data_nexus/app/models.py

from sqlalchemy import Column, Integer, String, Text, ForeignKey, DateTime, BigInteger, UniqueConstraint, JSON
from sqlalchemy.orm import relationship
from sqlalchemy.sql import func
from ulid import ULID
from app.database import Base


def generate_ulid() -> str:
    """Generate a new ULID string."""
    return str(ULID())


class Project(Base):
    __tablename__ = "projects"

    id = Column(String(26), primary_key=True, default=generate_ulid)
    project_name = Column(String(100), unique=True, nullable=False, index=True)
    description = Column(Text, nullable=True)
    created_at = Column(DateTime(timezone=True), server_default=func.now())

    versions = relationship("DataVersion", back_populates="project")


class DataVersion(Base):
    __tablename__ = "data_versions"

    id = Column(String(26), primary_key=True, default=generate_ulid)
    project_id = Column(String(26), ForeignKey("projects.id"))
    stage = Column(String(200), nullable=False)
    commit_id = Column(String(64), nullable=False)
    author = Column(String(50))
    commit_message = Column(Text)
    manifest_snapshot = Column(Text)
    created_at = Column(DateTime(timezone=True), server_default=func.now())

    __table_args__ = (
        UniqueConstraint('project_id', 'stage', 'commit_id', name='_project_stage_commit_uc'),
    )

    project = relationship("Project", back_populates="versions")
    files = relationship("DataFile", back_populates="version")


class DataFile(Base):
    __tablename__ = "data_files"

    id = Column(Integer, primary_key=True, index=True, autoincrement=True)
    version_id = Column(String(26), ForeignKey("data_versions.id"))
    relative_path = Column(String(255))
    storage_path = Column(String(500))
    file_size = Column(BigInteger)
    file_type = Column(String(20))
    file_sha = Column(String(64), index=True)  # Git Blob SHA for deduplication
    direction = Column(String(20), nullable=True)  # e.g., 'input' or 'output'
    label = Column(String(100), nullable=True)     # e.g., '帖子输入'
    extracted_value = Column(Text, nullable=True)  # extracted JSON value
    group_key = Column(String(255), nullable=True) # Used to group related inputs and outputs
    created_at = Column(DateTime(timezone=True), server_default=func.now())

    version = relationship("DataVersion", back_populates="files")

class DataRecord(Base):
    __tablename__ = "data_records"

    id = Column(String(26), primary_key=True, default=generate_ulid)
    project_id = Column(String(26), ForeignKey("projects.id"))
    version_id = Column(String(26), ForeignKey("data_versions.id"))
    stage = Column(String(200), index=True)
    commit_id = Column(String(64))
    commit_message = Column(Text)
    group_key = Column(String(255))
    
    inputs = Column(JSON)
    outputs = Column(JSON)
    
    content_hash = Column(String(64))
    
    author = Column(String(50))
    created_at = Column(DateTime(timezone=True), server_default=func.now())

    version = relationship("DataVersion")
    project = relationship("Project")