"""数据模型定义"""
from enum import Enum
from typing import List, Optional, Union
from pydantic import BaseModel
class VideoConcatMode(str, Enum):
"""视频拼接模式"""
random = "random"
sequential = "sequential"
class VideoTransitionMode(str, Enum):
"""视频转场模式"""
none = "none"
shuffle = "shuffle"
fade_in = "fade_in"
fade_out = "fade_out"
slide_in = "slide_in"
slide_out = "slide_out"
class VideoAspect(str, Enum):
"""视频比例"""
landscape = "16:9" # 横屏
portrait = "9:16" # 竖屏
square = "1:1" # 正方形
def to_resolution(self) -> tuple[int, int]:
"""转换为分辨率"""
if self == VideoAspect.landscape:
return 1920, 1080
elif self == VideoAspect.portrait:
return 1080, 1920
elif self == VideoAspect.square:
return 1080, 1080
return 1080, 1920
class MaterialInfo(BaseModel):
"""素材信息"""
provider: str = "pexels"
url: str = ""
duration: int = 0
class VideoParams(BaseModel):
"""视频生成参数"""
video_subject: str
video_script: Optional[str] = None
video_terms: Optional[Union[str, List[str]]] = None
video_aspect: VideoAspect = VideoAspect.portrait
video_concat_mode: VideoConcatMode = VideoConcatMode.random
video_transition_mode: Optional[VideoTransitionMode] = None
video_clip_duration: int = 5
video_count: int = 1
video_source: str = "pexels"
video_materials: Optional[List[MaterialInfo]] = None
custom_audio_file: Optional[str] = None
video_language: Optional[str] = ""
voice_name: str = "zh-CN-XiaoxiaoNeural-Female"
voice_volume: float = 1.0
voice_rate: float = 1.0
bgm_type: str = "random"
bgm_file: Optional[str] = None
bgm_volume: float = 0.2
subtitle_enabled: bool = True
subtitle_position: str = "bottom" # top, bottom, center
custom_position: float = 70.0
font_name: str = "STHeitiMedium.ttc"
text_fore_color: str = "#FFFFFF"
text_background_color: Union[bool, str] = True
font_size: int = 60
stroke_color: str = "#000000"
stroke_width: float = 1.5
n_threads: int = 2
paragraph_number: int = 1
