Skip to content

利用 AI 大模型和自动化工作流,根据主题或关键词一键生成高清短视频。Generate HD short videos from a topic or keyword with an automated AI workflow.

feat(video): add image motion options

harry070390f77011d7e3
diff --git a/app/models/schema.py b/app/models/schema.py
index e116ab42..d22355ab 100644
--- a/app/models/schema.py
+++ b/app/models/schema.py
@@ -53,6 +53,11 @@ class VideoFitMode(str, Enum):
contain = "contain"
+class ImageMotion(str, Enum):
+ static = "static"
+ zoom = "zoom"
+
+
SubtitleDisplayMode = Literal["sentence", "word_by_word"]
SubtitleAnimation = Literal["none", "pop_spring"]
_SUBTITLE_DISPLAY_MODES = ("sentence", "word_by_word")
@@ -108,6 +113,7 @@ class VideoParams(BaseModel):
video_terms: Optional[str | List[str]] = None # Keywords used to generate the video
video_aspect: Optional[VideoAspect] = VideoAspect.portrait.value
video_fit_mode: VideoFitMode = VideoFitMode.cover
+ image_motion: ImageMotion = ImageMotion.zoom
video_concat_mode: Optional[VideoConcatMode] = VideoConcatMode.random.value
video_transition_mode: Optional[VideoTransitionMode] = None
video_clip_duration: int = Field(default=5, ge=1)
diff --git a/app/services/material.py b/app/services/material.py
index 86955b83..8b045731 100644
--- a/app/services/material.py
+++ b/app/services/material.py
@@ -21,7 +21,7 @@ from moviepy.video.io.VideoFileClip import VideoFileClip
from PIL import Image, UnidentifiedImageError
from app.config import config
-from app.models.schema import MaterialInfo, VideoAspect, VideoConcatMode
+from app.models.schema import ImageMotion, MaterialInfo, VideoAspect, VideoConcatMode
from app.services import (
material_cache,
material_upload,
@@ -1737,14 +1737,18 @@ def generate_images_openai(
return [item]
-def _render_openai_image_video(image_path: str, clip_duration: int) -> str:
+def _render_openai_image_video(
+ image_path: str, clip_duration: int, image_motion: ImageMotion = ImageMotion.zoom
+) -> str:
"""
把生成的图片渲染成 mp4 片段,复用 local 素材的"图片 → 动态片段"管线。
渲染失败返回空字符串,由调用方停止后续付费请求。
"""
try:
- return video.render_image_zoom_video(image_path, clip_duration)
+ return video.render_image_zoom_video(
+ image_path, clip_duration, image_motion=image_motion
+ )
except Exception as e:
logger.error(
"failed to render generated image as a video clip: "
@@ -1761,6 +1765,7 @@ def _download_videos_openai_image_on_demand(
audio_duration: float,
max_clip_duration: int,
material_directory: str,
+ image_motion: ImageMotion = ImageMotion.zoom,
) -> List[str]:
"""
按脚本片段顺序逐张生成 OpenAI 兼容文生图素材,凑够所需总时长立即停止。
@@ -1771,6 +1776,8 @@ def _download_videos_openai_image_on_demand(
不再发起新的付费请求。明确拒绝的单张图片可跳过;请求结果不明、付费
结果下载失败或本地渲染失败时必须终止任务,避免后续关键词再次计费。
"""
+ # 直接调用素材服务也要在付费请求前校验,不能生成后才发现效果参数无效。
+ image_motion = ImageMotion(image_motion)
if not material_directory:
# 生成图片按任务计费且不可复用,默认落在任务目录便于追溯。
material_directory = utils.task_dir(task_id)
@@ -1805,7 +1812,9 @@ def _download_videos_openai_image_on_demand(
_persist_material_sources(task_id, material_sources)
raise
for item in items:
- video_file = _render_openai_image_video(item.url, max_clip_duration)
+ video_file = _render_openai_image_video(
+ item.url, max_clip_duration, image_motion=image_motion
+ )
if not video_file:
_persist_material_sources(task_id, material_sources)
raise OpenAIImagePaidResultError(
@@ -2159,6 +2168,7 @@ def download_videos(
max_clip_duration: int = 5,
match_script_order: bool = False,
progress_callback: Callable[[float], None] | None = None,
+ image_motion: ImageMotion = ImageMotion.zoom,
) -> List[str]:
"""
搜索并下载覆盖配音时长所需的素材,返回本地文件路径。
@@ -2265,6 +2275,7 @@ def download_videos(
audio_duration=audio_duration,
max_clip_duration=max_clip_duration,
material_directory=material_directory,
+ image_motion=image_motion,
)
if match_script_order:
diff --git a/app/services/task.py b/app/services/task.py
index 7a45dbb4..aa6a7754 100644
--- a/app/services/task.py
+++ b/app/services/task.py
@@ -730,7 +730,9 @@ def get_video_materials(
if params.video_source == "local":
logger.info("\n\n## preprocess local materials")
materials = video.preprocess_video(
- materials=params.video_materials, clip_duration=params.video_clip_duration
+ materials=params.video_materials,
+ clip_duration=params.video_clip_duration,
+ image_motion=params.image_motion,
)
if not materials:
_mark_task_failed(
@@ -819,6 +821,7 @@ def get_video_materials(
audio_duration=audio_duration * params.video_count,
max_clip_duration=params.video_clip_duration,
match_script_order=params.match_materials_to_script,
+ image_motion=params.image_motion,
# 素材阶段占用 40%~50%:每下完一个文件推进一次,慢速网络下
# 进度条不再长时间停在 40%。
progress_callback=_stage_progress_reporter(task_id, 40, 50),
diff --git a/app/services/video.py b/app/services/video.py
index c5f7a518..13e042c5 100644
--- a/app/services/video.py
+++ b/app/services/video.py
@@ -33,6 +33,7 @@ from PIL import Image, ImageColor, ImageDraw, ImageFont, ImageOps
from app.config import config
from app.models import const
from app.models.schema import (
+ ImageMotion,
MaterialInfo,
VideoAspect,
VideoConcatMode,
@@ -1890,31 +1891,31 @@ def generate_video(
return bgm_mix_succeeded
-def render_image_zoom_video(image_path: str, clip_duration: int = 5) -> str:
+def render_image_zoom_video(
+ image_path: str, clip_duration: int = 5, image_motion: ImageMotion = ImageMotion.zoom
+) -> str:
"""
- 将单张本地图片渲染为带缓慢放大效果的 mp4 片段,返回输出文件路径。
+ 将单张图片渲染为静止或缓慢放大的 mp4 片段,返回输出文件路径。
local 素材预处理和 OpenAI 兼容文生图素材共用这段"图片 → 片段"渲染
- 逻辑:ImageClip 按 clip_duration 固定时长播放,并叠加每秒约 3% 的
- 动态放大,避免静态画面在成片中显得呆板。渲染异常由调用方按各自
- 素材源的失败约定处理。
+ 逻辑:默认保留每秒约 3% 的动态放大,静止模式只固定播放时长。
+ 此设置与片段转场独立。渲染异常由调用方按各自素材源的失败约定处理。
"""
+ motion = ImageMotion(image_motion)
clip, _ = _open_image_clip_with_fallback(image_path)
clip = clip.with_duration(clip_duration).with_position("center")
temp_path = ""
try:
- # Apply a zoom effect using the resize method.
- # A lambda function is used to make the zoom effect dynamic over time.
- # The zoom effect starts from the original size and gradually scales up to 120%.
- # t represents the current time, and clip.duration is the total duration of the clip.
- # Note: 1 represents 100% size, so 1.2 represents 120%.
- zoom_clip = clip.resized(
- lambda t: 1 + (clip_duration * 0.03) * (t / clip.duration)
+ # 只有放大模式改变画面尺寸,静止模式保留原始构图及透明背景处理。
+ image_clip = (
+ clip.resized(lambda t: 1 + (clip_duration * 0.03) * (t / clip.duration))
+ if motion == ImageMotion.zoom
+ else clip
)
# Optionally, create a composite video clip containing the zoomed clip.
# This is useful if you want to add other elements to the video.
- final_clip = CompositeVideoClip([zoom_clip], bg_color=(0, 0, 0))
+ final_clip = CompositeVideoClip([image_clip], bg_color=(0, 0, 0))
try:
# The duration changes the rendered content, so it must be part of
# the output identity. Different tasks may render the same image
@@ -1928,7 +1929,8 @@ def render_image_zoom_video(image_path: str, clip_duration: int = 5) -> str:
source_identity = source_digest.hexdigest()
duration_identity = hashlib.sha256(str(clip_duration).encode()).hexdigest()[:16]
video_file = os.path.join(
- os.path.dirname(image_path), f"{source_identity}.zoom-{duration_identity}.mp4"
+ os.path.dirname(image_path),
+ f"{source_identity}.{motion.value}-{duration_identity}.mp4",
)
descriptor, temp_path = tempfile.mkstemp(
prefix=".image-zoom-",
@@ -1950,7 +1952,9 @@ def render_image_zoom_video(image_path: str, clip_duration: int = 5) -> str:
delete_files(temp_path)
-def preprocess_video(materials: List[MaterialInfo], clip_duration=4):
+def preprocess_video(
+ materials: List[MaterialInfo], clip_duration=4, image_motion: ImageMotion = ImageMotion.zoom
+):
# WebUI 在某些二次生成场景下可能传入空素材列表,这里直接返回空结果,避免抛出 NoneType 异常。
if not materials:
return []
@@ -2020,7 +2024,7 @@ def preprocess_video(materials: List[MaterialInfo], clip_duration=4):
# 用于导出的图片片段。
close_clip(clip)
video_file = render_image_zoom_video(
- material_source_path, clip_duration
+ material_source_path, clip_duration, image_motion=image_motion
)
material.url = video_file
logger.success(f"image processed: {video_file}")
diff --git a/cli.py b/cli.py
index b49c07ec..53ed0aee 100644
--- a/cli.py
+++ b/cli.py
@@ -221,6 +221,13 @@ def _video_fit_mode(value: str) -> str:
return normalized
+def _image_motion(value: str) -> str:
+ normalized = value.strip().lower()
+ if normalized not in {"static", "zoom"}:
+ raise argparse.ArgumentTypeError("image-motion must be one of: static, zoom")
+ return normalized
+
+
# 只有这些取值会请求外部音乐生成服务,因此只有它们消费配乐提示词。取值需与
# app/services/task.py 的 _VIDEO_MUSIC_PROVIDERS 保持一致;集中定义后,新增
# 配乐供应商不会再次遗漏 --bgm-type 校验、批量校验和提示词参数。
@@ -434,6 +441,13 @@ Batch manifests:
default=None,
help="source clip concatenation order (default: random)",
)
+ video_group.add_argument(
+ "--image-motion",
+ type=_image_motion,
+ choices=["static", "zoom"],
+ default=None,
+ help="motion for local or AI-generated images (default: zoom)",
+ )
video_group.add_argument(
"--video-transition-mode",
type=_transition_mode,
@@ -941,6 +955,7 @@ def build_video_params(args: argparse.Namespace) -> VideoParams:
"video_script_prompt",
"custom_system_prompt",
"video_fit_mode",
+ "image_motion",
"video_concat_mode",
"video_transition_mode",
"video_clip_duration",
@@ -979,6 +994,7 @@ def build_video_params(args: argparse.Namespace) -> VideoParams:
ui_defaults = (
("video_clip_speed", float, _clip_speed),
("video_fit_mode", str, _video_fit_mode),
+ ("image_motion", str, _image_motion),
("font_name", str, None),
("subtitle_display_mode", str, _subtitle_display_mode),
("subtitle_animation", str, _subtitle_animation),
diff --git a/test/services/test_image_motion.py b/test/services/test_image_motion.py
new file mode 100644
index 00000000..bbca94f8
--- /dev/null
+++ b/test/services/test_image_motion.py
@@ -0,0 +1,152 @@
+"""验证图片动态效果的兼容默认值、参数传递和真实渲染像素。"""
+
+from pathlib import Path
+from unittest.mock import patch
+
+import numpy as np
+import pytest
+from moviepy import VideoFileClip
+from PIL import Image, ImageDraw
+from pydantic import ValidationError
+
+import cli
+from app.config import config
+from app.models.schema import ImageMotion, MaterialInfo, TaskVideoRequest, VideoParams
+from app.services import material, task, video
+
+
+def test_old_requests_keep_zoom_and_invalid_motion_is_rejected():
+ assert TaskVideoRequest(video_subject="old task").image_motion == ImageMotion.zoom
+ assert TaskVideoRequest(video_subject="still", image_motion="static").image_motion == ImageMotion.static
+ with pytest.raises(ValidationError, match="image_motion"):
+ TaskVideoRequest(video_subject="bad", image_motion="unknown")
+
+
+def test_api_motion_reaches_task_queue_and_rejects_invalid_input():
+ from fastapi.testclient import TestClient
+ from app import asgi
+ from app.controllers.v1 import video as controller
+
+ with (
+ patch.dict(config.app, {"api_key": ""}),
+ patch.object(controller.task_manager, "add_task") as enqueue,
+ patch.object(controller.sm.state, "update_task"),
+ ):
+ client = TestClient(asgi.app)
+ for value, expected in ((None, "zoom"), ("static", "static"), ("zoom", "zoom")):
+ payload = {"video_subject": "test", "subtitle_enabled": False}
+ if value is not None:
+ payload["image_motion"] = value
+ response = client.post("/api/v1/videos", json=payload)
+ assert response.status_code == 200
+ assert enqueue.call_args.kwargs["params"].image_motion == expected
+ enqueue.reset_mock()
+ response = client.post("/api/v1/videos", json={"video_subject": "test", "image_motion": "bad"})
+ assert response.status_code == 400
+ enqueue.assert_not_called()
+
+
+def test_cli_motion_uses_saved_value_but_explicit_argument_wins():
+ args = ["--video-subject", "test"]
+ with patch.dict(config.ui, {}, clear=True):
+ assert cli.build_video_params(cli.parse_args(args)).image_motion == ImageMotion.zoom
+ with patch.dict(config.ui, {"image_motion": "static"}, clear=True):
+ assert cli.build_video_params(cli.parse_args(args)).image_motion == ImageMotion.static
+ explicit = cli.build_video_params(cli.parse_args(args + ["--image-motion", "zoom"]))
+ assert explicit.image_motion == ImageMotion.zoom
+ with patch.dict(config.ui, {"image_motion": "bad"}, clear=True):
+ assert cli.build_video_params(cli.parse_args(args)).image_motion == ImageMotion.zoom
+ with pytest.raises(SystemExit):
+ cli.parse_args(args + ["--image-motion", "bad"])
+
+
[email protected]("motion", [ImageMotion.static, ImageMotion.zoom])
+def test_task_passes_motion_to_local_and_generated_images(motion):
+ params = VideoParams(video_subject="test", video_source="local", image_motion=motion)
+ with patch.object(video, "preprocess_video", return_value=[MaterialInfo(url="local.mp4")]) as preprocess:
+ assert task.get_video_materials("motion-test", params, [], 2) == ["local.mp4"]
+ assert preprocess.call_args.kwargs["image_motion"] == motion
+ params.video_source = "openai_image"
+ with patch.object(material, "download_videos", return_value=["generated.mp4"]) as download:
+ assert task.get_video_materials("motion-test", params, ["scene"], 2) == ["generated.mp4"]
+ assert download.call_args.kwargs["image_motion"] == motion
+
+
+def test_generated_image_motion_reaches_renderer_without_extra_paid_requests(tmp_path):
+ item = MaterialInfo(provider="openai_image", url=str(tmp_path / "scene.png"), duration=2)
+ with (
+ patch.object(material, "generate_images_openai", return_value=[item]) as generate,
+ patch.object(video, "render_image_zoom_video", return_value="static.mp4") as render,
+ patch.object(material, "_persist_material_sources"),
+ ):
+ paths = material.download_videos(
+ "motion-test", ["first", "unused"], source="openai_image",
+ audio_duration=2, max_clip_duration=2, image_motion="static",
+ )
+ assert paths == ["static.mp4"]
+ generate.assert_called_once()
+ assert render.call_args.kwargs["image_motion"] == "static"
+
+
+def test_invalid_motion_is_rejected_before_paid_image_generation():
+ with patch.object(material, "generate_images_openai") as generate:
+ with pytest.raises(ValueError, match="ImageMotion"):
+ material.download_videos(
+ "motion-test", ["scene"], source="openai_image",
+ audio_duration=2, image_motion="invalid",
+ )
+ generate.assert_not_called()
+
+
+def _white_width(frame):
+ _, x = np.where(frame.min(axis=2) > 220)
+ return int(x.max() - x.min() + 1)
+
+
+def test_native_static_pixels_zoom_default_and_output_identity(tmp_path):
+ source = tmp_path / "scene.png"
+ image = Image.new("RGB", (256, 256), (220, 0, 0))
+ ImageDraw.Draw(image).rectangle((80, 80, 175, 175), fill="white")
+ image.save(source)
+ default = video.render_image_zoom_video(str(source), 3)
+ zoom_bytes = Path(default).read_bytes()
+ still = video.render_image_zoom_video(str(source), 3, image_motion="static")
+ assert still != default
+ assert Path(default).read_bytes() == zoom_bytes
+ with VideoFileClip(still, audio=False) as clip:
+ first, last = clip.get_frame(0), clip.get_frame(2.5)
+ assert clip.size == [256, 256]
+ assert clip.duration == pytest.approx(3, abs=0.04)
+ assert abs(_white_width(first) - _white_width(last)) <= 1
+ assert np.abs(first.astype(float) - last).mean() < 1
+ with VideoFileClip(default, audio=False) as clip:
+ assert _white_width(clip.get_frame(2.5)) > _white_width(clip.get_frame(0)) + 4
+ assert video.render_image_zoom_video(str(source), 3, image_motion="zoom") == default
+
+
+def test_static_image_preserves_alpha_composition(tmp_path):
+ source = tmp_path / "alpha.png"
+ Image.new("RGBA", (64, 64), (240, 0, 0, 128)).save(source)
+ output = video.render_image_zoom_video(str(source), 0.2, image_motion="static")
+ with VideoFileClip(output, audio=False) as clip:
+ pixel = clip.get_frame(0)[32, 32]
+ assert 110 <= pixel[0] <= 130
+ assert max(pixel[1:]) < 5
+
+
[email protected]("motion", ["static", "zoom"])
+def test_failed_motion_render_keeps_previous_output_and_removes_partial_file(tmp_path, motion):
+ source = tmp_path / "scene.png"
+ Image.new("RGB", (64, 64), "red").save(source)
+ output = video.render_image_zoom_video(str(source), 0.2, image_motion=motion)
+ completed = Path(output).read_bytes()
+
+ def interrupted_write(_clip, filename, **_kwargs):
+ Path(filename).write_bytes(b"partial video")
+ raise RuntimeError("encoder interrupted")
+
+ with patch.object(video.CompositeVideoClip, "write_videofile", interrupted_write):
+ with pytest.raises(RuntimeError, match="encoder interrupted"):
+ video.render_image_zoom_video(str(source), 0.2, image_motion=motion)
+ assert Path(output).read_bytes() == completed
+ assert not list(tmp_path.glob(".image-zoom-*.mp4"))
diff --git a/test/services/test_material_openai_image.py b/test/services/test_material_openai_image.py
index d954f7d5..9de8dec7 100644
--- a/test/services/test_material_openai_image.py
+++ b/test/services/test_material_openai_image.py
@@ -742,7 +742,7 @@ class TestOpenAIImageProvider(unittest.TestCase):
def fake_generate(search_term, minimum_duration, video_aspect, save_dir=""):
return generated[search_term]
- def fake_render(image_path, clip_duration):
+ def fake_render(image_path, clip_duration, image_motion="zoom"):
return f"{image_path}.mp4"
with (
@@ -895,7 +895,7 @@ class TestOpenAIImageProvider(unittest.TestCase):
def fake_generate(search_term, minimum_duration, video_aspect, save_dir=""):
return generated[search_term]
- def fake_render(image_path, clip_duration):
+ def fake_render(image_path, clip_duration, image_motion="zoom"):
return f"{image_path}.mp4"
with (
diff --git a/test/services/test_webui_image_motion.py b/test/services/test_webui_image_motion.py
new file mode 100644
index 00000000..0dacc803
--- /dev/null
+++ b/test/services/test_webui_image_motion.py
@@ -0,0 +1,113 @@
+"""用真实 Streamlit 控件验证图片效果选择、恢复与语言切换。"""
+
+from pathlib import Path
+from unittest.mock import patch
+
+import pytest
+from streamlit.testing.v1 import AppTest
+
+from app.config import config
+from app.models.schema import VideoParams
+from app.services import state as sm, version_checker
+
+
+WEBUI_MAIN = Path(__file__).resolve().parents[2] / "webui" / "Main.py"
+
+
+def _select(app, key):
+ return next(v for v in app.selectbox if v.key == key or str(v.key).startswith(key + "_"))
+
+
+def test_motion_widget_persists_and_restores_old_and_new_tasks(tmp_path):
+ with (
+ patch.object(config, "app", {"video_source": "local", "llm_provider": "openai"}),
+ patch.object(config, "ui", {"language": "zh", "voice_mode": "none"}),
+ patch.object(config, "try_save_config", return_value=True),
+ patch.object(sm.state, "get_all_tasks", return_value=([], 0)),
+ patch.object(version_checker, "poll_available_update", return_value=version_checker.UpdateCheckSnapshot(complete=True)),
+ patch("app.utils.utils.task_dir", return_value=str(tmp_path)),
+ ):
+ app = AppTest.from_file(str(WEBUI_MAIN), default_timeout=60)
+ app.session_state["ui_language"] = "zh"
+ app.run()
+ assert not app.exception
+ widget = _select(app, "image_motion_select")
+ assert widget.label == "图片动态效果"
+ assert widget.value == "zoom"
+ assert widget.options == ["静止", "缓慢放大"]
+ widget.select("static").run()
+ assert not app.exception
+ assert config.ui["image_motion"] == "static"
+ app.session_state["video_source_select_zh"] = "pexels"
+ app.run()
+ assert not app.exception
+ assert not any(str(v.key).startswith("image_motion_select") for v in app.selectbox)
+ app.session_state["video_source_select_zh"] = "openai_image"
+ app.run()
+ assert not app.exception
+ assert _select(app, "image_motion_select").value == "static"
+
+ # 模拟任务恢复:新增字段必须恢复,旧任务缺失字段必须回到原来的放大效果。
+ params = VideoParams(video_subject="restore", video_source="local", image_motion="static").model_dump(mode="json")
+ for expected in ("static", "zoom"):
+ if expected == "zoom":
+ params.pop("image_motion")
+ app.session_state["task_restore_payload"] = {"params": params, "task_id": "motion-test"}
+ app.run()
+ assert not app.exception
+ assert _select(app, "image_motion_select").value == expected
+
+ _select(app, "top_language_code_selector").select("en").run()
+ assert not app.exception
+ assert _select(app, "image_motion_select").label == "Image Motion"
+
+
[email protected]("rebuild", ["language", "source"])
+def test_motion_selection_survives_rebuild_while_generation_holds_config(tmp_path, rebuild):
+ import threading
+
+ locked = threading.Event()
+ release = threading.Event()
+
+ def active_generation():
+ with config.runtime_config_lock():
+ locked.set()
+ assert release.wait(30)
+
+ with (
+ patch.object(config, "app", {"video_source": "local", "llm_provider": "openai"}),
+ patch.object(config, "ui", {"language": "zh", "voice_mode": "none"}),
+ patch.object(config, "try_save_config", return_value=True),
+ patch.object(config, "save_config"),
+ patch.object(sm.state, "get_all_tasks", return_value=([], 0)),
+ patch.object(version_checker, "poll_available_update", return_value=version_checker.UpdateCheckSnapshot(complete=True)),
+ patch("app.utils.utils.task_dir", return_value=str(tmp_path)),
+ ):
+ app = AppTest.from_file(str(WEBUI_MAIN), default_timeout=60)
+ app.session_state["ui_language"] = "zh"
+ app.run()
+ assert not app.exception
+ worker = threading.Thread(target=active_generation)
+ worker.start()
+ try:
+ assert locked.wait(5)
+ _select(app, "image_motion_select").select("static").run()
+ assert not app.exception
+ assert _select(app, "image_motion_select").value == "static"
+ assert config.ui["image_motion"] == "zoom"
+ assert config.snapshot_config_with_pending(config.ui)["image_motion"] == "static"
+ if rebuild == "language":
+ _select(app, "top_language_code_selector").select("en").run()
+ else:
+ app.session_state["video_source_select_zh"] = "pexels"
+ app.run()
+ assert not app.exception
+ app.session_state["video_source_select_zh"] = "local"
+ app.run()
+ assert not app.exception
+ assert _select(app, "image_motion_select").value == "static"
+ assert config.snapshot_config_with_pending(config.ui)["image_motion"] == "static"
+ finally:
+ release.set()
+ worker.join(5)
+ assert config.ui["image_motion"] == "static"
diff --git a/test/services/test_webui_settings_transfer.py b/test/services/test_webui_settings_transfer.py
index a3901553..7007a4bd 100644
--- a/test/services/test_webui_settings_transfer.py
+++ b/test/services/test_webui_settings_transfer.py
@@ -158,6 +158,7 @@ def test_settings_preset_round_trip_preserves_generation_settings():
video_subject="a cat",
video_aspect="9:16",
video_fit_mode="contain",
+ image_motion="static",
font_size=48,
stroke_width=2.5,
voice_volume=0.8,
@@ -171,6 +172,7 @@ def test_settings_preset_round_trip_preserves_generation_settings():
assert restored["video_subject"] == "a cat"
assert restored["video_aspect"] == "9:16"
assert restored["video_fit_mode"] == "contain"
+ assert restored["image_motion"] == "static"
assert restored["font_size"] == 48
assert restored["stroke_width"] == 2.5
assert restored["voice_volume"] == 0.8
diff --git a/webui/Main.py b/webui/Main.py
index c7274c48..c61c44af 100644
--- a/webui/Main.py
+++ b/webui/Main.py
@@ -37,6 +37,7 @@ from app.models.llm_provider import (
normalize_provider_override,
)
from app.models.schema import (
+ ImageMotion,
MaterialInfo,
VideoAspect,
VideoConcatMode,
@@ -1562,6 +1563,9 @@ def _apply_restored_params(params):
"video_fit_mode_select",
params.get("video_fit_mode") or VideoFitMode.cover.value,
)
+ _set_stable_widget_value(
+ "image_motion_select", params.get("image_motion") or ImageMotion.zoom.value
+ )
_set_stable_widget_value(
"video_clip_duration_select", params.get("video_clip_duration", 3)
)
@@ -5404,6 +5408,28 @@ def _render_video_settings(panel, params):
"ui", "video_fit_mode", params.video_fit_mode.value
)
+ if params.video_source in {"local", "openai_image"}:
+ image_motion_labels = {
+ ImageMotion.static.value: tr("Static Image"),
+ ImageMotion.zoom.value: tr("Slow Zoom"),
+ }
+ params.image_motion = ImageMotion(
+ stable_selectbox(
+ tr("Image Motion"),
+ options=list(image_motion_labels),
+ default_value=_saved_ui_choice(
+ "image_motion", list(image_motion_labels), ImageMotion.zoom.value,
+ # 生成任务持锁时,切换语言或素材来源会重建控件;
+ # 此时必须恢复尚未落盘的最新选择,不能写回旧配置值。
+ section=config.snapshot_config_with_pending(config.ui),
+ ),
+ key="image_motion_select",
+ format_func=image_motion_labels.get,
+ help=tr("Image Motion Help"),
+ )
+ )
+ _set_runtime_config("ui", "image_motion", params.image_motion.value)
+
# MiniMax H3 的远端时长范围是 4~15 秒。选择秘塔时使用完整能力
# 范围,既避免 2/3 秒被按 4 秒计费,也让 WebUI 与 CLI、服务层一致。
if params.video_source == "metaso_minimax":
diff --git a/webui/i18n/az.json b/webui/i18n/az.json
index 130fa75d..9cb59a7d 100644
--- a/webui/i18n/az.json
+++ b/webui/i18n/az.json
@@ -100,6 +100,10 @@
"Fill and Crop": "Doldurma (mərkəzdən kəsmə)",
"Fit with Black Bars": "Sığdırma (qara zolaqlarla)",
"Video Fit Mode Help": "Doldurma rejimi bütün çıxış sahəsindən istifadə edir və artıq kənarları kəsir. Sığdırma rejimi mənbə kadrını bütöv saxlayır və lazım olduqda qara zolaqlar əlavə edir.",
+ "Image Motion": "Şəkil hərəkəti",
+ "Static Image": "Hərəkətsiz",
+ "Slow Zoom": "Yavaş böyütmə",
+ "Image Motion Help": "Yalnız şəkillərə təsir edir. Kliplərarası keçidlər ayrıca tənzimlənir.",
"Clip Duration": "Klipin maksimum müddəti (saniyə)",
"Clip Duration Help": "Videonun ümumi müddətini deyil, kompozisiyada istifadə olunan hər klipin müddətini tənzimləyir.",
"Clip Speed": "Klipin sürəti",
diff --git a/webui/i18n/ca.json b/webui/i18n/ca.json
index f9ddf4c1..09479c9c 100644
--- a/webui/i18n/ca.json
+++ b/webui/i18n/ca.json
@@ -106,6 +106,10 @@
"Fill and Crop": "Omple (retall central)",
"Fit with Black Bars": "Enquadra (franges negres)",
"Video Fit Mode Help": "Omple cobreix tota la imatge de sortida i retalla les vores sobrants. Enquadra conserva la imatge sencera i afegeix franges negres si cal.",
+ "Image Motion": "Moviment de les imatges",
+ "Static Image": "Estàtica",
+ "Slow Zoom": "Ampliació lenta",
+ "Image Motion Help": "Només afecta les imatges fixes. Les transicions entre clips es configuren per separat.",
"Clip Duration": "Durada màxima de cada clip (s)",
"Clip Duration Help": "Limita la durada de cada clip, no la durada total del vídeo.",
"Clip Speed": "Velocitat dels clips",
diff --git a/webui/i18n/de.json b/webui/i18n/de.json
index e753cfd1..5f99cfe5 100644
--- a/webui/i18n/de.json
+++ b/webui/i18n/de.json
@@ -106,6 +106,10 @@
"Fill and Crop": "Ausfüllen (zentriert zuschneiden)",
"Fit with Black Bars": "Einpassen (schwarze Balken)",
"Video Fit Mode Help": "Ausfüllen nutzt die gesamte Ausgabefläche und schneidet überstehende Ränder ab. Einpassen erhält das vollständige Quellbild und fügt bei Bedarf schwarze Balken hinzu.",
+ "Image Motion": "Bildbewegung",
+ "Static Image": "Statisch",
+ "Slow Zoom": "Langsamer Zoom",
+ "Image Motion Help": "Gilt nur für Standbilder. Übergänge zwischen Clips werden separat eingestellt.",
"Clip Duration": "Max. Dauer pro Clip (Sek.)",
"Clip Duration Help": "Legt die Dauer jedes verwendeten Clips fest, nicht die Gesamtlänge des Videos.",
"Clip Speed": "Clip-Geschwindigkeit",
diff --git a/webui/i18n/en.json b/webui/i18n/en.json
index 15c88f58..cf3d3d19 100644
--- a/webui/i18n/en.json
+++ b/webui/i18n/en.json
@@ -110,6 +110,10 @@
"Fill and Crop": "Fill (center crop)",
"Fit with Black Bars": "Fit (black bars)",
"Video Fit Mode Help": "Fill uses the entire output canvas and crops excess edges. Fit preserves the complete source frame and adds black bars when needed.",
+ "Image Motion": "Image Motion",
+ "Static Image": "Static",
+ "Slow Zoom": "Slow Zoom",
+ "Image Motion Help": "Only affects still images. Clip transitions are configured separately.",
"Clip Duration": "Maximum Clip Duration (seconds)",
"Clip Duration Help": "Controls the duration of each clip used in the composition, not the total video length.",
"Clip Speed": "Clip Speed",
diff --git a/webui/i18n/es.json b/webui/i18n/es.json
index 6d74deff..400a3ab1 100644
--- a/webui/i18n/es.json
+++ b/webui/i18n/es.json
@@ -106,6 +106,10 @@
"Fill and Crop": "Rellenar (recorte central)",
"Fit with Black Bars": "Ajustar (barras negras)",
"Video Fit Mode Help": "Rellenar ocupa todo el lienzo de salida y recorta los bordes sobrantes. Ajustar conserva el encuadre completo y añade barras negras cuando sea necesario.",
+ "Image Motion": "Movimiento de imágenes",
+ "Static Image": "Estática",
+ "Slow Zoom": "Zoom lento",
+ "Image Motion Help": "Solo afecta a las imágenes fijas. Las transiciones entre clips se configuran por separado.",
"Clip Duration": "Duración máxima por clip (s)",
"Clip Duration Help": "Controla la duración de cada clip utilizado, no la duración total del vídeo.",
"Clip Speed": "Velocidad del clip",
diff --git a/webui/i18n/fr.json b/webui/i18n/fr.json
index 4b03707e..54462726 100644
--- a/webui/i18n/fr.json
+++ b/webui/i18n/fr.json
@@ -106,6 +106,10 @@
"Fill and Crop": "Remplir (recadrage centré)",
"Fit with Black Bars": "Ajuster (bandes noires)",
"Video Fit Mode Help": "Remplir utilise tout le cadre de sortie et coupe les bords excédentaires. Ajuster conserve l’image source complète et ajoute des bandes noires si nécessaire.",
+ "Image Motion": "Animation des images",
+ "Static Image": "Image fixe",
+ "Slow Zoom": "Zoom lent",
+ "Image Motion Help": "Concerne uniquement les images fixes. Les transitions entre les clips se règlent séparément.",
"Clip Duration": "Durée maximale d'un extrait (secondes)",
"Clip Duration Help": "Contrôle la durée de chaque extrait utilisé dans la composition, et non la durée totale du rendu.",
"Clip Speed": "Vitesse de l'extrait",
diff --git a/webui/i18n/id.json b/webui/i18n/id.json
index 26bdbbdd..5351b35d 100644
--- a/webui/i18n/id.json
+++ b/webui/i18n/id.json
@@ -106,6 +106,10 @@
"Fill and Crop": "Penuhi (potong tengah)",
"Fit with Black Bars": "Sesuaikan (bilah hitam)",
"Video Fit Mode Help": "Penuhi memakai seluruh kanvas keluaran dan memotong tepi berlebih. Sesuaikan mempertahankan bingkai sumber lengkap dan menambahkan bilah hitam bila diperlukan.",
+ "Image Motion": "Gerakan gambar",
+ "Static Image": "Diam",
+ "Slow Zoom": "Perbesar perlahan",
+ "Image Motion Help": "Hanya memengaruhi gambar diam. Transisi antarklip diatur secara terpisah.",
"Clip Duration": "Durasi Maks. per Klip (detik)",
"Clip Duration Help": "Menentukan durasi setiap klip yang digunakan, bukan durasi total video.",
"Clip Speed": "Kecepatan Klip",
diff --git a/webui/i18n/it.json b/webui/i18n/it.json
index c3e808c4..88d9631c 100644
--- a/webui/i18n/it.json
+++ b/webui/i18n/it.json
@@ -106,6 +106,10 @@
"Fill and Crop": "Riempi (ritaglio centrale)",
"Fit with Black Bars": "Adatta (bande nere)",
"Video Fit Mode Help": "Riempi usa l’intera area di uscita e ritaglia i bordi in eccesso. Adatta conserva l’inquadratura sorgente completa e aggiunge bande nere quando necessario.",
+ "Image Motion": "Movimento delle immagini",
+ "Static Image": "Statica",
+ "Slow Zoom": "Zoom lento",
+ "Image Motion Help": "Si applica solo alle immagini fisse. Le transizioni tra clip si impostano separatamente.",
"Clip Duration": "Durata massima per clip (secondi)",
"Clip Duration Help": "Controlla la durata di ciascuna clip utilizzata nella composizione, non la durata totale del video.",
"Clip Speed": "Velocità della clip",
diff --git a/webui/i18n/ko.json b/webui/i18n/ko.json
index 1860a1e2..c85bd30e 100644
--- a/webui/i18n/ko.json
+++ b/webui/i18n/ko.json
@@ -106,6 +106,10 @@
"Fill and Crop": "채우기 (가운데 자르기)",
"Fit with Black Bars": "맞추기 (검은 여백)",
"Video Fit Mode Help": "채우기는 출력 화면 전체를 사용하고 넘치는 가장자리를 자릅니다. 맞추기는 원본 화면 전체를 유지하고 필요할 때 검은 여백을 추가합니다.",
+ "Image Motion": "이미지 움직임",
+ "Static Image": "정지",
+ "Slow Zoom": "느린 확대",
+ "Image Motion Help": "정지 이미지에만 적용됩니다. 클립 간 전환은 별도로 설정합니다.",
"Clip Duration": "클립 최대 길이 (초)",
"Clip Duration Help": "합성에 사용되는 각 클립의 길이를 조절합니다. 전체 영상 길이가 아닙니다.",
"Clip Speed": "클립 재생 속도",
diff --git a/webui/i18n/pt.json b/webui/i18n/pt.json
index f174e4c7..ea839af7 100644
--- a/webui/i18n/pt.json
+++ b/webui/i18n/pt.json
@@ -106,6 +106,10 @@
"Fill and Crop": "Preencher (recorte central)",
"Fit with Black Bars": "Ajustar (barras pretas)",
"Video Fit Mode Help": "Preencher ocupa toda a tela e recorta o excesso nas bordas. Ajustar preserva o quadro completo e adiciona barras pretas quando necessário.",
+ "Image Motion": "Movimento das imagens",
+ "Static Image": "Estática",
+ "Slow Zoom": "Zoom lento",
+ "Image Motion Help": "Afeta apenas imagens estáticas. As transições entre clipes são configuradas separadamente.",
"Clip Duration": "Duração máxima por clipe (s)",
"Clip Duration Help": "Define a duração de cada clipe utilizado, não a duração total do vídeo.",
"Clip Speed": "Velocidade do clipe",
diff --git a/webui/i18n/ru.json b/webui/i18n/ru.json
index 3745587f..eaec6bda 100644
--- a/webui/i18n/ru.json
+++ b/webui/i18n/ru.json
@@ -108,6 +108,10 @@
"Fill and Crop": "Заполнить (обрезка по центру)",
"Fit with Black Bars": "Вместить (чёрные полосы)",
"Video Fit Mode Help": "Заполнение использует весь выходной кадр и обрезает лишние края. Вмещение сохраняет исходный кадр целиком и при необходимости добавляет чёрные полосы.",
+ "Image Motion": "Движение изображения",
+ "Static Image": "Без движения",
+ "Slow Zoom": "Медленное увеличение",
+ "Image Motion Help": "Применяется только к изображениям. Переходы между клипами настраиваются отдельно.",
"Clip Duration": "Макс. длительность фрагмента (сек.)",
"Clip Duration Help": "Определяет длительность каждого используемого фрагмента, а не общую длину видео.",
"Clip Speed": "Скорость фрагментов",
diff --git a/webui/i18n/tr.json b/webui/i18n/tr.json
index 1128e6e0..7096d420 100644
--- a/webui/i18n/tr.json
+++ b/webui/i18n/tr.json
@@ -106,6 +106,10 @@
"Fill and Crop": "Doldur (ortadan kırp)",
"Fit with Black Bars": "Sığdır (siyah çubuklar)",
"Video Fit Mode Help": "Doldur, çıkış alanının tamamını kullanır ve taşan kenarları kırpar. Sığdır, kaynak kadrajın tamamını korur ve gerektiğinde siyah çubuklar ekler.",
+ "Image Motion": "Görüntü hareketi",
+ "Static Image": "Sabit",
+ "Slow Zoom": "Yavaş yakınlaştırma",
+ "Image Motion Help": "Yalnızca sabit görüntüleri etkiler. Klip geçişleri ayrıca ayarlanır.",
"Clip Duration": "Maks. Klip Süresi (sn.)",
"Clip Duration Help": "Toplam video süresini değil, kullanılan her klibin süresini belirler.",
"Clip Speed": "Klip Hızı",
diff --git a/webui/i18n/vi.json b/webui/i18n/vi.json
index c5e576fa..3a302202 100644
--- a/webui/i18n/vi.json
+++ b/webui/i18n/vi.json
@@ -106,6 +106,10 @@
"Fill and Crop": "Lấp đầy (cắt giữa)",
"Fit with Black Bars": "Vừa khung (viền đen)",
"Video Fit Mode Help": "Lấp đầy sử dụng toàn bộ khung đầu ra và cắt phần cạnh dư. Vừa khung giữ nguyên toàn bộ hình nguồn và thêm viền đen khi cần.",
+ "Image Motion": "Chuyển động hình ảnh",
+ "Static Image": "Tĩnh",
+ "Slow Zoom": "Phóng to chậm",
+ "Image Motion Help": "Chỉ áp dụng cho ảnh tĩnh. Chuyển cảnh giữa các đoạn được cài đặt riêng.",
"Clip Duration": "Thời Lượng Tối Đa Mỗi Clip (giây)",
"Clip Duration Help": "Quy định thời lượng của từng clip được sử dụng, không phải tổng thời lượng video.",
"Clip Speed": "Tốc độ clip",
diff --git a/webui/i18n/zh.json b/webui/i18n/zh.json
index 6cf48c0f..62f021a3 100644
--- a/webui/i18n/zh.json
+++ b/webui/i18n/zh.json
@@ -110,6 +110,10 @@
"Fill and Crop": "填满(居中裁剪)",
"Fit with Black Bars": "完整显示(黑边)",
"Video Fit Mode Help": "填满会占满整个输出画面并裁掉多余边缘;完整显示会保留全部素材画面,并在需要时添加黑边。",
+ "Image Motion": "图片动态效果",
+ "Static Image": "静止",
+ "Slow Zoom": "缓慢放大",
+ "Image Motion Help": "仅影响图片素材;片段之间的转场需单独设置。",
"Clip Duration": "单个片段最大时长(秒)",
"Clip Duration Help": "控制合成时使用的每个视频片段时长,不是视频总时长。",
"Clip Speed": "片段播放速度",