llm_engineering/domain/video_chunks.py

from typing import List, Optional
from pydantic import Field
from llm_engineering.domain.embedded_chunks import EmbeddedChunk

class EmbeddedVideoChunk(EmbeddedChunk):
    # Video-specific required fields
    video_id: str
    start_time: float
    end_time: float
    frame_paths: List[str]
    
    # Inherited fields with video defaults
    platform: str = Field(default="video")
    document_type: str = Field(default="video_segment")
    
    # Optional inherited fields
    document_id: Optional[str] = None
    author_id: Optional[str] = None 
    author_full_name: Optional[str] = None
    
    # Frame embeddings (initialize empty)
    frame_embeddings: List[List[float]] = Field(default_factory=list)
    
    # Main content embedding
    embedding: List[float]
    
    # Add explicit content field override
    content: str = Field(..., description="Text content from subtitles")