FreeVideo 0.4.24

MANIFEST.in

@@ -11,6 +11,7 @@ recursive-include web/fonts *.woff2
 recursive-include web/tutorial *.webp *.mp4
 recursive-include freevideo_engine/assets *.png *.ico
 recursive-include freevideo_engine/launcher *.qml
+recursive-include freevideo_engine/agent/skill *.txt
 recursive-include freevideo_engine/launcher/licenses *.txt
 recursive-include freevideo_engine/licenses *.txt
 recursive-include constraints *.txt

THIRD_PARTY_NOTICES.md

@@ -5,6 +5,7 @@ FreeVideo code uses [Apache-2.0](LICENSE). Its dependencies and models retain th
 | Component | License |
 | --- | --- |
 | [Three.js](https://threejs.org/) r160, © 2010–2023 three.js authors (bundled whitebox renderer) | [MIT](web/vendor/three/LICENSE) |
+| [Depth Anything V2 Small](https://github.com/DepthAnything/Depth-Anything-V2/tree/a561b849ebae10a6f5ef49e26c83cbbcd36c71bf), © 2024 Depth Anything authors; DINOv2 © Meta Platforms, Inc. and affiliates. Small-only DPT/DINOv2 inference and preprocessing are adapted and bundled; Small weights are an optional download pinned at `03876f8651c73a60fe4c2c48294e09fcb6838fcf`. | [Apache-2.0 (code and Small weights)](freevideo_engine/licenses/Depth-Anything-V2-Apache-2.0.txt) |
 | Qt, PySide6 and Shiboken6 | LGPL-3.0; [component notices](freevideo_engine/launcher/licenses/NOTICE.txt) |
 | PyInstaller bootloader | GPL-2.0 with the bootloader exception |
 | CPython / psutil | PSF / BSD-3-Clause |

docs/zh-CN/使用说明.md

@@ -32,6 +32,10 @@
 - 编号顺序是：先是上传的素材，再是从上到下连接的图片，最后是音频。参考视频的声音也占一个音频编号。
 - 删除或关闭某张素材后，描述里指向它的标记会改成“[素材已移除]”，其余标记按新编号更新。
 
+## 用一句话做白模
+
+未连接外部 Agent 时，可开启聊天框中的白模按钮，用一句话描述镜头。有合适的现成场景时，会先生成白模供您确认。
+
 ## 做更长的视频
 
 有三种做法，可以单独用，也可以搭配。
@@ -46,7 +50,7 @@
 
 ### 限制
 
-- 两种做法一样，一条视频最多 8 段（第一段加 7 次接着生成）；创作面板里时长最多填 60 秒。接续的每一段至少 10 秒，整条约 110 秒。剪辑视图里逐段生成同样适用。段数越多，衔接处越容易出现变化，适合切换镜头处，可直接切镜头，用对应的参考图开始新的一段。
+- 两种做法一样，一条视频最多 8 段（第一段加 7 次接着生成），有短于 10 秒的段时最多 4 段；创作面板里时长最多填 60 秒。接续的每一段至少 5 秒，每段都是 10 秒以上时整条约 110 秒。剪辑视图里逐段生成同样适用。段数越多，衔接处越容易出现变化，适合切换镜头处，可直接切镜头，用对应的参考图开始新的一段。
 - 接着生成使用轻量档，需要 NVIDIA 显卡或 Mac，也不能和“继续二采”同时使用。
 - 单次生成的最长时长随画幅和显卡而不同，在创作面板的“可生成范围”中可查看各尺寸的最长时长和预计用时。
 - 分段生成同样需要 NVIDIA 显卡或 Mac。

freevideo_engine/agent/cli.py

@@ -16,10 +16,11 @@
 import sys
 
 from .client import ApiError, Client
-from . import schema, spec, errors
+from . import errors
 
 
 def read(path):
+    from . import schema
     text = sys.stdin.read() if path == '-' else Path(path).read_text(encoding='utf-8')
     value = json.loads(text)
     schema.validate({'type': 'object'}, value)
@@ -40,13 +41,16 @@ def error(self, message):
 
 
 def build_parser():
+    from . import schema
     parser = Parser(prog='./freevideo agent', description='Drive FreeVideo through its agent API.')
     parser.add_argument('--root', type=Path)
     parser.add_argument('--url')
     commands = parser.add_subparsers(dest='command', required=True)
     commands.add_parser('status', help='Readiness, limits, LoRAs, VLM and queue')
     commands.add_parser('guide', help='How to write H3 prompts')
     commands.add_parser('schema', help='JSON Schemas of jobs')
+    command = commands.add_parser('install-skill', help='Copy the FreeVideo skill where Codex, Gemini CLI, Cursor and Claude Code read skills')
+    command.add_argument('--target', choices=('agents', 'claude'), action='append', help='Default: both')
     command = commands.add_parser('capabilities', help='Read the studio capability cache')
     schema.cli_arguments(command, schema.contracts()['GET', '/capabilities']['query'])
     command = commands.add_parser('asset', help='Download a saved library video or thumbnail over HTTP')
@@ -159,10 +163,39 @@ def build_parser():
     return parser
 
 
+def install_skill(argv):
+    """Run `install-skill` without the job schemas, whose import loads the image libraries (numpy, Pillow).
+
+    A deep installation folder on Windows cannot load those libraries by their full path, and copying
+    the skill needs none of them. None where the command line is anything else."""
+    parser = argparse.ArgumentParser(add_help=False)
+    parser.add_argument('--root', type=Path)
+    parser.add_argument('--url')
+    parser.add_argument('command', nargs='?')
+    known, rest = parser.parse_known_args(argv)
+    if known.command != 'install-skill':
+        return None
+    parser = argparse.ArgumentParser(add_help=False, exit_on_error=False)
+    parser.add_argument('--target', choices=('agents', 'claude'), action='append')
+    try:
+        args, extra = parser.parse_known_args(rest)
+    except argparse.ArgumentError:
+        return None
+    if extra:
+        return None
+    from . import skill_install
+    print(json.dumps({'installed': skill_install.install(tuple(args.target or ('agents', 'claude')))}, indent=2))
+    return 0
+
+
 def main(argv=None):
     for stream in (sys.stdin, sys.stdout):
         if hasattr(stream, 'reconfigure'):
             stream.reconfigure(encoding='utf-8')
+    done = install_skill(sys.argv[1:] if argv is None else argv)
+    if done is not None:
+        return done
+    from . import schema, spec
     parser = build_parser()
     args = parser.parse_args(argv)
     values = vars(args)

freevideo_engine/agent/drama.py

@@ -20,7 +20,7 @@
 import time
 
 from . import spec as specs
-from ..chain_plan import MIN_PART_SECONDS
+from ..chain_plan import CONTINUE_SECONDS
 
 EXAMPLE = {
     'title': '末班车', 'name': 'the-last-train', 'aspect': '9:16', 'quality': 'light',
@@ -451,9 +451,9 @@ def intent(play, shot):
 
 
 def shot_seconds(play, shot):
-    """A shot says its length; else the film does; else 6 s, and 10 s for a shot that continues the one before
-    (a part that continues another video is at least 10 s)."""
-    return shot.get('seconds', play.get('seconds', MIN_PART_SECONDS if shot.get('continue') else 6))
+    """A shot says its length; else the film does; else 6 s, and CONTINUE_SECONDS for a shot that continues the
+    one before (a full continuing part, which keeps the video to the full number of parts)."""
+    return shot.get('seconds', play.get('seconds', CONTINUE_SECONDS if shot.get('continue') else 6))
 
 
 def shot_job(play, shot, index, ref):

freevideo_engine/agent/editor_runtime.py

@@ -16,7 +16,7 @@
 # Definition order only; evaluation starts at editor_actions like the page's
 # import does. editor_model and editor_text_model import each other: the
 # module still loading is reached only through a namespace import.
-MODULES = ('editor_picture_fields', 'editor_transport_model', 'editor_audio_model', 'editor_speed_model',
+MODULES = ('chain_rules', 'editor_picture_fields', 'editor_transport_model', 'editor_audio_model', 'editor_speed_model',
            'editor_text_model', 'editor_model', 'editor_room', 'editor_actions', 'editor_generation', 'sampling_effort',
            'chain_support', 'editor_store')
 TIME_LIMIT = 5.0

freevideo_engine/agent/errors.py

@@ -67,6 +67,7 @@
     'chain_missing': (409, True, 'The saved chain receipt is unavailable.', 'Restore the chain receipt and retry resume, or submit a new job.'),
     'join_failed': (409, True, 'The generated blocks could not be joined; completed blocks are retained.', 'POST /freevideo/agent/v1/jobs/{id}/resume to retry joining the completed blocks.'),
     'chain_too_long': (422, False, 'One continuous shot can have up to 8 parts. Make it shorter.', 'Submit a new job with a smaller seconds, or cut to a new shot with a new job without continue_from; continue_from counts the parts of the video it continues.'),
+    'chain_short_parts': (422, False, 'With a part shorter than 10 seconds, one continuous shot can have up to 4 parts. Make every part at least 10 seconds long, or use fewer parts.', 'Submit a new job whose continuing parts are at least 10 seconds, with fewer parts, or with first_frame_reference true; continue_from counts the parts of the video it continues.'),
     'chain_rejected': (409, False, 'The continuous backend rejected this generation.', 'Read GET /freevideo/agent/v1/capabilities and submit a supported generation request.'),
     'invalid_wait': (422, False, 'wait must be a finite nonnegative number of seconds.', 'Use wait=0 or wait=25; larger valid values are clamped to 25 seconds.'),
     'invalid_resume': (422, False, 'resume accepts an empty object or an optional ref only.', 'Send {} or {"ref":"resume-1"}; submit a new job to change generation parameters.'),
@@ -302,6 +303,7 @@ def safe_details(value):
     'invalid_dimensions': '图片尺寸超出了支持的范围',
     'too_many_images': '参考图片数量超过了支持的范围',
     'chain_too_long': '接续后的视频超过了支持的段数',
+    'chain_short_parts': '有短于 10 秒的段时，接续后的视频超过了支持的段数',
     'chain_rejected': '生成服务无法使用这组参数',
     'generation_failed': '生成没有完成',
     'internal_error': '没有完成',
@@ -350,6 +352,7 @@ def safe_details(value):
     'invalid_dimensions': 'The picture size is outside the supported range',
     'too_many_images': 'There are more reference pictures than supported',
     'chain_too_long': 'The continued video has more parts than supported',
+    'chain_short_parts': 'With a part shorter than 10 seconds, the continued video has more parts than supported',
     'chain_rejected': 'The generation service cannot use these settings',
     'generation_failed': 'Generation did not finish',
     'internal_error': 'Did not finish',

freevideo_engine/agent/guide.py

@@ -99,7 +99,7 @@
     'timing': [
         'Single-block frames are 24 fps on a 17n+5 grid (6 s -> 6.583 s, 10 s -> 10.125 s). '
         'Under 30 s defaults to one generation when the device supports this size and quality; unknown limits use 15 s. '
-        'Otherwise it uses as few parts as possible, each at least 10 s; how many depends on how long this computer can generate in one run, and the parts continue and join automatically. '
+        'Otherwise it uses as few parts as possible, each continuing part 10 s or longer where the length allows and never under 5 s; how many depends on how long this computer can generate in one run, and the parts continue and join automatically. '
         'Multiple blocks or continue_from need two-pass text generation: quality "light" or "medium" '
         'or sampling.two_pass=true and no preview. Their only media is media.references with at most 3 pictures, '
         'which every continuing part also sees (keeps the same people); first_frame_reference then gives way to them. '
@@ -250,9 +250,10 @@
     'Cut by default. Wherever the film can change shot, make the new shot from the reference pictures it should '
     'have (cast cards, scene pictures) instead of continuing the previous video. Continue a video only for a '
     'long take that cannot be cut, an action that must stay unbroken, or sound that must stay unbroken, and '
-    'continuing again and again is not recommended. One video has at most 8 parts (the first and 7 '
-    'continuations), each continued part at least 10 seconds, about 110 seconds in all; filling that is not '
-    'recommended. Past it, cut to a new shot.',
+    'continuing again and again is not recommended. Each continued part is at least 5 seconds. One video has at '
+    'most 8 parts (the first and 7 continuations, about 110 seconds in all) when every continued part is at '
+    'least 10 seconds; with a continued part shorter than 10 seconds it has at most 4 parts, unless '
+    'first_frame_reference is true. Filling that is not recommended. Past it, cut to a new shot.',
     'A question is not a request. When the user asks about progress or status ("is it working?", "what happened?") '
     'answer it in one sentence from what you know or from the job and white-model status, and start nothing; never '
     'use a question as a prompt. When a row carries "whitebox", the user was starting a white-model scene that is '
@@ -287,7 +288,8 @@
     '先用 people（角色图和各自的位置）做出首帧，再用它生成视频。',
     '默认切镜头。能切镜头的地方就切：新镜头直接用对应的参考图（人物卡、场景图）生成，不接上一段视频。'
     '只有不能切镜的长镜头、必须连贯的动作、必须连贯的声音，才接着上一段生成，而且不建议一段接一段地连续接。'
-    '一条视频最多 8 段（第一段加 7 次接着生成），接着生成的每一段至少 10 秒，总长约 110 秒；不建议接满，超过就切镜头。',
+    '接着生成的每一段至少 5 秒。每段都不短于 10 秒时，一条视频最多 8 段（第一段加 7 次接着生成，总长约 110 秒）；'
+    '有短于 10 秒的段时最多 4 段，first_frame_reference 为 true 时除外。不建议接满，超过就切镜头。',
     '问话不是要求。用户问进度或状态（“在做吗”“怎么了”）时，用一句话照实回答，依据是你所知道的、任务和白模的状态，'
     '什么都不要开始；不要把问句当提示词。某条 inbox 带有 whitebox 时，说明用户刚启动了白模场景，'
     '它还在准备或已经失败：照实说是哪一种，再等用户回话。',
@@ -573,8 +575,9 @@ def connection_text(value):
     'before uploading. Unknown fields are rejected and every batch is validated before it is recorded.',
     'POST /plan gives the length supported for the current size and quality. Without continuation, only a '
     'single block can run; unknown device limits use 15 s. Multiple blocks return continuation_unavailable.',
-    'One video has at most 8 parts (60 s in one request). continue_from counts the parts of the video it '
-    'continues; plan and submit refuse a longer one and name the longest seconds that still fits.',
+    'One video has at most 8 parts (60 s in one request), or 4 when a continued part is shorter than 10 s and '
+    'first_frame_reference is not true. continue_from counts the parts of the video it continues; plan and '
+    'submit refuse a longer one and name the longest seconds that still fits.',
     'features in the status reply names what this installation offers; false means not usable now (for example '
     'chain.auto while continuation is false), and a missing name means not offered here.',
 ]