890-4d0e-8f53-60443d713db8/final.mp4 differ diff --git a/outputs/be02bd5e-7890-4d0e-8f53-60443d713db8/scenes.json b/outputs/be02bd5e-7890-4d0e-8f53-60443d713db8/scenes.json new file mode 100644 index 0000000..e36e51c --- /dev/null +++ b/outputs/be02bd5e-7890-4d0e-8f53-60443d713db8/scenes.json @@ -0,0 +1,19 @@ +{ + "scenes": [ + { + "image_prompt": "电影感夜景，湿润街道反射霓虹，小林独自走在街头，米色羊毛大衣，灰色围巾，背景虚化光斑，35mm 镜头", + "video_motion": "低角度慢速跟拍，镜头随人物步伐轻微晃动，霓虹光影流动", + "narration": "城市的夜晚有些冷，霓虹灯在风中闪烁。" + }, + { + "image_prompt": "特写镜头，小林双手捧着热纸杯，蒸汽升腾，街边摊位暖光照明，米色大衣袖口可见，电影级布光", + "video_motion": "固定镜头，焦点从热气缓缓移至人物眼睛，蒸汽向上飘散", + "narration": "街角摊贩的热气，温暖了冻僵的双手。" + }, + { + "image_prompt": "中景背影，小林走向远处灯火，米色大衣，灰色围巾，城市背景光斑模糊，暖色调夜景，景深效果", + "video_motion": "镜头缓慢后拉，人物走入景深深处，背景光斑逐渐扩大", + "narration": "捧着这份温度，回家的路不再漫长。" + } + ] +} \ No newline at end of file diff --git a/outputs/be02bd5e-7890-4d0e-8f53-60443d713db8/task.json b/outputs/be02bd5e-7890-4d0e-8f53-60443d713db8/task.json new file mode 100644 index 0000000..d83dfef --- /dev/null +++ b/outputs/be02bd5e-7890-4d0e-8f53-60443d713db8/task.json @@ -0,0 +1,18 @@ +{ + "task_id": "be02bd5e-7890-4d0e-8f53-60443d713db8", + "status": "done", + "shots": [ + { + "shot_id": "scene_01_01", + "status": "done" + }, + { + "shot_id": "scene_02_01", + "status": "done" + }, + { + "shot_id": "scene_03_01", + "status": "done" + } + ] +} \ No newline at end of file diff --git a/outputs/c0a5eb46-51ce-4e55-b653-541103917c20/scenes.json b/outputs/c0a5eb46-51ce-4e55-b653-541103917c20/scenes.json new file mode 100644 index 0000000..75b4740 --- /dev/null +++ b/outputs/c0a5eb46-51ce-4e55-b653-541103917c20/scenes.json @@ -0,0 +1,19 @@ +{ + "scenes": [ + { + "image_prompt": "写一个温暖的城市夜景故事\n\n\n[Global Constraints]\n- Global Style: 电影感\n请严格遵守上述全局信息，并保持三分镜主角一致。，城市夜景，霓虹灯，电影感", + "video_motion": "缓慢推进镜头，轻微摇镜", + "narration": "夜色温柔落在街灯上" + }, + { + "image_prompt": "写一个温暖的城市夜景故事\n\n\n[Global Constraints]\n- Global Style: 电影感\n请严格遵守上述全局信息，并保持三分镜主角一致。，咖啡店窗边，暖光，细雨", + "video_motion": "侧向平移，人物轻轻抬头", + "narration": "雨声里藏着一段回忆" + }, + { + "image_prompt": "写一个温暖的城市夜景故事\n\n\n[Global Constraints]\n- Global Style: 电影感\n请严格遵守上述全局信息，并保持三分镜主角一致。，桥上远景，车流光轨，温暖", + "video_motion": "拉远全景，光轨流动", + "narration": "我们在光里学会告别" + } + ] +} \ No newline at end of file diff --git a/outputs/c0a5eb46-51ce-4e55-b653-541103917c20/shot_5b88a978a09a416e83380540a100f578.png b/outputs/c0a5eb46-51ce-4e55-b653-541103917c20/shot_5b88a978a09a416e83380540a100f578.png new file mode 100644 index 0000000..4c86bf9 Binary files /dev/null and b/outputs/c0a5eb46-51ce-4e55-b653-541103917c20/shot_5b88a978a09a416e83380540a100f578.png differ diff --git a/outputs/c0a5eb46-51ce-4e55-b653-541103917c20/shot_5bb49e0bc63649c386901afba4579880.png b/outputs/c0a5eb46-51ce-4e55-b653-541103917c20/shot_5bb49e0bc63649c386901afba4579880.png new file mode 100644 index 0000000..4c86bf9 Binary files /dev/null and b/outputs/c0a5eb46-51ce-4e55-b653-541103917c20/shot_5bb49e0bc63649c386901afba4579880.png differ diff --git a/outputs/c0a5eb46-51ce-4e55-b653-541103917c20/shot_b25526b451b646ab95eb45b23a52cb2f.png b/outputs/c0a5eb46-51ce-4e55-b653-541103917c20/shot_b25526b451b646ab95eb45b23a52cb2f.png new file mode 100644 index 0000000..4c86bf9 Binary files /dev/null and b/outputs/c0a5eb46-51ce-4e55-b653-541103917c20/shot_b25526b451b646ab95eb45b23a52cb2f.png differ diff --git a/outputs/d4de4ab3-29bb-41b1-bc97-daa3d131baaf/scenes.json b/outputs/d4de4ab3-29bb-41b1-bc97-daa3d131baaf/scenes.json new file mode 100644 index 0000000..c4da354 --- /dev/null +++ b/outputs/d4de4ab3-29bb-41b1-bc97-daa3d131baaf/scenes.json @@ -0,0 +1,19 @@ +{ + "scenes": [ + { + "image_prompt": "hello，城市夜景，霓虹灯，电影感", + "video_motion": "缓慢推进镜头，轻微摇镜", + "narration": "夜色温柔落在街灯上" + }, + { + "image_prompt": "hello，咖啡店窗边，暖光，细雨", + "video_motion": "侧向平移，人物轻轻抬头", + "narration": "雨声里藏着一段回忆" + }, + { + "image_prompt": "hello，桥上远景，车流光轨，温暖", + "video_motion": "拉远全景，光轨流动", + "narration": "我们在光里学会告别" + } + ] +} \ No newline at end of file diff --git a/outputs/f2810b54-03a6-45e9-9032-9e6b2e33c104/audio/shot_scene_01_01.mp3 b/outputs/f2810b54-03a6-45e9-9032-9e6b2e33c104/audio/shot_scene_01_01.mp3 new file mode 100644 index 0000000..36c5dc2 Binary files /dev/null and b/outputs/f2810b54-03a6-45e9-9032-9e6b2e33c104/audio/shot_scene_01_01.mp3 differ diff --git a/outputs/f2810b54-03a6-45e9-9032-9e6b2e33c104/audio/shot_scene_02_01.mp3 b/outputs/f2810b54-03a6-45e9-9032-9e6b2e33c104/audio/shot_scene_02_01.mp3 new file mode 100644 index 0000000..cee3639 Binary files /dev/null and b/outputs/f2810b54-03a6-45e9-9032-9e6b2e33c104/audio/shot_scene_02_01.mp3 differ diff --git a/outputs/f2810b54-03a6-45e9-9032-9e6b2e33c104/audio/shot_scene_03_01.mp3 b/outputs/f2810b54-03a6-45e9-9032-9e6b2e33c104/audio/shot_scene_03_01.mp3 new file mode 100644 index 0000000..e9e8d2a Binary files /dev/null and b/outputs/f2810b54-03a6-45e9-9032-9e6b2e33c104/audio/shot_scene_03_01.mp3 differ diff --git a/outputs/f2810b54-03a6-45e9-9032-9e6b2e33c104/clips/shot_scene_01_01.mp4 b/outputs/f2810b54-03a6-45e9-9032-9e6b2e33c104/clips/shot_scene_01_01.mp4 new file mode 100644 index 0000000..c9faece Binary files /dev/null and b/outputs/f2810b54-03a6-45e9-9032-9e6b2e33c104/clips/shot_scene_01_01.mp4 differ diff --git a/outputs/f2810b54-03a6-45e9-9032-9e6b2e33c104/clips/shot_scene_02_01.mp4 b/outputs/f2810b54-03a6-45e9-9032-9e6b2e33c104/clips/shot_scene_02_01.mp4 new file mode 100644 index 0000000..681f666 Binary files /dev/null and b/outputs/f2810b54-03a6-45e9-9032-9e6b2e33c104/clips/shot_scene_02_01.mp4 differ diff --git a/outputs/f2810b54-03a6-45e9-9032-9e6b2e33c104/clips/shot_scene_03_01.mp4 b/outputs/f2810b54-03a6-45e9-9032-9e6b2e33c104/clips/shot_scene_03_01.mp4 new file mode 100644 index 0000000..9c2119b Binary files /dev/null and b/outputs/f2810b54-03a6-45e9-9032-9e6b2e33c104/clips/shot_scene_03_01.mp4 differ diff --git a/outputs/f2810b54-03a6-45e9-9032-9e6b2e33c104/final.mp4 b/outputs/f2810b54-03a6-45e9-9032-9e6b2e33c104/final.mp4 new file mode 100644 index 0000000..c4f4c32 Binary files /dev/null and b/outputs/f2810b54-03a6-45e9-9032-9e6b2e33c104/final.mp4 differ diff --git a/outputs/f2810b54-03a6-45e9-9032-9e6b2e33c104/scenes.json b/outputs/f2810b54-03a6-45e9-9032-9e6b2e33c104/scenes.json new file mode 100644 index 0000000..5b08e9f --- /dev/null +++ b/outputs/f2810b54-03a6-45e9-9032-9e6b2e33c104/scenes.json @@ -0,0 +1,19 @@ +{ + "scenes": [ + { + "image_prompt": "写一个温暖的城市夜景故事，城市夜景，霓虹灯，电影感", + "video_motion": "缓慢推进镜头，轻微摇镜", + "narration": "夜色温柔落在街灯上" + }, + { + "image_prompt": "写一个温暖的城市夜景故事，咖啡店窗边，暖光，细雨", + "video_motion": "侧向平移，人物轻轻抬头", + "narration": "雨声里藏着一段回忆" + }, + { + "image_prompt": "写一个温暖的城市夜景故事，桥上远景，车流光轨，温暖", + "video_motion": "拉远全景，光轨流动", + "narration": "我们在光里学会告别" + } + ] +} \ No newline at end of file diff --git a/outputs/f2810b54-03a6-45e9-9032-9e6b2e33c104/task.json b/outputs/f2810b54-03a6-45e9-9032-9e6b2e33c104/task.json new file mode 100644 index 0000000..fcfde4d --- /dev/null +++ b/outputs/f2810b54-03a6-45e9-9032-9e6b2e33c104/task.json @@ -0,0 +1,18 @@ +{ + "task_id": "f2810b54-03a6-45e9-9032-9e6b2e33c104", + "status": "done", + "shots": [ + { + "shot_id": "scene_01_01", + "status": "done" + }, + { + "shot_id": "scene_02_01", + "status": "done" + }, + { + "shot_id": "scene_03_01", + "status": "done" + } + ] +} \ No newline at end of file diff --git a/outputs/ff47613e-6516-4e5d-a9e8-fd0c7ef1ba77/audio/shot_scene_01_01.mp3 b/outputs/ff47613e-6516-4e5d-a9e8-fd0c7ef1ba77/audio/shot_scene_01_01.mp3 new file mode 100644 index 0000000..36c5dc2 Binary files /dev/null and b/outputs/ff47613e-6516-4e5d-a9e8-fd0c7ef1ba77/audio/shot_scene_01_01.mp3 differ diff --git a/outputs/ff47613e-6516-4e5d-a9e8-fd0c7ef1ba77/audio/shot_scene_02_01.mp3 b/outputs/ff47613e-6516-4e5d-a9e8-fd0c7ef1ba77/audio/shot_scene_02_01.mp3 new file mode 100644 index 0000000..cee3639 Binary files /dev/null and b/outputs/ff47613e-6516-4e5d-a9e8-fd0c7ef1ba77/audio/shot_scene_02_01.mp3 differ diff --git a/outputs/ff47613e-6516-4e5d-a9e8-fd0c7ef1ba77/audio/shot_scene_03_01.mp3 b/outputs/ff47613e-6516-4e5d-a9e8-fd0c7ef1ba77/audio/shot_scene_03_01.mp3 new file mode 100644 index 0000000..e9e8d2a Binary files /dev/null and b/outputs/ff47613e-6516-4e5d-a9e8-fd0c7ef1ba77/audio/shot_scene_03_01.mp3 differ diff --git a/outputs/ff47613e-6516-4e5d-a9e8-fd0c7ef1ba77/clips/shot_scene_01_01.mp4 b/outputs/ff47613e-6516-4e5d-a9e8-fd0c7ef1ba77/clips/shot_scene_01_01.mp4 new file mode 100644 index 0000000..a78a1c3 Binary files /dev/null and b/outputs/ff47613e-6516-4e5d-a9e8-fd0c7ef1ba77/clips/shot_scene_01_01.mp4 differ diff --git a/outputs/ff47613e-6516-4e5d-a9e8-fd0c7ef1ba77/clips/shot_scene_02_01.mp4 b/outputs/ff47613e-6516-4e5d-a9e8-fd0c7ef1ba77/clips/shot_scene_02_01.mp4 new file mode 100644 index 0000000..5a3388e Binary files /dev/null and b/outputs/ff47613e-6516-4e5d-a9e8-fd0c7ef1ba77/clips/shot_scene_02_01.mp4 differ diff --git a/outputs/ff47613e-6516-4e5d-a9e8-fd0c7ef1ba77/clips/shot_scene_03_01.mp4 b/outputs/ff47613e-6516-4e5d-a9e8-fd0c7ef1ba77/clips/shot_scene_03_01.mp4 new file mode 100644 index 0000000..5390725 Binary files /dev/null and b/outputs/ff47613e-6516-4e5d-a9e8-fd0c7ef1ba77/clips/shot_scene_03_01.mp4 differ diff --git a/outputs/ff47613e-6516-4e5d-a9e8-fd0c7ef1ba77/final.mp4 b/outputs/ff47613e-6516-4e5d-a9e8-fd0c7ef1ba77/final.mp4 new file mode 100644 index 0000000..1af94fe Binary files /dev/null and b/outputs/ff47613e-6516-4e5d-a9e8-fd0c7ef1ba77/final.mp4 differ diff --git a/outputs/ff47613e-6516-4e5d-a9e8-fd0c7ef1ba77/scenes.json b/outputs/ff47613e-6516-4e5d-a9e8-fd0c7ef1ba77/scenes.json new file mode 100644 index 0000000..5b08e9f --- /dev/null +++ b/outputs/ff47613e-6516-4e5d-a9e8-fd0c7ef1ba77/scenes.json @@ -0,0 +1,19 @@ +{ + "scenes": [ + { + "image_prompt": "写一个温暖的城市夜景故事，城市夜景，霓虹灯，电影感", + "video_motion": "缓慢推进镜头，轻微摇镜", + "narration": "夜色温柔落在街灯上" + }, + { + "image_prompt": "写一个温暖的城市夜景故事，咖啡店窗边，暖光，细雨", + "video_motion": "侧向平移，人物轻轻抬头", + "narration": "雨声里藏着一段回忆" + }, + { + "image_prompt": "写一个温暖的城市夜景故事，桥上远景，车流光轨，温暖", + "video_motion": "拉远全景，光轨流动", + "narration": "我们在光里学会告别" + } + ] +} \ No newline at end of file diff --git a/outputs/ff47613e-6516-4e5d-a9e8-fd0c7ef1ba77/shot_raw_scene_01_01.mp4 b/outputs/ff47613e-6516-4e5d-a9e8-fd0c7ef1ba77/shot_raw_scene_01_01.mp4 new file mode 100644 index 0000000..fe4e343 Binary files /dev/null and b/outputs/ff47613e-6516-4e5d-a9e8-fd0c7ef1ba77/shot_raw_scene_01_01.mp4 differ diff --git a/outputs/ff47613e-6516-4e5d-a9e8-fd0c7ef1ba77/shot_raw_scene_02_01.mp4 b/outputs/ff47613e-6516-4e5d-a9e8-fd0c7ef1ba77/shot_raw_scene_02_01.mp4 new file mode 100644 index 0000000..fe4e343 Binary files /dev/null and b/outputs/ff47613e-6516-4e5d-a9e8-fd0c7ef1ba77/shot_raw_scene_02_01.mp4 differ diff --git a/outputs/ff47613e-6516-4e5d-a9e8-fd0c7ef1ba77/shot_raw_scene_03_01.mp4 b/outputs/ff47613e-6516-4e5d-a9e8-fd0c7ef1ba77/shot_raw_scene_03_01.mp4 new file mode 100644 index 0000000..fe4e343 Binary files /dev/null and b/outputs/ff47613e-6516-4e5d-a9e8-fd0c7ef1ba77/shot_raw_scene_03_01.mp4 differ diff --git a/outputs/ff47613e-6516-4e5d-a9e8-fd0c7ef1ba77/task.json b/outputs/ff47613e-6516-4e5d-a9e8-fd0c7ef1ba77/task.json new file mode 100644 index 0000000..f4619ed --- /dev/null +++ b/outputs/ff47613e-6516-4e5d-a9e8-fd0c7ef1ba77/task.json @@ -0,0 +1,18 @@ +{ + "task_id": "ff47613e-6516-4e5d-a9e8-fd0c7ef1ba77", + "status": "done", + "shots": [ + { + "shot_id": "scene_01_01", + "status": "done" + }, + { + "shot_id": "scene_02_01", + "status": "done" + }, + { + "shot_id": "scene_03_01", + "status": "done" + } + ] +} \ No newline at end of file diff --git a/outputs/smoke-fallback-hard/scenes.json b/outputs/smoke-fallback-hard/scenes.json new file mode 100644 index 0000000..41d4636 --- /dev/null +++ b/outputs/smoke-fallback-hard/scenes.json @@ -0,0 +1,19 @@ +{ + "scenes": [ + { + "image_prompt": "fallback-hard-test，城市夜景，霓虹灯，电影感", + "video_motion": "缓慢推进镜头，轻微摇镜", + "narration": "夜色温柔落在街灯上" + }, + { + "image_prompt": "fallback-hard-test，咖啡店窗边，暖光，细雨", + "video_motion": "侧向平移，人物轻轻抬头", + "narration": "雨声里藏着一段回忆" + }, + { + "image_prompt": "fallback-hard-test，桥上远景，车流光轨，温暖", + "video_motion": "拉远全景，光轨流动", + "narration": "我们在光里学会告别" + } + ] +} \ No newline at end of file diff --git a/outputs/smoke-fallback-hard/shot_4f0d67d8750444418ff2192abf0bcf7a.png b/outputs/smoke-fallback-hard/shot_4f0d67d8750444418ff2192abf0bcf7a.png new file mode 100644 index 0000000..4c86bf9 Binary files /dev/null and b/outputs/smoke-fallback-hard/shot_4f0d67d8750444418ff2192abf0bcf7a.png differ diff --git a/outputs/smoke-fallback-hard/shot_7c95c8ed44144ebca4d2a338339c1b62.png b/outputs/smoke-fallback-hard/shot_7c95c8ed44144ebca4d2a338339c1b62.png new file mode 100644 index 0000000..4c86bf9 Binary files /dev/null and b/outputs/smoke-fallback-hard/shot_7c95c8ed44144ebca4d2a338339c1b62.png differ diff --git a/outputs/smoke-fallback-hard/shot_f5aa777799864916915a176883fb5ab2.png b/outputs/smoke-fallback-hard/shot_f5aa777799864916915a176883fb5ab2.png new file mode 100644 index 0000000..4c86bf9 Binary files /dev/null and b/outputs/smoke-fallback-hard/shot_f5aa777799864916915a176883fb5ab2.png differ diff --git a/outputs/smoke-multi-1/audio/shot_scene_01_01.mp3 b/outputs/smoke-multi-1/audio/shot_scene_01_01.mp3 new file mode 100644 index 0000000..173c951 Binary files /dev/null and b/outputs/smoke-multi-1/audio/shot_scene_01_01.mp3 differ diff --git a/outputs/smoke-multi-1/clips/shot_scene_01_01.mp4 b/outputs/smoke-multi-1/clips/shot_scene_01_01.mp4 new file mode 100644 index 0000000..e12eccd Binary files /dev/null and b/outputs/smoke-multi-1/clips/shot_scene_01_01.mp4 differ diff --git a/outputs/smoke-multi-1/final.mp4 b/outputs/smoke-multi-1/final.mp4 new file mode 100644 index 0000000..5d3c539 Binary files /dev/null and b/outputs/smoke-multi-1/final.mp4 differ diff --git a/outputs/smoke-multi-1/shot_7db8c0342c0b406b84e2919fc9271ba3.png b/outputs/smoke-multi-1/shot_7db8c0342c0b406b84e2919fc9271ba3.png new file mode 100644 index 0000000..4c86bf9 Binary files /dev/null and b/outputs/smoke-multi-1/shot_7db8c0342c0b406b84e2919fc9271ba3.png differ diff --git a/outputs/smoke-multi-1/task.json b/outputs/smoke-multi-1/task.json new file mode 100644 index 0000000..5b40339 --- /dev/null +++ b/outputs/smoke-multi-1/task.json @@ -0,0 +1,10 @@ +{ + "task_id": "smoke-multi-1", + "status": "done", + "shots": [ + { + "shot_id": "scene_01_01", + "status": "done" + } + ] +} \ No newline at end of file diff --git a/outputs/smoke-refine-1/refine_scene_1.json b/outputs/smoke-refine-1/refine_scene_1.json new file mode 100644 index 0000000..4fcda21 --- /dev/null +++ b/outputs/smoke-refine-1/refine_scene_1.json @@ -0,0 +1,6 @@ +{ + "index": 1, + "image_prompt": "a warm city night", + "video_motion": "缓慢推进镜头", + "narration": "夜色温柔（更凝练）" +} \ No newline at end of file diff --git a/outputs/smoke-refine-sse/refine_scene_1.json b/outputs/smoke-refine-sse/refine_scene_1.json new file mode 100644 index 0000000..a93b224 --- /dev/null +++ b/outputs/smoke-refine-sse/refine_scene_1.json @@ -0,0 +1,6 @@ +{ + "index": 1, + "image_prompt": "test", + "video_motion": "move", + "narration": "旁白（更凝练）" +} \ No newline at end of file diff --git a/outputs/smoke-refine-sse/scenes.json b/outputs/smoke-refine-sse/scenes.json new file mode 100644 index 0000000..87bf95e --- /dev/null +++ b/outputs/smoke-refine-sse/scenes.json @@ -0,0 +1,19 @@ +{ + "scenes": [ + { + "image_prompt": "refine-sse-test，城市夜景，霓虹灯，电影感", + "video_motion": "缓慢推进镜头，轻微摇镜", + "narration": "夜色温柔落在街灯上" + }, + { + "image_prompt": "refine-sse-test，咖啡店窗边，暖光，细雨", + "video_motion": "侧向平移，人物轻轻抬头", + "narration": "雨声里藏着一段回忆" + }, + { + "image_prompt": "refine-sse-test，桥上远景，车流光轨，温暖", + "video_motion": "拉远全景，光轨流动", + "narration": "我们在光里学会告别" + } + ] +} \ No newline at end of file diff --git a/outputs/smoke-refine-sse/shot_2bc87cbac84f4848bce8bf7a4d28e0d1.png b/outputs/smoke-refine-sse/shot_2bc87cbac84f4848bce8bf7a4d28e0d1.png new file mode 100644 index 0000000..4c86bf9 Binary files /dev/null and b/outputs/smoke-refine-sse/shot_2bc87cbac84f4848bce8bf7a4d28e0d1.png differ diff --git a/outputs/smoke-refine-sse/shot_81e1f292749d4631ad90f399978d7031.png b/outputs/smoke-refine-sse/shot_81e1f292749d4631ad90f399978d7031.png new file mode 100644 index 0000000..4c86bf9 Binary files /dev/null and b/outputs/smoke-refine-sse/shot_81e1f292749d4631ad90f399978d7031.png differ diff --git a/outputs/smoke-refine-sse/shot_9ee28ddcf6c44fc59ad5d7a5b4d23b9e.png b/outputs/smoke-refine-sse/shot_9ee28ddcf6c44fc59ad5d7a5b4d23b9e.png new file mode 100644 index 0000000..4c86bf9 Binary files /dev/null and b/outputs/smoke-refine-sse/shot_9ee28ddcf6c44fc59ad5d7a5b4d23b9e.png differ diff --git a/outputs/smoke-refine-sse/shot_a85bc815e422450393e0815bb3dd7d41.png b/outputs/smoke-refine-sse/shot_a85bc815e422450393e0815bb3dd7d41.png new file mode 100644 index 0000000..4c86bf9 Binary files /dev/null and b/outputs/smoke-refine-sse/shot_a85bc815e422450393e0815bb3dd7d41.png differ diff --git a/outputs/smoke-script-1/scenes.json b/outputs/smoke-script-1/scenes.json new file mode 100644 index 0000000..c4da354 --- /dev/null +++ b/outputs/smoke-script-1/scenes.json @@ -0,0 +1,19 @@ +{ + "scenes": [ + { + "image_prompt": "hello，城市夜景，霓虹灯，电影感", + "video_motion": "缓慢推进镜头，轻微摇镜", + "narration": "夜色温柔落在街灯上" + }, + { + "image_prompt": "hello，咖啡店窗边，暖光，细雨", + "video_motion": "侧向平移，人物轻轻抬头", + "narration": "雨声里藏着一段回忆" + }, + { + "image_prompt": "hello，桥上远景，车流光轨，温暖", + "video_motion": "拉远全景，光轨流动", + "narration": "我们在光里学会告别" + } + ] +} \ No newline at end of file diff --git a/outputs/smoke-test-1/audio/shot_scene_01_01.mp3 b/outputs/smoke-test-1/audio/shot_scene_01_01.mp3 new file mode 100644 index 0000000..173c951 Binary files /dev/null and b/outputs/smoke-test-1/audio/shot_scene_01_01.mp3 differ diff --git a/outputs/smoke-test-1/audio/shot_scene_02_01.mp3 b/outputs/smoke-test-1/audio/shot_scene_02_01.mp3 new file mode 100644 index 0000000..5ffa77b Binary files /dev/null and b/outputs/smoke-test-1/audio/shot_scene_02_01.mp3 differ diff --git a/outputs/smoke-test-1/clips/shot_scene_01_01.mp4 b/outputs/smoke-test-1/clips/shot_scene_01_01.mp4 new file mode 100644 index 0000000..e12eccd Binary files /dev/null and b/outputs/smoke-test-1/clips/shot_scene_01_01.mp4 differ diff --git a/outputs/smoke-test-1/clips/shot_scene_02_01.mp4 b/outputs/smoke-test-1/clips/shot_scene_02_01.mp4 new file mode 100644 index 0000000..243a7b5 Binary files /dev/null and b/outputs/smoke-test-1/clips/shot_scene_02_01.mp4 differ diff --git a/outputs/smoke-test-1/final.mp4 b/outputs/smoke-test-1/final.mp4 new file mode 100644 index 0000000..1a88b15 Binary files /dev/null and b/outputs/smoke-test-1/final.mp4 differ diff --git a/outputs/smoke-test-1/task.json b/outputs/smoke-test-1/task.json new file mode 100644 index 0000000..3d6eac1 --- /dev/null +++ b/outputs/smoke-test-1/task.json @@ -0,0 +1,14 @@ +{ + "task_id": "smoke-test-1", + "status": "done", + "shots": [ + { + "shot_id": "scene_01_01", + "status": "done" + }, + { + "shot_id": "scene_02_01", + "status": "done" + } + ] +} \ No newline at end of file diff --git a/outputs/smoke-test-2/audio/shot_scene_01_01.mp3 b/outputs/smoke-test-2/audio/shot_scene_01_01.mp3 new file mode 100644 index 0000000..173c951 Binary files /dev/null and b/outputs/smoke-test-2/audio/shot_scene_01_01.mp3 differ diff --git a/outputs/smoke-test-2/audio/shot_scene_02_01.mp3 b/outputs/smoke-test-2/audio/shot_scene_02_01.mp3 new file mode 100644 index 0000000..5ffa77b Binary files /dev/null and b/outputs/smoke-test-2/audio/shot_scene_02_01.mp3 differ diff --git a/outputs/smoke-test-2/clips/shot_scene_01_01.mp4 b/outputs/smoke-test-2/clips/shot_scene_01_01.mp4 new file mode 100644 index 0000000..e12eccd Binary files /dev/null and b/outputs/smoke-test-2/clips/shot_scene_01_01.mp4 differ diff --git a/outputs/smoke-test-2/clips/shot_scene_02_01.mp4 b/outputs/smoke-test-2/clips/shot_scene_02_01.mp4 new file mode 100644 index 0000000..243a7b5 Binary files /dev/null and b/outputs/smoke-test-2/clips/shot_scene_02_01.mp4 differ diff --git a/outputs/smoke-test-2/final.mp4 b/outputs/smoke-test-2/final.mp4 new file mode 100644 index 0000000..1a88b15 Binary files /dev/null and b/outputs/smoke-test-2/final.mp4 differ diff --git a/outputs/smoke-test-2/task.json b/outputs/smoke-test-2/task.json new file mode 100644 index 0000000..462095e --- /dev/null +++ b/outputs/smoke-test-2/task.json @@ -0,0 +1,14 @@ +{ + "task_id": "smoke-test-2", + "status": "done", + "shots": [ + { + "shot_id": "scene_01_01", + "status": "done" + }, + { + "shot_id": "scene_02_01", + "status": "done" + } + ] +} \ No newline at end of file diff --git a/outputs/smoke-test-3/audio/shot_scene_01_01.mp3 b/outputs/smoke-test-3/audio/shot_scene_01_01.mp3 new file mode 100644 index 0000000..173c951 Binary files /dev/null and b/outputs/smoke-test-3/audio/shot_scene_01_01.mp3 differ diff --git a/outputs/smoke-test-3/clips/shot_scene_01_01.mp4 b/outputs/smoke-test-3/clips/shot_scene_01_01.mp4 new file mode 100644 index 0000000..e12eccd Binary files /dev/null and b/outputs/smoke-test-3/clips/shot_scene_01_01.mp4 differ diff --git a/outputs/smoke-test-3/final.mp4 b/outputs/smoke-test-3/final.mp4 new file mode 100644 index 0000000..5d3c539 Binary files /dev/null and b/outputs/smoke-test-3/final.mp4 differ diff --git a/outputs/smoke-test-3/shot_813d7ff6067346c2875b688874c3ed54.png b/outputs/smoke-test-3/shot_813d7ff6067346c2875b688874c3ed54.png new file mode 100644 index 0000000..4c86bf9 Binary files /dev/null and b/outputs/smoke-test-3/shot_813d7ff6067346c2875b688874c3ed54.png differ diff --git a/outputs/smoke-test-3/task.json b/outputs/smoke-test-3/task.json new file mode 100644 index 0000000..8a2977b --- /dev/null +++ b/outputs/smoke-test-3/task.json @@ -0,0 +1,10 @@ +{ + "task_id": "smoke-test-3", + "status": "done", + "shots": [ + { + "shot_id": "scene_01_01", + "status": "done" + } + ] +} \ No newline at end of file diff --git a/outputs/smoke-ui-1/scenes.json b/outputs/smoke-ui-1/scenes.json new file mode 100644 index 0000000..9a2bb52 --- /dev/null +++ b/outputs/smoke-ui-1/scenes.json @@ -0,0 +1,19 @@ +{ + "scenes": [ + { + "image_prompt": "自动分镜测试，城市夜景，霓虹灯，电影感", + "video_motion": "缓慢推进镜头，轻微摇镜", + "narration": "夜色温柔落在街灯上" + }, + { + "image_prompt": "自动分镜测试，咖啡店窗边，暖光，细雨", + "video_motion": "侧向平移，人物轻轻抬头", + "narration": "雨声里藏着一段回忆" + }, + { + "image_prompt": "自动分镜测试，桥上远景，车流光轨，温暖", + "video_motion": "拉远全景，光轨流动", + "narration": "我们在光里学会告别" + } + ] +} \ No newline at end of file diff --git a/outputs/smoke-ui-1/shot_19fd419d32b046b2ab9652d013616106.png b/outputs/smoke-ui-1/shot_19fd419d32b046b2ab9652d013616106.png new file mode 100644 index 0000000..4c86bf9 Binary files /dev/null and b/outputs/smoke-ui-1/shot_19fd419d32b046b2ab9652d013616106.png differ diff --git a/outputs/smoke-ui-1/shot_acb4acee6a51464b9af2f83226c21f0c.png b/outputs/smoke-ui-1/shot_acb4acee6a51464b9af2f83226c21f0c.png new file mode 100644 index 0000000..4c86bf9 Binary files /dev/null and b/outputs/smoke-ui-1/shot_acb4acee6a51464b9af2f83226c21f0c.png differ diff --git a/outputs/smoke-ui-1/shot_f0cdc48d52de427fb8665382c139d290.png b/outputs/smoke-ui-1/shot_f0cdc48d52de427fb8665382c139d290.png new file mode 100644 index 0000000..4c86bf9 Binary files /dev/null and b/outputs/smoke-ui-1/shot_f0cdc48d52de427fb8665382c139d290.png differ diff --git a/outputs/smoke-ui-2/scenes.json b/outputs/smoke-ui-2/scenes.json new file mode 100644 index 0000000..19c7638 --- /dev/null +++ b/outputs/smoke-ui-2/scenes.json @@ -0,0 +1,19 @@ +{ + "scenes": [ + { + "image_prompt": "状态机测试，城市夜景，霓虹灯，电影感", + "video_motion": "缓慢推进镜头，轻微摇镜", + "narration": "夜色温柔落在街灯上" + }, + { + "image_prompt": "状态机测试，咖啡店窗边，暖光，细雨", + "video_motion": "侧向平移，人物轻轻抬头", + "narration": "雨声里藏着一段回忆" + }, + { + "image_prompt": "状态机测试，桥上远景，车流光轨，温暖", + "video_motion": "拉远全景，光轨流动", + "narration": "我们在光里学会告别" + } + ] +} \ No newline at end of file diff --git a/outputs/smoke-ui-2/shot_665af13cab614672b5934a54bf2a5970.png b/outputs/smoke-ui-2/shot_665af13cab614672b5934a54bf2a5970.png new file mode 100644 index 0000000..4c86bf9 Binary files /dev/null and b/outputs/smoke-ui-2/shot_665af13cab614672b5934a54bf2a5970.png differ diff --git a/outputs/smoke-ui-2/shot_9e9813f55d824ae88f0eaa2e35be4a45.png b/outputs/smoke-ui-2/shot_9e9813f55d824ae88f0eaa2e35be4a45.png new file mode 100644 index 0000000..4c86bf9 Binary files /dev/null and b/outputs/smoke-ui-2/shot_9e9813f55d824ae88f0eaa2e35be4a45.png differ diff --git a/outputs/smoke-ui-2/shot_caadfc482f8c421982ea21853ceba306.png b/outputs/smoke-ui-2/shot_caadfc482f8c421982ea21853ceba306.png new file mode 100644 index 0000000..4c86bf9 Binary files /dev/null and b/outputs/smoke-ui-2/shot_caadfc482f8c421982ea21853ceba306.png differ diff --git a/outputs/smoke-ui-3/scenes.json b/outputs/smoke-ui-3/scenes.json new file mode 100644 index 0000000..725e6cb --- /dev/null +++ b/outputs/smoke-ui-3/scenes.json @@ -0,0 +1,19 @@ +{ + "scenes": [ + { + "image_prompt": "schema版本验证，城市夜景，霓虹灯，电影感", + "video_motion": "缓慢推进镜头，轻微摇镜", + "narration": "夜色温柔落在街灯上" + }, + { + "image_prompt": "schema版本验证，咖啡店窗边，暖光，细雨", + "video_motion": "侧向平移，人物轻轻抬头", + "narration": "雨声里藏着一段回忆" + }, + { + "image_prompt": "schema版本验证，桥上远景，车流光轨，温暖", + "video_motion": "拉远全景，光轨流动", + "narration": "我们在光里学会告别" + } + ] +} \ No newline at end of file diff --git a/outputs/smoke-ui-3/shot_354103ba62a64dff8177c05b4c247541.png b/outputs/smoke-ui-3/shot_354103ba62a64dff8177c05b4c247541.png new file mode 100644 index 0000000..4c86bf9 Binary files /dev/null and b/outputs/smoke-ui-3/shot_354103ba62a64dff8177c05b4c247541.png differ diff --git a/outputs/smoke-ui-3/shot_71005f19e64c4c81a3d7b7f35c1a4cb9.png b/outputs/smoke-ui-3/shot_71005f19e64c4c81a3d7b7f35c1a4cb9.png new file mode 100644 index 0000000..4c86bf9 Binary files /dev/null and b/outputs/smoke-ui-3/shot_71005f19e64c4c81a3d7b7f35c1a4cb9.png differ diff --git a/outputs/smoke-ui-3/shot_fa0a1de7cd9c4a0993ca681830d5a6e1.png b/outputs/smoke-ui-3/shot_fa0a1de7cd9c4a0993ca681830d5a6e1.png new file mode 100644 index 0000000..4c86bf9 Binary files /dev/null and b/outputs/smoke-ui-3/shot_fa0a1de7cd9c4a0993ca681830d5a6e1.png differ diff --git a/server/index.js b/server/index.js index 10a0480..4714761 100644 --- a/server/index.js +++ b/server/index.js @@ -59,6 +59,25 @@ function sseSend(res, event, data) { res.write("\n"); } +function sseStageUpdate(res, payload) { + // Unified schema for frontend stage rendering. + const safe = { + schema_version: 1, + stage: String(payload && payload.stage ? payload.stage : "Unknown"), + progress: + payload && typeof payload.progress === "number" && Number.isFinite(payload.progress) + ? Math.max(0, Math.min(1, payload.progress)) + : null, + scene_index: payload && Number.isFinite(payload.scene_index) ? Number(payload.scene_index) : null, + scene_json: payload && payload.scene_json && typeof payload.scene_json === "object" ? payload.scene_json : null, + shot_id: payload && payload.shot_id ? String(payload.shot_id) : null, + shot_status: payload && payload.shot_status ? String(payload.shot_status) : null, + message: payload && payload.message ? String(payload.message) : "", + timestamp: Date.now(), + }; + sseSend(res, "stage_update", JSON.stringify(safe)); +} + function newTaskId() { // `crypto.randomUUID()` exists on newer Node versions; fall back for older runtimes. if (crypto && typeof crypto.randomUUID === "function") return crypto.randomUUID(); @@ -81,7 +100,19 @@ function ensureTaskDir(taskId) { return dir; } -function spawnPythonStep({ step, prompt, configPath, mock, globalStyle, character, taskId, sceneIndex }) { +function spawnPythonStep({ + step, + prompt, + configPath, + mock, + globalStyle, + character, + taskId, + sceneIndex, + llmProvider, + imageProvider, + imageFallbackProvider, +}) { const py = process.env.PYTHON_BIN || "python3.10"; const args = [ "-m", @@ -99,7 +130,11 @@ function spawnPythonStep({ step, prompt, configPath, mock, globalStyle, characte if (globalStyle) args.push("--global-style", globalStyle); if (character) args.push("--character", character); if (mock) args.push("--mock"); - return spawn(py, args, { cwd: repoRoot, env: process.env, stdio: ["pipe", "pipe", "pipe"] }); + const childEnv = { ...process.env }; + if (llmProvider) childEnv.ENGINE_LLM_PROVIDER = String(llmProvider).trim(); + if (imageProvider) childEnv.ENGINE_IMAGE_PROVIDER = String(imageProvider).trim(); + if (imageFallbackProvider) childEnv.ENGINE_IMAGE_FALLBACK_PROVIDER = String(imageFallbackProvider).trim(); + return spawn(py, args, { cwd: repoRoot, env: childEnv, stdio: ["pipe", "pipe", "pipe"] }); } app.get("/api/script", (req, res) => { @@ -108,6 +143,9 @@ app.get("/api/script", (req, res) => { const globalStyle = String(req.query.global_style || "").trim(); const character = String(req.query.character || "").trim(); const configPath = String(req.query.config || "./configs/config.yaml"); + const llmProvider = String(req.query.llm_provider || "").trim(); + const imageProvider = String(req.query.image_provider || "").trim(); + const imageFallbackProvider = String(req.query.image_fallback_provider || "").trim(); if (!prompt) { res.status(400).json({ error: "missing prompt" }); @@ -129,9 +167,13 @@ app.get("/api/script", (req, res) => { globalStyle, character, taskId, + llmProvider, + imageProvider, + imageFallbackProvider, }); let buf = ""; + let sceneCount = 0; child.stdout.setEncoding("utf8"); child.stdout.on("data", (chunk) => { buf += chunk; @@ -139,15 +181,37 @@ app.get("/api/script", (req, res) => { buf = parts.pop() || ""; for (const line of parts) { if (!line) continue; - if (line.startsWith("SCENE_JSON ")) sseSend(res, "scene", line.slice("SCENE_JSON ".length)); - else if (line.startsWith("PROG ")) sseSend(res, "prog", line.slice("PROG ".length)); - else sseSend(res, "line", line); + if (line.startsWith("SCENE_JSON ")) { + try { + const scene = JSON.parse(line.slice("SCENE_JSON ".length)); + sceneCount += 1; + sseStageUpdate(res, { + stage: "Script", + scene_index: Number(scene.index || sceneCount) - 1, + scene_json: scene, + progress: Math.min(0.9, sceneCount / 3), + message: "scene_generated", + }); + } catch { + sseSend(res, "line", line); + } + } else if (line.startsWith("PROG ")) { + try { + const p = JSON.parse(line.slice("PROG ".length)); + sseStageUpdate(res, { stage: "Script", progress: Number(p.p || 0), message: p.msg || "" }); + } catch { + sseSend(res, "line", line); + } + } else { + sseSend(res, "line", line); + } } }); child.stderr.setEncoding("utf8"); child.stderr.on("data", (chunk) => { - sseSend(res, "error", chunk); + // stderr can contain non-fatal logs/warnings; keep as a normal line event. + sseSend(res, "line", "[stderr] " + chunk); }); req.on("close", () => { @@ -156,6 +220,7 @@ app.get("/api/script", (req, res) => { child.on("exit", (code) => { if (buf.trim()) sseSend(res, "line", buf.trim()); + if (code !== 0) sseSend(res, "error", `[ERROR] python exit_code=${code}`); sseSend(res, "done", String(code != null ? code : 0)); res.end(); }); @@ -170,6 +235,9 @@ app.post("/api/refine", (req, res) => { const globalStyle = String((req.body && req.body.global_style) || "").trim(); const character = String((req.body && req.body.character) || "").trim(); const configPath = String((req.body && req.body.config) || "./configs/config.yaml"); + const llmProvider = String((req.body && req.body.llm_provider) || "").trim(); + const imageProvider = String((req.body && req.body.image_provider) || "").trim(); + const imageFallbackProvider = String((req.body && req.body.image_fallback_provider) || "").trim(); const taskId = String((req.body && req.body.task_id) || "").trim() || newTaskId(); if (!prompt) return res.status(400).json({ error: "missing prompt" }); @@ -178,6 +246,9 @@ app.post("/api/refine", (req, res) => { return res.status(400).json({ error: "missing scene or scenes[]" }); } ensureTaskDir(taskId); + sseHeaders(res); + sseSend(res, "task", JSON.stringify({ task_id: taskId })); + sseStageUpdate(res, { stage: "Refine", progress: 0.05, message: "refine_start" }); const child = spawnPythonStep({ step: "refine", @@ -188,6 +259,9 @@ app.post("/api/refine", (req, res) => { character, taskId, sceneIndex, + llmProvider, + imageProvider, + imageFallbackProvider, }); if (Array.isArray(scenes)) { child.stdin.end(JSON.stringify({ scenes })); @@ -197,19 +271,53 @@ app.post("/api/refine", (req, res) => { let out = ""; let err = ""; + let buf = ""; child.stdout.setEncoding("utf8"); child.stderr.setEncoding("utf8"); - child.stdout.on("data", (c) => (out += c)); + child.stdout.on("data", (chunk) => { + out += chunk; + buf += chunk; + const parts = buf.split(/\r?\n/); + buf = parts.pop() || ""; + for (const line of parts) { + if (!line) continue; + if (line.startsWith("SCENE_JSON ")) { + try { + const scenePayload = JSON.parse(line.slice("SCENE_JSON ".length)); + sseStageUpdate(res, { + stage: "Refine", + progress: 1, + scene_index: Number(scenePayload.index || sceneIndex) - 1, + scene_json: scenePayload, + message: "scene_refined", + }); + } catch { + sseSend(res, "line", line); + } + } else if (line.startsWith("PROG ")) { + try { + const p = JSON.parse(line.slice("PROG ".length)); + sseStageUpdate(res, { stage: "Refine", progress: Number(p.p || 0), message: p.msg || "" }); + } catch { + sseSend(res, "line", line); + } + } else { + sseSend(res, "line", line); + } + } + }); child.stderr.on("data", (c) => (err += c)); child.on("exit", (code) => { - if (code !== 0) return res.status(500).json({ error: "python failed", stderr: err, stdout: out }); - const line = out - .split(/\r?\n/) - .map((s) => s.trim()) - .find((s) => s.startsWith("SCENE_JSON ")); - if (!line) return res.status(500).json({ error: "no SCENE_JSON", stderr: err, stdout: out }); - const payload = JSON.parse(line.slice("SCENE_JSON ".length)); - return res.json({ task_id: taskId, scene: payload, stderr: err }); + if (buf.trim()) sseSend(res, "line", buf.trim()); + if (err.trim()) sseSend(res, "line", "[stderr] " + err.trim()); + if (code !== 0) { + sseStageUpdate(res, { stage: "Refine", progress: null, message: `refine_failed(exit=${code})` }); + sseSend(res, "error", `[ERROR] python exit_code=${code}`); + return res.end(); + } + sseStageUpdate(res, { stage: "Refine", progress: 1, message: "refine_done" }); + sseSend(res, "done", JSON.stringify({ exit_code: code != null ? code : 0, task_id: taskId })); + return res.end(); }); }); @@ -222,6 +330,9 @@ app.post("/api/render", (req, res) => { const globalStyle = String((req.body && req.body.global_style) || "").trim(); const character = String((req.body && req.body.character) || "").trim(); const configPath = String((req.body && req.body.config) || "./configs/config.yaml"); + const llmProvider = String((req.body && req.body.llm_provider) || "").trim(); + const imageProvider = String((req.body && req.body.image_provider) || "").trim(); + const imageFallbackProvider = String((req.body && req.body.image_fallback_provider) || "").trim(); const taskId = String((req.body && req.body.task_id) || "").trim() || newTaskId(); if (!prompt) return res.status(400).json({ error: "missing prompt" }); @@ -245,6 +356,9 @@ app.post("/api/render", (req, res) => { globalStyle, character, taskId, + llmProvider, + imageProvider, + imageFallbackProvider, }); child.stdin.end(JSON.stringify({ scenes })); @@ -258,14 +372,26 @@ app.post("/api/render", (req, res) => { buf = parts.pop() || ""; for (const line of parts) { if (!line) continue; - if (line.startsWith("PROG ")) sseSend(res, "prog", line.slice("PROG ".length)); - else if (line.startsWith("PROG_SHOT ")) { + if (line.startsWith("PROG ")) { + try { + const p = JSON.parse(line.slice("PROG ".length)); + sseStageUpdate(res, { stage: "Render", progress: Number(p.p || 0), message: p.msg || "" }); + } catch { + sseSend(res, "line", line); + } + } else if (line.startsWith("PROG_SHOT ")) { const rest = line.slice("PROG_SHOT ".length).trim(); const firstSpace = rest.indexOf(" "); if (firstSpace > 0) { const shotId = rest.slice(0, firstSpace).trim(); const status = rest.slice(firstSpace + 1).trim(); - sseSend(res, "shot_progress", JSON.stringify({ shot_id: shotId, status })); + sseStageUpdate(res, { + stage: "Render", + progress: null, + shot_id: shotId, + shot_status: status, + message: "shot_progress", + }); } else { sseSend(res, "line", line); } @@ -276,7 +402,8 @@ app.post("/api/render", (req, res) => { }); child.stderr.on("data", (chunk) => { - sseSend(res, "error", chunk); + // stderr can contain non-fatal logs/warnings; keep as a normal line event. + sseSend(res, "line", "[stderr] " + chunk); }); req.on("close", () => { diff --git a/server/public/index.html b/server/public/index.html index 66df199..c4bb3e4 100644 --- a/server/public/index.html +++ b/server/public/index.html @@ -63,78 +63,82 @@ const [globalStyle, setGlobalStyle] = useState("电影感"); const [characterPreset, setCharacterPreset] = useState(""); const [mock, setMock] = useState(true); - const [logs, setLogs] = useState(""); - const [scenes, setScenes] = useState([null, null, null]); + const [llmProvider, setLlmProvider] = useState("mock"); + const [imageProvider, setImageProvider] = useState("mock"); + const [imageFallbackProvider, setImageFallbackProvider] = useState("mock"); + const [scenes, setScenes] = useState([]); + const [stageLogs, setStageLogs] = useState({ Script: [], Refine: [], Render: [] }); + const [stageState, setStageState] = useState({ + Script: "pending", + Refine: "pending", + Render: "pending", + }); + const [activeLogStage, setActiveLogStage] = useState("Script"); + const [renderProgress, setRenderProgress] = useState(0); const [canRender, setCanRender] = useState(false); const [finalVideoUrl, setFinalVideoUrl] = useState(""); const [taskId, setTaskId] = useState(""); const [toast, setToast] = useState(""); + const [scriptRunning, setScriptRunning] = useState(false); const esRef = useRef(null); const logRef = useRef(null); + const autoTimerRef = useRef(null); - const appendLog = (line) => { - setLogs((prev) => prev + line + "\n"); + const appendStageLog = (stage, line) => { + setStageLogs((prev) => ({ ...prev, [stage]: [...(prev[stage] || []), String(line)] })); }; const showToast = (msg) => { setToast(String(msg || "发生错误")); - // auto hide setTimeout(() => setToast(""), 6000); }; useEffect(() => { if (!logRef.current) return; logRef.current.scrollTop = logRef.current.scrollHeight; - }, [logs]); + }, [stageLogs]); - const startScript = () => { - stopScript(); - setLogs(""); - setScenes([null, null, null]); - setCanRender(false); - setFinalVideoUrl(""); - setTaskId(""); + const normalizeScene = (raw, idx) => { + const s = raw || {}; + return { + index: Number(s.index || idx + 1), + image_prompt: String(s.image_prompt || ""), + video_motion: String(s.video_motion || ""), + narration: String(s.narration || ""), + preview_url: String(s.preview_url || ""), + motion_camera: String(s.motion_camera || ""), + motion_direction: String(s.motion_direction || ""), + motion_speed: String(s.motion_speed || "normal"), + }; + }; - const url = `/api/script?prompt=${encodeURIComponent(prompt.trim())}&mock=${mock ? "1" : "0"}&global_style=${encodeURIComponent(globalStyle)}&character=${encodeURIComponent(characterPreset)}`; - const es = new EventSource(url); - esRef.current = es; + const parseStageUpdate = (raw) => { + if (!raw || typeof raw !== "object") return null; + if (Number(raw.schema_version) !== 1) return null; + const stage = String(raw.stage || ""); + if (!["Script", "Refine", "Render"].includes(stage)) return null; + return raw; + }; - es.addEventListener("status", (e) => appendLog("[status] " + e.data)); - es.addEventListener("error", (e) => { - const m = (e && e.data) ? e.data : "连接或后端错误"; - appendLog("[ERROR] " + m); - showToast(m); - }); - es.addEventListener("task", (e) => { - try { setTaskId(JSON.parse(e.data).task_id || ""); } catch (err) { } - }); - es.addEventListener("done", (e) => { - appendLog("[done] exit_code=" + e.data); - stopScript(); - }); - es.addEventListener("scene", (e) => { - try { - const obj = JSON.parse(e.data); - setScenes((prev) => { - const next = [...prev]; - next[obj.index - 1] = { - index: obj.index, - image_prompt: obj.image_prompt || "", - video_motion: obj.video_motion || "", - narration: obj.narration || "", - }; - return next; - }); - } catch (err) { - appendLog("[parse_error] " + err); - } - }); - es.addEventListener("line", (e) => { - appendLog(e.data); - if (e.data === "SCRIPT_END") setCanRender(true); - }); - es.onerror = () => appendLog("[error] connection error"); + const applyStageUpdate = (u) => { + const stage = String(u.stage || "Script"); + setStageState((prev) => ({ ...prev, [stage]: prev[stage] === "done" ? "done" : "running" })); + if (typeof u.progress === "number" && stage === "Render") { + setRenderProgress(Math.max(0, Math.min(1, u.progress))); + } + if (u.scene_json) { + const scene = u.scene_json; + const idx = Math.max(0, Number(scene.index || 1) - 1); + setScenes((prev) => { + const next = [...prev]; + next[idx] = normalizeScene(scene, idx); + return next; + }); + setCanRender(true); + } + if (u.message) appendStageLog(stage, u.message); + if (u.shot_id && u.shot_status) appendStageLog(stage, `${u.shot_id} -> ${u.shot_status}`); }; const stopScript = () => { @@ -142,68 +146,116 @@ esRef.current.close(); esRef.current = null; } + setScriptRunning(false); + }; + + const startScript = () => { + const p = prompt.trim(); + if (!p) return; + stopScript(); + setScriptRunning(true); + setCanRender(false); + setFinalVideoUrl(""); + setRenderProgress(0); + setScenes([]); + setStageLogs({ Script: [], Refine: [], Render: [] }); + setStageState({ Script: "running", Refine: "pending", Render: "pending" }); + setActiveLogStage("Script"); + + const q = new URLSearchParams({ + prompt: p, + mock: mock ? "1" : "0", + global_style: globalStyle, + character: characterPreset, + llm_provider: llmProvider, + image_provider: imageProvider, + image_fallback_provider: imageFallbackProvider, + }); + const es = new EventSource(`/api/script?${q.toString()}`); + esRef.current = es; + + es.addEventListener("task", (e) => { + try { setTaskId(JSON.parse(e.data).task_id || ""); } catch (err) { } + }); + es.addEventListener("stage_update", (e) => { + try { + const parsed = parseStageUpdate(JSON.parse(e.data)); + if (!parsed) return appendStageLog("Script", "[schema_error] invalid stage_update payload"); + applyStageUpdate(parsed); + } catch (err) { appendStageLog("Script", "[parse_error] " + err); } + }); + es.addEventListener("line", (e) => appendStageLog("Script", e.data)); + es.addEventListener("error", (e) => { + setStageState((prev) => ({ ...prev, Script: "failed" })); + showToast((e && e.data) ? e.data : "script连接错误"); + }); + es.addEventListener("done", () => { + stopScript(); + setCanRender(true); + setStageState((prev) => ({ ...prev, Script: "done" })); + }); + }; + + useEffect(() => { + if (autoTimerRef.current) clearTimeout(autoTimerRef.current); + autoTimerRef.current = setTimeout(() => { + if (!prompt.trim()) return; + startScript(); + }, 700); + return () => { + if (autoTimerRef.current) clearTimeout(autoTimerRef.current); + }; + }, [prompt, globalStyle, characterPreset, llmProvider, imageProvider, imageFallbackProvider, mock]); + + const updateMotionField = (idx, field, value) => { + setScenes((prev) => { + const next = [...prev]; + const cur = normalizeScene(next[idx], idx); + const merged = { ...cur, [field]: value }; + merged.video_motion = [merged.motion_camera, merged.motion_direction, merged.motion_speed].filter(Boolean).join(" | "); + next[idx] = merged; + return next; + }); }; const onEdit = (idx, field, value) => { setScenes((prev) => { const next = [...prev]; - const cur = next[idx] || { index: idx + 1, image_prompt: "", video_motion: "", narration: "" }; + const cur = normalizeScene(next[idx], idx); next[idx] = { ...cur, [field]: value }; return next; }); }; const refineOne = async (sceneIndex) => { - appendLog(`[refine] scene ${sceneIndex}...`); + setStageState((prev) => ({ ...prev, Refine: "running" })); + appendStageLog("Refine", `scene ${sceneIndex} refining...`); const s0 = scenes[sceneIndex - 1] || {}; - const payloadScene = { - image_prompt: s0.image_prompt || "", - video_motion: s0.video_motion || "", - narration: s0.narration || "", - }; const resp = await fetch("/api/refine", { method: "POST", headers: { "Content-Type": "application/json" }, - body: JSON.stringify({ prompt, scene: payloadScene, scene_index: sceneIndex, mock, global_style: globalStyle, character: characterPreset, task_id: taskId }), - }); - const data = await resp.json(); - if (!resp.ok) { - appendLog("[refine_error] " + JSON.stringify(data)); - showToast((data && (data.error || data.msg)) || "润色失败"); - return; - } - const s = data.scene; - setScenes((prev) => { - const next = [...prev]; - next[s.index - 1] = { - index: s.index, - image_prompt: s.image_prompt || "", - video_motion: s.video_motion || "", - narration: s.narration || "", - }; - return next; - }); - appendLog(`[refine] scene ${sceneIndex} done`); - }; - - const renderVideo = async () => { - appendLog("[render] start..."); - const payloadScenes = scenes.map((s, i) => ({ - image_prompt: (s && s.image_prompt) || "", - video_motion: (s && s.video_motion) || "", - narration: (s && s.narration) || "", - })); - const resp = await fetch("/api/render", { - method: "POST", - headers: { "Content-Type": "application/json" }, - body: JSON.stringify({ prompt, scenes: payloadScenes, mock, global_style: globalStyle, character: characterPreset, task_id: taskId }), + body: JSON.stringify({ + prompt, + scene: { + image_prompt: s0.image_prompt || "", + video_motion: s0.video_motion || "", + narration: s0.narration || "", + }, + scene_index: sceneIndex, + mock, + global_style: globalStyle, + character: characterPreset, + task_id: taskId, + llm_provider: llmProvider, + image_provider: imageProvider, + image_fallback_provider: imageFallbackProvider, + }), }); if (!resp.ok) { - appendLog("[render_error] http " + resp.status); - showToast("渲染请求失败（HTTP " + resp.status + "）"); + setStageState((prev) => ({ ...prev, Refine: "failed" })); + showToast("润色请求失败（HTTP " + resp.status + "）"); return; } - // Parse SSE from fetch (POST) const reader = resp.body.getReader(); const decoder = new TextDecoder("utf-8"); let buf = ""; @@ -224,40 +276,146 @@ const data = dataLines.join("\n"); if (event === "task") { try { setTaskId(JSON.parse(data).task_id || ""); } catch (err) { } - } else if (event === "prog") { - appendLog("[prog] " + data); + } else if (event === "stage_update") { + try { + const parsed = parseStageUpdate(JSON.parse(data)); + if (!parsed) return appendStageLog("Refine", "[schema_error] invalid stage_update payload"); + applyStageUpdate(parsed); + } catch (err) { appendStageLog("Refine", "[parse_error] " + err); } } else if (event === "error") { - appendLog("[ERROR] " + data); + appendStageLog("Refine", "[ERROR] " + data); + setStageState((prev) => ({ ...prev, Refine: "failed" })); + showToast(data); + } else if (event === "done") { + appendStageLog("Refine", `scene ${sceneIndex} done`); + setStageState((prev) => ({ ...prev, Refine: "done" })); + } else if (event === "line") { + appendStageLog("Refine", data); + } + } + } + }; + + const renderVideo = async () => { + setStageState((prev) => ({ ...prev, Render: "running" })); + setRenderProgress(0); + appendStageLog("Render", "render start..."); + const payloadScenes = scenes.map((s, i) => ({ + image_prompt: (s && s.image_prompt) || "", + video_motion: (s && s.video_motion) || "", + narration: (s && s.narration) || "", + })); + const resp = await fetch("/api/render", { + method: "POST", + headers: { "Content-Type": "application/json" }, + body: JSON.stringify({ + prompt, + scenes: payloadScenes, + mock, + global_style: globalStyle, + character: characterPreset, + task_id: taskId, + llm_provider: llmProvider, + image_provider: imageProvider, + image_fallback_provider: imageFallbackProvider, + }), + }); + if (!resp.ok) { + setStageState((prev) => ({ ...prev, Render: "failed" })); + showToast("渲染请求失败（HTTP " + resp.status + "）"); + return; + } + const reader = resp.body.getReader(); + const decoder = new TextDecoder("utf-8"); + let buf = ""; + while (true) { + const { value, done } = await reader.read(); + if (done) break; + buf += decoder.decode(value, { stream: true }); + const chunks = buf.split("\n\n"); + buf = chunks.pop() || ""; + for (const c of chunks) { + const lines = c.split("\n").filter(Boolean); + let event = "message"; + const dataLines = []; + for (const line of lines) { + if (line.startsWith("event:")) event = line.slice(6).trim(); + else if (line.startsWith("data:")) dataLines.push(line.slice(5).trim()); + } + const data = dataLines.join("\n"); + if (event === "task") { + try { setTaskId(JSON.parse(data).task_id || ""); } catch (err) { } + } else if (event === "stage_update") { + try { + const parsed = parseStageUpdate(JSON.parse(data)); + if (!parsed) return appendStageLog("Render", "[schema_error] invalid stage_update payload"); + applyStageUpdate(parsed); + } catch (err) { appendStageLog("Render", "[parse_error] " + err); } + } else if (event === "error") { + appendStageLog("Render", "[ERROR] " + data); + setStageState((prev) => ({ ...prev, Render: "failed" })); showToast(data); } else if (event === "done") { try { const obj = JSON.parse(data); const file = String(obj.output || "").split("/").pop() || "final.mp4"; const tid = taskId || (obj.task_id || ""); - appendLog("[render] done: " + file); if (tid) setFinalVideoUrl(`/api/static/${encodeURIComponent(tid)}/${encodeURIComponent(file)}?t=${Date.now()}`); + setRenderProgress(1); + setStageState((prev) => ({ ...prev, Render: "done" })); } catch (e) { - appendLog("[render_done_parse_error] " + e); + setStageState((prev) => ({ ...prev, Render: "failed" })); showToast("渲染完成消息解析失败"); } - } else { - appendLog(data); + } else if (event === "line") { + appendStageLog("Render", data); } } } }; + const activeLogs = useMemo(() => (stageLogs[activeLogStage] || []).join("\n"), [stageLogs, activeLogStage]); + const stageColor = (s) => { + if (s === "done") return "#16a34a"; + if (s === "running") return "#2563eb"; + if (s === "failed") return "#dc2626"; + return "#9ca3af"; + }; + return (
-

AiVideo POC：双向交互手搓平台

-

分镜可编辑、可单条润色，渲染完成后可直接预览与下载。

+

AiVideo POC：多后端模型平台

+

自动拆分 Prompt，SSE 实时反馈，Render 全链路可视化。

- setPrompt(e.target.value)} - /> + setPrompt(e.target.value)} /> +
+ +
+ + + +
@@ -271,26 +429,42 @@ - - + - {canRender ? ( - - ) : null} + {canRender ? : null} {taskId ? task_id: {taskId} : null}
+
+ {["Script", "Refine", "Render"].map((s) => ( + + ))} +
+ +
+
Render Progress: {(renderProgress * 100).toFixed(0)}%
+
+
+
+
+
{scenes.map((s, idx) => (
@@ -298,21 +472,27 @@ Scene {idx + 1}
+ {s && s.preview_url ? ( +
+ +
+ ) : null}
image_prompt
-