Track bundled vendor runtime sources
Some checks failed
CI / main (push) Has been cancelled
CI / release-e2e (push) Has been cancelled

This commit is contained in:
2026-07-07 10:05:50 +08:00
parent fbe179ab53
commit 52636c91ae
2111 changed files with 1850012 additions and 14 deletions

View File

@@ -0,0 +1,243 @@
from __future__ import annotations
import uuid
from datetime import datetime
from typing import Any
from sqlalchemy import (
JSON,
BigInteger,
Boolean,
CheckConstraint,
DateTime,
ForeignKey,
Index,
Integer,
Numeric,
String,
Text,
)
from sqlalchemy.orm import Mapped, foreign, mapped_column, relationship
from app.assets.helpers import get_utc_now
from app.database.models import Base
class Asset(Base):
__tablename__ = "assets"
id: Mapped[str] = mapped_column(
String(36), primary_key=True, default=lambda: str(uuid.uuid4())
)
hash: Mapped[str | None] = mapped_column(String(256), nullable=True)
size_bytes: Mapped[int] = mapped_column(BigInteger, nullable=False, default=0)
mime_type: Mapped[str | None] = mapped_column(String(255))
created_at: Mapped[datetime] = mapped_column(
DateTime(timezone=False), nullable=False, default=get_utc_now
)
references: Mapped[list[AssetReference]] = relationship(
"AssetReference",
back_populates="asset",
primaryjoin=lambda: Asset.id == foreign(AssetReference.asset_id),
foreign_keys=lambda: [AssetReference.asset_id],
cascade="all,delete-orphan",
passive_deletes=True,
)
# preview_id on AssetReference is a self-referential FK to asset_references.id
__table_args__ = (
Index("uq_assets_hash", "hash", unique=True),
Index("ix_assets_mime_type", "mime_type"),
CheckConstraint("size_bytes >= 0", name="ck_assets_size_nonneg"),
)
def __repr__(self) -> str:
return f"<Asset id={self.id} hash={(self.hash or '')[:12]}>"
class AssetReference(Base):
"""Unified model combining file cache state and user-facing metadata.
Each row represents either:
- A filesystem reference (file_path is set) with cache state
- An API-created reference (file_path is NULL) without cache state
"""
__tablename__ = "asset_references"
id: Mapped[str] = mapped_column(
String(36), primary_key=True, default=lambda: str(uuid.uuid4())
)
asset_id: Mapped[str] = mapped_column(
String(36), ForeignKey("assets.id", ondelete="CASCADE"), nullable=False
)
# Cache state fields (from former AssetCacheState)
file_path: Mapped[str | None] = mapped_column(Text, nullable=True)
mtime_ns: Mapped[int | None] = mapped_column(BigInteger, nullable=True)
needs_verify: Mapped[bool] = mapped_column(Boolean, nullable=False, default=False)
is_missing: Mapped[bool] = mapped_column(Boolean, nullable=False, default=False)
enrichment_level: Mapped[int] = mapped_column(Integer, nullable=False, default=0)
# Info fields (from former AssetInfo)
owner_id: Mapped[str] = mapped_column(String(128), nullable=False, default="")
name: Mapped[str] = mapped_column(String(512), nullable=False)
preview_id: Mapped[str | None] = mapped_column(
String(36), ForeignKey("asset_references.id", ondelete="SET NULL")
)
user_metadata: Mapped[dict[str, Any] | None] = mapped_column(
JSON(none_as_null=True)
)
system_metadata: Mapped[dict[str, Any] | None] = mapped_column(
JSON(none_as_null=True), nullable=True, default=None
)
job_id: Mapped[str | None] = mapped_column(String(36), nullable=True, default=None)
created_at: Mapped[datetime] = mapped_column(
DateTime(timezone=False), nullable=False, default=get_utc_now
)
updated_at: Mapped[datetime] = mapped_column(
DateTime(timezone=False), nullable=False, default=get_utc_now
)
last_access_time: Mapped[datetime] = mapped_column(
DateTime(timezone=False), nullable=False, default=get_utc_now
)
deleted_at: Mapped[datetime | None] = mapped_column(
DateTime(timezone=False), nullable=True, default=None
)
asset: Mapped[Asset] = relationship(
"Asset",
back_populates="references",
foreign_keys=[asset_id],
lazy="selectin",
)
preview_ref: Mapped[AssetReference | None] = relationship(
"AssetReference",
foreign_keys=[preview_id],
remote_side=lambda: [AssetReference.id],
)
metadata_entries: Mapped[list[AssetReferenceMeta]] = relationship(
back_populates="asset_reference",
cascade="all,delete-orphan",
passive_deletes=True,
)
tag_links: Mapped[list[AssetReferenceTag]] = relationship(
back_populates="asset_reference",
cascade="all,delete-orphan",
passive_deletes=True,
overlaps="tags,asset_references",
)
tags: Mapped[list[Tag]] = relationship(
secondary="asset_reference_tags",
back_populates="asset_references",
lazy="selectin",
viewonly=True,
overlaps="tag_links,asset_reference_links,asset_references,tag",
)
__table_args__ = (
Index("uq_asset_references_file_path", "file_path", unique=True),
Index("ix_asset_references_asset_id", "asset_id"),
Index("ix_asset_references_owner_id", "owner_id"),
Index("ix_asset_references_name", "name"),
Index("ix_asset_references_is_missing", "is_missing"),
Index("ix_asset_references_enrichment_level", "enrichment_level"),
Index("ix_asset_references_created_at", "created_at"),
Index("ix_asset_references_last_access_time", "last_access_time"),
Index("ix_asset_references_deleted_at", "deleted_at"),
Index("ix_asset_references_preview_id", "preview_id"),
Index("ix_asset_references_owner_name", "owner_id", "name"),
CheckConstraint(
"(mtime_ns IS NULL) OR (mtime_ns >= 0)", name="ck_ar_mtime_nonneg"
),
CheckConstraint(
"enrichment_level >= 0 AND enrichment_level <= 2",
name="ck_ar_enrichment_level_range",
),
)
def __repr__(self) -> str:
path_part = f" path={self.file_path!r}" if self.file_path else ""
return f"<AssetReference id={self.id} name={self.name!r}{path_part}>"
class AssetReferenceMeta(Base):
__tablename__ = "asset_reference_meta"
asset_reference_id: Mapped[str] = mapped_column(
String(36),
ForeignKey("asset_references.id", ondelete="CASCADE"),
primary_key=True,
)
key: Mapped[str] = mapped_column(String(256), primary_key=True)
ordinal: Mapped[int] = mapped_column(Integer, primary_key=True, default=0)
val_str: Mapped[str | None] = mapped_column(String(2048), nullable=True)
val_num: Mapped[float | None] = mapped_column(Numeric(38, 10), nullable=True)
val_bool: Mapped[bool | None] = mapped_column(Boolean, nullable=True)
val_json: Mapped[Any | None] = mapped_column(JSON(none_as_null=True), nullable=True)
asset_reference: Mapped[AssetReference] = relationship(
back_populates="metadata_entries"
)
__table_args__ = (
Index("ix_asset_reference_meta_key", "key"),
Index("ix_asset_reference_meta_key_val_str", "key", "val_str"),
Index("ix_asset_reference_meta_key_val_num", "key", "val_num"),
Index("ix_asset_reference_meta_key_val_bool", "key", "val_bool"),
CheckConstraint(
"val_str IS NOT NULL OR val_num IS NOT NULL OR val_bool IS NOT NULL OR val_json IS NOT NULL",
name="has_value",
),
)
class AssetReferenceTag(Base):
__tablename__ = "asset_reference_tags"
asset_reference_id: Mapped[str] = mapped_column(
String(36),
ForeignKey("asset_references.id", ondelete="CASCADE"),
primary_key=True,
)
tag_name: Mapped[str] = mapped_column(
String(512), ForeignKey("tags.name", ondelete="RESTRICT"), primary_key=True
)
origin: Mapped[str] = mapped_column(String(32), nullable=False, default="manual")
added_at: Mapped[datetime] = mapped_column(
DateTime(timezone=False), nullable=False, default=get_utc_now
)
asset_reference: Mapped[AssetReference] = relationship(back_populates="tag_links")
tag: Mapped[Tag] = relationship(back_populates="asset_reference_links")
__table_args__ = (
Index("ix_asset_reference_tags_tag_name", "tag_name"),
Index("ix_asset_reference_tags_asset_reference_id", "asset_reference_id"),
)
class Tag(Base):
__tablename__ = "tags"
name: Mapped[str] = mapped_column(String(512), primary_key=True)
asset_reference_links: Mapped[list[AssetReferenceTag]] = relationship(
back_populates="tag",
overlaps="asset_references,tags",
)
asset_references: Mapped[list[AssetReference]] = relationship(
secondary="asset_reference_tags",
back_populates="tags",
viewonly=True,
overlaps="asset_reference_links,tag_links,tags,asset_reference",
)
def __repr__(self) -> str:
return f"<Tag {self.name}>"

View File

@@ -0,0 +1,137 @@
from app.assets.database.queries.asset import (
asset_exists_by_hash,
bulk_insert_assets,
create_stub_asset,
get_asset_by_hash,
get_existing_asset_ids,
reassign_asset_references,
update_asset_hash_and_mime,
upsert_asset,
)
from app.assets.database.queries.asset_reference import (
CacheStateRow,
UnenrichedReferenceRow,
bulk_insert_references_ignore_conflicts,
bulk_update_enrichment_level,
count_active_siblings,
bulk_update_is_missing,
bulk_update_needs_verify,
convert_metadata_to_rows,
delete_assets_by_ids,
delete_orphaned_seed_asset,
delete_reference_by_id,
delete_references_by_ids,
fetch_reference_and_asset,
fetch_reference_asset_and_tags,
get_or_create_reference,
get_reference_by_file_path,
get_reference_by_id,
get_reference_with_owner_check,
get_reference_ids_by_ids,
get_references_by_paths_and_asset_ids,
get_references_for_prefixes,
get_unenriched_references,
get_unreferenced_unhashed_asset_ids,
insert_reference,
list_all_file_paths_by_asset_id,
list_references_by_asset_id,
list_references_page,
mark_references_missing_outside_prefixes,
rebuild_metadata_projection,
reference_exists,
reference_exists_for_asset_id,
restore_references_by_paths,
set_reference_metadata,
set_reference_preview,
set_reference_system_metadata,
soft_delete_reference_by_id,
update_reference_access_time,
update_reference_name,
update_is_missing_by_asset_id,
update_reference_timestamps,
update_reference_updated_at,
upsert_reference,
)
from app.assets.database.queries.tags import (
AddTagsResult,
RemoveTagsResult,
SetTagsResult,
add_missing_tag_for_asset_id,
add_tags_to_reference,
bulk_insert_tags_and_meta,
ensure_tags_exist,
get_reference_tags,
list_tag_counts_for_filtered_assets,
list_tags_with_usage,
remove_missing_tag_for_asset_id,
remove_tags_from_reference,
set_reference_tags,
validate_tags_exist,
)
__all__ = [
"AddTagsResult",
"CacheStateRow",
"RemoveTagsResult",
"SetTagsResult",
"UnenrichedReferenceRow",
"add_missing_tag_for_asset_id",
"add_tags_to_reference",
"asset_exists_by_hash",
"bulk_insert_assets",
"bulk_insert_references_ignore_conflicts",
"bulk_insert_tags_and_meta",
"bulk_update_enrichment_level",
"count_active_siblings",
"create_stub_asset",
"bulk_update_is_missing",
"bulk_update_needs_verify",
"convert_metadata_to_rows",
"delete_assets_by_ids",
"delete_orphaned_seed_asset",
"delete_reference_by_id",
"delete_references_by_ids",
"ensure_tags_exist",
"fetch_reference_and_asset",
"fetch_reference_asset_and_tags",
"get_asset_by_hash",
"get_existing_asset_ids",
"get_or_create_reference",
"get_reference_by_file_path",
"get_reference_by_id",
"get_reference_with_owner_check",
"get_reference_ids_by_ids",
"get_reference_tags",
"get_references_by_paths_and_asset_ids",
"get_references_for_prefixes",
"get_unenriched_references",
"get_unreferenced_unhashed_asset_ids",
"insert_reference",
"list_all_file_paths_by_asset_id",
"list_references_by_asset_id",
"list_references_page",
"list_tag_counts_for_filtered_assets",
"list_tags_with_usage",
"mark_references_missing_outside_prefixes",
"reassign_asset_references",
"rebuild_metadata_projection",
"reference_exists",
"reference_exists_for_asset_id",
"remove_missing_tag_for_asset_id",
"remove_tags_from_reference",
"restore_references_by_paths",
"set_reference_metadata",
"set_reference_preview",
"set_reference_system_metadata",
"soft_delete_reference_by_id",
"set_reference_tags",
"update_asset_hash_and_mime",
"update_is_missing_by_asset_id",
"update_reference_access_time",
"update_reference_name",
"update_reference_timestamps",
"update_reference_updated_at",
"upsert_asset",
"upsert_reference",
"validate_tags_exist",
]

View File

@@ -0,0 +1,152 @@
import sqlalchemy as sa
from sqlalchemy import select
from sqlalchemy.dialects import sqlite
from sqlalchemy.orm import Session
from app.assets.database.models import Asset, AssetReference
from app.assets.database.queries.common import MAX_BIND_PARAMS, calculate_rows_per_statement, iter_chunks
def asset_exists_by_hash(
session: Session,
asset_hash: str,
) -> bool:
"""
Check if an asset with a given hash exists in database.
"""
row = (
session.execute(
select(sa.literal(True))
.select_from(Asset)
.where(Asset.hash == asset_hash)
.limit(1)
)
).first()
return row is not None
def get_asset_by_hash(
session: Session,
asset_hash: str,
) -> Asset | None:
return (
(session.execute(select(Asset).where(Asset.hash == asset_hash).limit(1)))
.scalars()
.first()
)
def upsert_asset(
session: Session,
asset_hash: str,
size_bytes: int,
mime_type: str | None = None,
) -> tuple[Asset, bool, bool]:
"""Upsert an Asset by hash. Returns (asset, created, updated)."""
vals = {"hash": asset_hash, "size_bytes": int(size_bytes)}
if mime_type:
vals["mime_type"] = mime_type
ins = (
sqlite.insert(Asset)
.values(**vals)
.on_conflict_do_nothing(index_elements=[Asset.hash])
)
res = session.execute(ins)
created = int(res.rowcount or 0) > 0
asset = (
session.execute(select(Asset).where(Asset.hash == asset_hash).limit(1))
.scalars()
.first()
)
if not asset:
raise RuntimeError("Asset row not found after upsert.")
updated = False
if not created:
changed = False
if asset.size_bytes != int(size_bytes) and int(size_bytes) > 0:
asset.size_bytes = int(size_bytes)
changed = True
if mime_type and not asset.mime_type:
asset.mime_type = mime_type
changed = True
if changed:
updated = True
return asset, created, updated
def create_stub_asset(
session: Session,
size_bytes: int,
mime_type: str | None = None,
) -> Asset:
"""Create a new asset with no hash (stub for later enrichment)."""
asset = Asset(size_bytes=size_bytes, mime_type=mime_type, hash=None)
session.add(asset)
session.flush()
return asset
def bulk_insert_assets(
session: Session,
rows: list[dict],
) -> None:
"""Bulk insert Asset rows with ON CONFLICT DO NOTHING on hash."""
if not rows:
return
ins = sqlite.insert(Asset).on_conflict_do_nothing(index_elements=[Asset.hash])
for chunk in iter_chunks(rows, calculate_rows_per_statement(5)):
session.execute(ins, chunk)
def get_existing_asset_ids(
session: Session,
asset_ids: list[str],
) -> set[str]:
"""Return the subset of asset_ids that exist in the database."""
if not asset_ids:
return set()
found: set[str] = set()
for chunk in iter_chunks(asset_ids, MAX_BIND_PARAMS):
rows = session.execute(
select(Asset.id).where(Asset.id.in_(chunk))
).fetchall()
found.update(row[0] for row in rows)
return found
def update_asset_hash_and_mime(
session: Session,
asset_id: str,
asset_hash: str | None = None,
mime_type: str | None = None,
) -> bool:
"""Update asset hash and/or mime_type. Returns True if asset was found."""
asset = session.get(Asset, asset_id)
if not asset:
return False
if asset_hash is not None:
asset.hash = asset_hash
if mime_type is not None and not asset.mime_type:
asset.mime_type = mime_type
return True
def reassign_asset_references(
session: Session,
from_asset_id: str,
to_asset_id: str,
reference_id: str,
) -> None:
"""Reassign a reference from one asset to another.
Used when merging a stub asset into an existing asset with the same hash.
"""
ref = session.get(AssetReference, reference_id)
if ref and ref.asset_id == from_asset_id:
ref.asset_id = to_asset_id
session.flush()

File diff suppressed because it is too large Load Diff

View File

@@ -0,0 +1,127 @@
"""Shared utilities for database query modules."""
import os
from decimal import Decimal
from typing import Iterable, Sequence
import sqlalchemy as sa
from sqlalchemy import exists
from app.assets.database.models import AssetReference, AssetReferenceMeta, AssetReferenceTag
from app.assets.helpers import escape_sql_like_string, normalize_tags
MAX_BIND_PARAMS = 800
def calculate_rows_per_statement(cols: int) -> int:
"""Calculate how many rows can fit in one statement given column count."""
return max(1, MAX_BIND_PARAMS // max(1, cols))
def iter_chunks(seq, n: int):
"""Yield successive n-sized chunks from seq."""
for i in range(0, len(seq), n):
yield seq[i : i + n]
def iter_row_chunks(rows: list[dict], cols_per_row: int) -> Iterable[list[dict]]:
"""Yield chunks of rows sized to fit within bind param limits."""
if not rows:
return
yield from iter_chunks(rows, calculate_rows_per_statement(cols_per_row))
def build_visible_owner_clause(owner_id: str) -> sa.sql.ClauseElement:
"""Build owner visibility predicate for reads.
Owner-less rows are visible to everyone.
"""
owner_id = (owner_id or "").strip()
if owner_id == "":
return AssetReference.owner_id == ""
return AssetReference.owner_id.in_(["", owner_id])
def build_prefix_like_conditions(
prefixes: list[str],
) -> list[sa.sql.ColumnElement]:
"""Build LIKE conditions for matching file paths under directory prefixes."""
conds = []
for p in prefixes:
base = os.path.abspath(p)
if not base.endswith(os.sep):
base += os.sep
escaped, esc = escape_sql_like_string(base)
conds.append(AssetReference.file_path.like(escaped + "%", escape=esc))
return conds
def apply_tag_filters(
stmt: sa.sql.Select,
include_tags: Sequence[str] | None = None,
exclude_tags: Sequence[str] | None = None,
) -> sa.sql.Select:
"""include_tags: every tag must be present; exclude_tags: none may be present."""
include_tags = normalize_tags(include_tags)
exclude_tags = normalize_tags(exclude_tags)
if include_tags:
for tag_name in include_tags:
stmt = stmt.where(
exists().where(
(AssetReferenceTag.asset_reference_id == AssetReference.id)
& (AssetReferenceTag.tag_name == tag_name)
)
)
if exclude_tags:
stmt = stmt.where(
~exists().where(
(AssetReferenceTag.asset_reference_id == AssetReference.id)
& (AssetReferenceTag.tag_name.in_(exclude_tags))
)
)
return stmt
def apply_metadata_filter(
stmt: sa.sql.Select,
metadata_filter: dict | None = None,
) -> sa.sql.Select:
"""Apply filters using asset_reference_meta projection table."""
if not metadata_filter:
return stmt
def _exists_for_pred(key: str, *preds) -> sa.sql.ClauseElement:
return sa.exists().where(
AssetReferenceMeta.asset_reference_id == AssetReference.id,
AssetReferenceMeta.key == key,
*preds,
)
def _exists_clause_for_value(key: str, value) -> sa.sql.ClauseElement:
if value is None:
return sa.not_(
sa.exists().where(
AssetReferenceMeta.asset_reference_id == AssetReference.id,
AssetReferenceMeta.key == key,
)
)
if isinstance(value, bool):
return _exists_for_pred(key, AssetReferenceMeta.val_bool == bool(value))
if isinstance(value, (int, float, Decimal)):
num = value if isinstance(value, Decimal) else Decimal(str(value))
return _exists_for_pred(key, AssetReferenceMeta.val_num == num)
if isinstance(value, str):
return _exists_for_pred(key, AssetReferenceMeta.val_str == value)
return _exists_for_pred(key, AssetReferenceMeta.val_json == value)
for k, v in metadata_filter.items():
if isinstance(v, list):
ors = [_exists_clause_for_value(k, elem) for elem in v]
if ors:
stmt = stmt.where(sa.or_(*ors))
else:
stmt = stmt.where(_exists_clause_for_value(k, v))
return stmt

View File

@@ -0,0 +1,415 @@
from dataclasses import dataclass
from typing import Iterable, Sequence
import sqlalchemy as sa
from sqlalchemy import delete, func, select
from sqlalchemy.dialects import sqlite
from sqlalchemy.exc import IntegrityError
from sqlalchemy.orm import Session
from app.assets.database.models import (
Asset,
AssetReference,
AssetReferenceMeta,
AssetReferenceTag,
Tag,
)
from app.assets.database.queries.common import (
apply_metadata_filter,
apply_tag_filters,
build_visible_owner_clause,
iter_row_chunks,
)
from app.assets.helpers import escape_sql_like_string, get_utc_now, normalize_tags
@dataclass(frozen=True)
class AddTagsResult:
added: list[str]
already_present: list[str]
total_tags: list[str]
@dataclass(frozen=True)
class RemoveTagsResult:
removed: list[str]
not_present: list[str]
total_tags: list[str]
@dataclass(frozen=True)
class SetTagsResult:
added: list[str]
removed: list[str]
total: list[str]
def validate_tags_exist(session: Session, tags: list[str]) -> None:
"""Raise ValueError if any of the given tag names do not exist."""
existing_tag_names = set(
name
for (name,) in session.execute(select(Tag.name).where(Tag.name.in_(tags))).all()
)
missing = [t for t in tags if t not in existing_tag_names]
if missing:
raise ValueError(f"Unknown tags: {missing}")
def ensure_tags_exist(session: Session, names: Iterable[str]) -> None:
wanted = normalize_tags(list(names))
if not wanted:
return
rows = [{"name": n} for n in list(dict.fromkeys(wanted))]
ins = (
sqlite.insert(Tag)
.values(rows)
.on_conflict_do_nothing(index_elements=[Tag.name])
)
session.execute(ins)
def get_reference_tags(session: Session, reference_id: str) -> list[str]:
return [
tag_name
for (tag_name,) in (
session.execute(
select(AssetReferenceTag.tag_name)
.where(AssetReferenceTag.asset_reference_id == reference_id)
.order_by(AssetReferenceTag.tag_name.asc())
)
).all()
]
def set_reference_tags(
session: Session,
reference_id: str,
tags: Sequence[str],
origin: str = "manual",
) -> SetTagsResult:
desired = normalize_tags(tags)
current = set(get_reference_tags(session, reference_id))
to_add = [t for t in desired if t not in current]
to_remove = [t for t in current if t not in desired]
if to_add:
ensure_tags_exist(session, to_add)
session.add_all(
[
AssetReferenceTag(
asset_reference_id=reference_id,
tag_name=t,
origin=origin,
added_at=get_utc_now(),
)
for t in to_add
]
)
session.flush()
if to_remove:
session.execute(
delete(AssetReferenceTag).where(
AssetReferenceTag.asset_reference_id == reference_id,
AssetReferenceTag.tag_name.in_(to_remove),
)
)
session.flush()
return SetTagsResult(added=sorted(to_add), removed=sorted(to_remove), total=sorted(desired))
def add_tags_to_reference(
session: Session,
reference_id: str,
tags: Sequence[str],
origin: str = "manual",
create_if_missing: bool = True,
reference_row: AssetReference | None = None,
) -> AddTagsResult:
if not reference_row:
ref = session.get(AssetReference, reference_id)
if not ref:
raise ValueError(f"AssetReference {reference_id} not found")
norm = normalize_tags(tags)
if not norm:
total = get_reference_tags(session, reference_id=reference_id)
return AddTagsResult(added=[], already_present=[], total_tags=total)
if create_if_missing:
ensure_tags_exist(session, norm)
current = set(get_reference_tags(session, reference_id))
want = set(norm)
to_add = sorted(want - current)
if to_add:
with session.begin_nested() as nested:
try:
session.add_all(
[
AssetReferenceTag(
asset_reference_id=reference_id,
tag_name=t,
origin=origin,
added_at=get_utc_now(),
)
for t in to_add
]
)
session.flush()
except IntegrityError:
nested.rollback()
after = set(get_reference_tags(session, reference_id=reference_id))
return AddTagsResult(
added=sorted(((after - current) & want)),
already_present=sorted(want & current),
total_tags=sorted(after),
)
def remove_tags_from_reference(
session: Session,
reference_id: str,
tags: Sequence[str],
) -> RemoveTagsResult:
ref = session.get(AssetReference, reference_id)
if not ref:
raise ValueError(f"AssetReference {reference_id} not found")
norm = normalize_tags(tags)
if not norm:
total = get_reference_tags(session, reference_id=reference_id)
return RemoveTagsResult(removed=[], not_present=[], total_tags=total)
existing = set(get_reference_tags(session, reference_id))
to_remove = sorted(set(t for t in norm if t in existing))
not_present = sorted(set(t for t in norm if t not in existing))
if to_remove:
session.execute(
delete(AssetReferenceTag).where(
AssetReferenceTag.asset_reference_id == reference_id,
AssetReferenceTag.tag_name.in_(to_remove),
)
)
session.flush()
total = get_reference_tags(session, reference_id=reference_id)
return RemoveTagsResult(removed=to_remove, not_present=not_present, total_tags=total)
def add_missing_tag_for_asset_id(
session: Session,
asset_id: str,
origin: str = "automatic",
) -> None:
select_rows = (
sa.select(
AssetReference.id.label("asset_reference_id"),
sa.literal("missing").label("tag_name"),
sa.literal(origin).label("origin"),
sa.literal(get_utc_now()).label("added_at"),
)
.where(AssetReference.asset_id == asset_id)
.where(
sa.not_(
sa.exists().where(
(AssetReferenceTag.asset_reference_id == AssetReference.id)
& (AssetReferenceTag.tag_name == "missing")
)
)
)
)
session.execute(
sqlite.insert(AssetReferenceTag)
.from_select(
["asset_reference_id", "tag_name", "origin", "added_at"],
select_rows,
)
.on_conflict_do_nothing(
index_elements=[
AssetReferenceTag.asset_reference_id,
AssetReferenceTag.tag_name,
]
)
)
def remove_missing_tag_for_asset_id(
session: Session,
asset_id: str,
) -> None:
session.execute(
sa.delete(AssetReferenceTag).where(
AssetReferenceTag.asset_reference_id.in_(
sa.select(AssetReference.id).where(AssetReference.asset_id == asset_id)
),
AssetReferenceTag.tag_name == "missing",
)
)
def list_tags_with_usage(
session: Session,
prefix: str | None = None,
limit: int = 100,
offset: int = 0,
include_zero: bool = True,
order: str = "count_desc",
owner_id: str = "",
) -> tuple[list[tuple[str, str, int]], int]:
counts_sq = (
select(
AssetReferenceTag.tag_name.label("tag_name"),
func.count(AssetReferenceTag.asset_reference_id).label("cnt"),
)
.select_from(AssetReferenceTag)
.join(AssetReference, AssetReference.id == AssetReferenceTag.asset_reference_id)
.where(build_visible_owner_clause(owner_id))
.where(
sa.or_(
AssetReference.is_missing == False, # noqa: E712
AssetReferenceTag.tag_name == "missing",
)
)
.where(AssetReference.deleted_at.is_(None))
.group_by(AssetReferenceTag.tag_name)
.subquery()
)
q = (
select(
Tag.name,
func.coalesce(counts_sq.c.cnt, 0).label("count"),
)
.select_from(Tag)
.join(counts_sq, counts_sq.c.tag_name == Tag.name, isouter=True)
)
if prefix:
escaped, esc = escape_sql_like_string(prefix.strip().lower())
q = q.where(Tag.name.like(escaped + "%", escape=esc))
if not include_zero:
q = q.where(func.coalesce(counts_sq.c.cnt, 0) > 0)
if order == "name_asc":
q = q.order_by(Tag.name.asc())
else:
q = q.order_by(func.coalesce(counts_sq.c.cnt, 0).desc(), Tag.name.asc())
total_q = select(func.count()).select_from(Tag)
if prefix:
escaped, esc = escape_sql_like_string(prefix.strip().lower())
total_q = total_q.where(Tag.name.like(escaped + "%", escape=esc))
if not include_zero:
visible_tags_sq = (
select(AssetReferenceTag.tag_name)
.join(AssetReference, AssetReference.id == AssetReferenceTag.asset_reference_id)
.where(build_visible_owner_clause(owner_id))
.where(
sa.or_(
AssetReference.is_missing == False, # noqa: E712
AssetReferenceTag.tag_name == "missing",
)
)
.where(AssetReference.deleted_at.is_(None))
.group_by(AssetReferenceTag.tag_name)
)
total_q = total_q.where(Tag.name.in_(visible_tags_sq))
rows = (session.execute(q.limit(limit).offset(offset))).all()
total = (session.execute(total_q)).scalar_one()
rows_norm = [(name, int(count or 0)) for (name, count) in rows]
return rows_norm, int(total or 0)
def list_tag_counts_for_filtered_assets(
session: Session,
owner_id: str = "",
include_tags: Sequence[str] | None = None,
exclude_tags: Sequence[str] | None = None,
name_contains: str | None = None,
metadata_filter: dict | None = None,
limit: int = 100,
) -> dict[str, int]:
"""Return tag counts for assets matching the given filters.
Uses the same filtering logic as list_references_page but returns
{tag_name: count} instead of paginated references.
"""
# Build a subquery of matching reference IDs
ref_sq = (
select(AssetReference.id)
.join(Asset, Asset.id == AssetReference.asset_id)
.where(build_visible_owner_clause(owner_id))
.where(AssetReference.is_missing == False) # noqa: E712
.where(AssetReference.deleted_at.is_(None))
)
if name_contains:
escaped, esc = escape_sql_like_string(name_contains)
ref_sq = ref_sq.where(AssetReference.name.ilike(f"%{escaped}%", escape=esc))
ref_sq = apply_tag_filters(ref_sq, include_tags, exclude_tags)
ref_sq = apply_metadata_filter(ref_sq, metadata_filter)
ref_sq = ref_sq.subquery()
# Count tags across those references
q = (
select(
AssetReferenceTag.tag_name,
func.count(AssetReferenceTag.asset_reference_id).label("cnt"),
)
.where(AssetReferenceTag.asset_reference_id.in_(select(ref_sq.c.id)))
.group_by(AssetReferenceTag.tag_name)
.order_by(func.count(AssetReferenceTag.asset_reference_id).desc(), AssetReferenceTag.tag_name.asc())
.limit(limit)
)
rows = session.execute(q).all()
return {tag_name: int(cnt) for tag_name, cnt in rows}
def bulk_insert_tags_and_meta(
session: Session,
tag_rows: list[dict],
meta_rows: list[dict],
) -> None:
"""Batch insert into asset_reference_tags and asset_reference_meta.
Uses ON CONFLICT DO NOTHING.
Args:
session: Database session
tag_rows: Dicts with: asset_reference_id, tag_name, origin, added_at
meta_rows: Dicts with: asset_reference_id, key, ordinal, val_*
"""
if tag_rows:
ins_tags = sqlite.insert(AssetReferenceTag).on_conflict_do_nothing(
index_elements=[
AssetReferenceTag.asset_reference_id,
AssetReferenceTag.tag_name,
]
)
for chunk in iter_row_chunks(tag_rows, cols_per_row=4):
session.execute(ins_tags, chunk)
if meta_rows:
ins_meta = sqlite.insert(AssetReferenceMeta).on_conflict_do_nothing(
index_elements=[
AssetReferenceMeta.asset_reference_id,
AssetReferenceMeta.key,
AssetReferenceMeta.ordinal,
]
)
for chunk in iter_row_chunks(meta_rows, cols_per_row=7):
session.execute(ins_meta, chunk)