global refactoring; add support for Assets without the computed hash

This commit is contained in:
bigcat88
2025-09-12 18:14:52 +03:00
parent 934377ac1e
commit bb9ed04758
27 changed files with 2380 additions and 1930 deletions

View File

@@ -1,27 +1,26 @@
import uuid
from datetime import datetime
from typing import Any, Optional
import uuid
from sqlalchemy import (
Integer,
JSON,
BigInteger,
Boolean,
CheckConstraint,
DateTime,
ForeignKey,
Index,
UniqueConstraint,
JSON,
Integer,
Numeric,
String,
Text,
CheckConstraint,
Numeric,
Boolean,
UniqueConstraint,
)
from sqlalchemy.dialects.postgresql import JSONB
from sqlalchemy.orm import DeclarativeBase, Mapped, mapped_column, relationship, foreign
from sqlalchemy.orm import DeclarativeBase, Mapped, foreign, mapped_column, relationship
from .timeutil import utcnow
JSONB_V = JSON(none_as_null=True).with_variant(JSONB(none_as_null=True), 'postgresql')
@@ -46,7 +45,8 @@ def to_dict(obj: Any, include_none: bool = False) -> dict[str, Any]:
class Asset(Base):
__tablename__ = "assets"
hash: Mapped[str] = mapped_column(String(256), primary_key=True)
id: Mapped[str] = mapped_column(String(36), primary_key=True, default=lambda: str(uuid.uuid4()))
hash: Mapped[Optional[str]] = mapped_column(String(256), nullable=True)
size_bytes: Mapped[int] = mapped_column(BigInteger, nullable=False, default=0)
mime_type: Mapped[Optional[str]] = mapped_column(String(255))
created_at: Mapped[datetime] = mapped_column(
@@ -56,8 +56,8 @@ class Asset(Base):
infos: Mapped[list["AssetInfo"]] = relationship(
"AssetInfo",
back_populates="asset",
primaryjoin=lambda: Asset.hash == foreign(AssetInfo.asset_hash),
foreign_keys=lambda: [AssetInfo.asset_hash],
primaryjoin=lambda: Asset.id == foreign(AssetInfo.asset_id),
foreign_keys=lambda: [AssetInfo.asset_id],
cascade="all,delete-orphan",
passive_deletes=True,
)
@@ -65,8 +65,8 @@ class Asset(Base):
preview_of: Mapped[list["AssetInfo"]] = relationship(
"AssetInfo",
back_populates="preview_asset",
primaryjoin=lambda: Asset.hash == foreign(AssetInfo.preview_hash),
foreign_keys=lambda: [AssetInfo.preview_hash],
primaryjoin=lambda: Asset.id == foreign(AssetInfo.preview_id),
foreign_keys=lambda: [AssetInfo.preview_id],
viewonly=True,
)
@@ -76,36 +76,32 @@ class Asset(Base):
passive_deletes=True,
)
locations: Mapped[list["AssetLocation"]] = relationship(
back_populates="asset",
cascade="all, delete-orphan",
passive_deletes=True,
)
__table_args__ = (
Index("ix_assets_mime_type", "mime_type"),
CheckConstraint("size_bytes >= 0", name="ck_assets_size_nonneg"),
)
def to_dict(self, include_none: bool = False) -> dict[str, Any]:
return to_dict(self, include_none=include_none)
def __repr__(self) -> str:
return f"<Asset hash={self.hash[:12]}>"
return f"<Asset id={self.id} hash={(self.hash or '')[:12]}>"
class AssetCacheState(Base):
__tablename__ = "asset_cache_state"
id: Mapped[int] = mapped_column(Integer, primary_key=True, autoincrement=True)
asset_hash: Mapped[str] = mapped_column(String(256), ForeignKey("assets.hash", ondelete="CASCADE"), nullable=False)
asset_id: Mapped[str] = mapped_column(String(36), ForeignKey("assets.id", ondelete="CASCADE"), nullable=False)
file_path: Mapped[str] = mapped_column(Text, nullable=False)
mtime_ns: Mapped[Optional[int]] = mapped_column(BigInteger, nullable=True)
needs_verify: Mapped[bool] = mapped_column(Boolean, nullable=False, default=False)
asset: Mapped["Asset"] = relationship(back_populates="cache_states")
__table_args__ = (
Index("ix_asset_cache_state_file_path", "file_path"),
Index("ix_asset_cache_state_asset_hash", "asset_hash"),
Index("ix_asset_cache_state_asset_id", "asset_id"),
CheckConstraint("(mtime_ns IS NULL) OR (mtime_ns >= 0)", name="ck_acs_mtime_nonneg"),
UniqueConstraint("file_path", name="uq_asset_cache_state_file_path"),
)
@@ -114,27 +110,7 @@ class AssetCacheState(Base):
return to_dict(self, include_none=include_none)
def __repr__(self) -> str:
return f"<AssetCacheState id={self.id} hash={self.asset_hash[:12]} path={self.file_path!r}>"
class AssetLocation(Base):
__tablename__ = "asset_locations"
id: Mapped[int] = mapped_column(Integer, primary_key=True, autoincrement=True)
asset_hash: Mapped[str] = mapped_column(String(256), ForeignKey("assets.hash", ondelete="CASCADE"), nullable=False)
provider: Mapped[str] = mapped_column(String(32), nullable=False) # "gcs"
locator: Mapped[str] = mapped_column(Text, nullable=False) # "gs://bucket/object"
expected_size_bytes: Mapped[Optional[int]] = mapped_column(BigInteger, nullable=True)
etag: Mapped[Optional[str]] = mapped_column(String(256), nullable=True)
last_modified: Mapped[Optional[str]] = mapped_column(String(128), nullable=True)
asset: Mapped["Asset"] = relationship(back_populates="locations")
__table_args__ = (
UniqueConstraint("asset_hash", "provider", "locator", name="uq_asset_locations_triplet"),
Index("ix_asset_locations_hash", "asset_hash"),
Index("ix_asset_locations_provider", "provider"),
)
return f"<AssetCacheState id={self.id} asset_id={self.asset_id} path={self.file_path!r}>"
class AssetInfo(Base):
@@ -143,31 +119,23 @@ class AssetInfo(Base):
id: Mapped[str] = mapped_column(String(36), primary_key=True, default=lambda: str(uuid.uuid4()))
owner_id: Mapped[str] = mapped_column(String(128), nullable=False, default="")
name: Mapped[str] = mapped_column(String(512), nullable=False)
asset_hash: Mapped[str] = mapped_column(
String(256), ForeignKey("assets.hash", ondelete="RESTRICT"), nullable=False
)
preview_hash: Mapped[Optional[str]] = mapped_column(String(256), ForeignKey("assets.hash", ondelete="SET NULL"))
asset_id: Mapped[str] = mapped_column(String(36), ForeignKey("assets.id", ondelete="RESTRICT"), nullable=False)
preview_id: Mapped[Optional[str]] = mapped_column(String(36), ForeignKey("assets.id", ondelete="SET NULL"))
user_metadata: Mapped[Optional[dict[str, Any]]] = mapped_column(JSON(none_as_null=True))
created_at: Mapped[datetime] = mapped_column(
DateTime(timezone=False), nullable=False, default=utcnow
)
updated_at: Mapped[datetime] = mapped_column(
DateTime(timezone=False), nullable=False, default=utcnow
)
last_access_time: Mapped[datetime] = mapped_column(
DateTime(timezone=False), nullable=False, default=utcnow
)
created_at: Mapped[datetime] = mapped_column(DateTime(timezone=False), nullable=False, default=utcnow)
updated_at: Mapped[datetime] = mapped_column(DateTime(timezone=False), nullable=False, default=utcnow)
last_access_time: Mapped[datetime] = mapped_column(DateTime(timezone=False), nullable=False, default=utcnow)
# Relationships
asset: Mapped[Asset] = relationship(
"Asset",
back_populates="infos",
foreign_keys=[asset_hash],
foreign_keys=[asset_id],
lazy="selectin",
)
preview_asset: Mapped[Optional[Asset]] = relationship(
"Asset",
back_populates="preview_of",
foreign_keys=[preview_hash],
foreign_keys=[preview_id],
)
metadata_entries: Mapped[list["AssetInfoMeta"]] = relationship(
@@ -186,16 +154,16 @@ class AssetInfo(Base):
tags: Mapped[list["Tag"]] = relationship(
secondary="asset_info_tags",
back_populates="asset_infos",
lazy="joined",
lazy="selectin",
viewonly=True,
overlaps="tag_links,asset_info_links,asset_infos,tag",
)
__table_args__ = (
UniqueConstraint("asset_hash", "owner_id", "name", name="uq_assets_info_hash_owner_name"),
UniqueConstraint("asset_id", "owner_id", "name", name="uq_assets_info_asset_owner_name"),
Index("ix_assets_info_owner_name", "owner_id", "name"),
Index("ix_assets_info_owner_id", "owner_id"),
Index("ix_assets_info_asset_hash", "asset_hash"),
Index("ix_assets_info_asset_id", "asset_id"),
Index("ix_assets_info_name", "name"),
Index("ix_assets_info_created_at", "created_at"),
Index("ix_assets_info_last_access_time", "last_access_time"),
@@ -207,7 +175,7 @@ class AssetInfo(Base):
return data
def __repr__(self) -> str:
return f"<AssetInfo id={self.id} name={self.name!r} hash={self.asset_hash[:12]}>"
return f"<AssetInfo id={self.id} name={self.name!r} asset_id={self.asset_id}>"
class AssetInfoMeta(Base):