MCPeasy

""" YouTube chunk model for vector similarity search """ from sqlalchemy import Column, Integer, String, Text, REAL from pgvector.sqlalchemy import Vector from .base import Base class YouTubeChunk(Base): """YouTube video chunk with embeddings for similarity search""" __tablename__ = "youtube_chunks" id = Column(Integer, primary_key=True, autoincrement=True) project_slug = Column(String(255), nullable=False, index=True) video_id = Column(String(255), nullable=False) chunk_index = Column(Integer, nullable=False) title = Column(Text, nullable=False) start_timestamp = Column(REAL, nullable=False) end_timestamp = Column(REAL, nullable=False) text = Column(Text, nullable=False) word_count = Column(Integer, nullable=False) sentence_count = Column(Integer, nullable=False) embedding = Column(Vector(1024), nullable=True) def __repr__(self): return f"<YouTubeChunk(project_slug='{self.project_slug}', video_id='{self.video_id}', chunk_index={self.chunk_index})>" def get_youtube_url(self) -> str: """Get YouTube URL with timestamp""" timestamp_seconds = int(self.start_timestamp) return f"https://www.youtube.com/watch?v={self.video_id}&t={timestamp_seconds}s" def to_dict(self) -> dict: """Convert to dictionary for JSON serialization""" return { "id": self.id, "project_slug": self.project_slug, "video_id": self.video_id, "chunk_index": self.chunk_index, "title": self.title, "start_timestamp": self.start_timestamp, "end_timestamp": self.end_timestamp, "text": self.text, "word_count": self.word_count, "sentence_count": self.sentence_count, "youtube_url": self.get_youtube_url() }