Coverage for python/lsst/daf/butler/registry/bridge/monolithic.py : 82%

Hot-keys on this page
r m x p toggle line displays
j k next/prev highlighted chunk
0 (zero) top of page
1 (one) first highlighted chunk
1# This file is part of daf_butler.
2#
3# Developed for the LSST Data Management System.
4# This product includes software developed by the LSST Project
5# (http://www.lsst.org).
6# See the COPYRIGHT file at the top-level directory of this distribution
7# for details of code ownership.
8#
9# This program is free software: you can redistribute it and/or modify
10# it under the terms of the GNU General Public License as published by
11# the Free Software Foundation, either version 3 of the License, or
12# (at your option) any later version.
13#
14# This program is distributed in the hope that it will be useful,
15# but WITHOUT ANY WARRANTY; without even the implied warranty of
16# MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
17# GNU General Public License for more details.
18#
19# You should have received a copy of the GNU General Public License
20# along with this program. If not, see <http://www.gnu.org/licenses/>.
21from __future__ import annotations
23__all__ = ("MonolithicDatastoreRegistryBridgeManager", "MonolithicDatastoreRegistryBridge")
25from collections import namedtuple
26from contextlib import contextmanager
27import copy
28from typing import cast, Dict, Iterable, Iterator, List, Type, TYPE_CHECKING
30import sqlalchemy
32from lsst.daf.butler import DatasetRef, ddl
33from lsst.daf.butler.core.utils import NamedValueSet
34from lsst.daf.butler.registry.interfaces import (
35 DatasetIdRef,
36 DatastoreRegistryBridge,
37 DatastoreRegistryBridgeManager,
38 FakeDatasetRef,
39)
40from lsst.daf.butler.registry.bridge.ephemeral import EphemeralDatastoreRegistryBridge
42if TYPE_CHECKING: 42 ↛ 43line 42 didn't jump to line 43, because the condition on line 42 was never true
43 from lsst.daf.butler import DimensionUniverse
44 from lsst.daf.butler.registry.interfaces import (
45 Database,
46 DatasetRecordStorageManager,
47 OpaqueTableStorageManager,
48 StaticTablesContext,
49 )
52_TablesTuple = namedtuple(
53 "_TablesTuple",
54 [
55 "dataset_location",
56 "dataset_location_trash",
57 ]
58)
61def _makeTableSpecs(datasets: Type[DatasetRecordStorageManager]) -> _TablesTuple:
62 """Construct specifications for tables used by the monolithic datastore
63 bridge classes.
65 Parameters
66 ----------
67 universe : `DimensionUniverse`
68 All dimensions known to the `Registry`.
69 datasets : subclass of `DatasetRecordStorageManager`
70 Manager class for datasets; used only to create foreign key fields.
72 Returns
73 -------
74 specs : `_TablesTuple`
75 A named tuple containing `ddl.TableSpec` instances.
76 """
77 # We want the dataset_location and dataset_location_trash tables
78 # to have the same definition, aside from the behavior of their link
79 # to the dataset table: the trash table has no foreign key constraint.
80 dataset_location_spec = ddl.TableSpec(
81 doc=(
82 "A table that provides information on whether a dataset is stored in "
83 "one or more Datastores. The presence or absence of a record in this "
84 "table itself indicates whether the dataset is present in that "
85 "Datastore. "
86 ),
87 fields=NamedValueSet([
88 ddl.FieldSpec(
89 name="datastore_name",
90 dtype=sqlalchemy.String,
91 length=256,
92 primaryKey=True,
93 nullable=False,
94 doc="Name of the Datastore this entry corresponds to.",
95 ),
96 ]),
97 )
98 dataset_location = copy.deepcopy(dataset_location_spec)
99 datasets.addDatasetForeignKey(dataset_location, primaryKey=True)
100 dataset_location_trash = copy.deepcopy(dataset_location_spec)
101 datasets.addDatasetForeignKey(dataset_location_trash, primaryKey=True, constraint=False)
102 return _TablesTuple(
103 dataset_location=dataset_location,
104 dataset_location_trash=dataset_location_trash,
105 )
108class MonolithicDatastoreRegistryBridge(DatastoreRegistryBridge):
109 """An implementation of `DatastoreRegistryBridge` that uses the same two
110 tables for all non-ephemeral datastores.
112 Parameters
113 ----------
114 datastoreName : `str`
115 Name of the `Datastore` as it should appear in `Registry` tables
116 referencing it.
117 db : `Database`
118 Object providing a database connection and generic distractions.
119 tables : `_TablesTuple`
120 Named tuple containing `sqlalchemy.schema.Table` instances.
121 """
122 def __init__(self, datastoreName: str, *, db: Database, tables: _TablesTuple):
123 super().__init__(datastoreName)
124 self._db = db
125 self._tables = tables
127 def _refsToRows(self, refs: Iterable[DatasetIdRef]) -> List[dict]:
128 """Transform an iterable of `DatasetRef` or `FakeDatasetRef` objects to
129 a list of dictionaries that match the schema of the tables used by this
130 class.
132 Parameters
133 ----------
134 refs : `Iterable` [ `DatasetRef` or `FakeDatasetRef` ]
135 Datasets to transform.
137 Returns
138 -------
139 rows : `list` [ `dict` ]
140 List of dictionaries, with "datastoreName" and "dataset_id" keys.
141 """
142 return [{"datastore_name": self.datastoreName, "dataset_id": ref.getCheckedId()} for ref in refs]
144 def insert(self, refs: Iterable[DatasetIdRef]) -> None:
145 # Docstring inherited from DatastoreRegistryBridge
146 self._db.insert(self._tables.dataset_location, *self._refsToRows(refs))
148 def moveToTrash(self, refs: Iterable[DatasetIdRef]) -> None:
149 # Docstring inherited from DatastoreRegistryBridge
150 # TODO: avoid self.check() call via queries like
151 # INSERT INTO dataset_location_trash
152 # SELECT datastore_name, dataset_id FROM dataset_location
153 # WHERE datastore_name=? AND dataset_id IN (?);
154 # DELETE FROM dataset_location
155 # WHERE datastore_name=? AND dataset_id IN (?);
156 # ...but the Database interface doesn't support those kinds of queries
157 # right now.
158 rows = self._refsToRows(self.check(refs))
159 with self._db.transaction():
160 self._db.delete(self._tables.dataset_location, ["datastore_name", "dataset_id"], *rows)
161 self._db.insert(self._tables.dataset_location_trash, *rows)
163 def check(self, refs: Iterable[DatasetIdRef]) -> Iterable[DatasetIdRef]:
164 # Docstring inherited from DatastoreRegistryBridge
165 byId = {ref.getCheckedId(): ref for ref in refs}
166 sql = sqlalchemy.sql.select(
167 [self._tables.dataset_location.columns.dataset_id]
168 ).select_from(
169 self._tables.dataset_location
170 ).where(
171 sqlalchemy.sql.and_(
172 self._tables.dataset_location.columns.datastore_name == self.datastoreName,
173 self._tables.dataset_location.columns.dataset_id.in_(byId.keys())
174 )
175 )
176 for row in self._db.query(sql).fetchall():
177 yield byId[row["dataset_id"]]
179 @contextmanager
180 def emptyTrash(self) -> Iterator[Iterable[DatasetIdRef]]:
181 # Docstring inherited from DatastoreRegistryBridge
182 sql = sqlalchemy.sql.select(
183 [self._tables.dataset_location_trash.columns.dataset_id]
184 ).select_from(
185 self._tables.dataset_location_trash
186 ).where(
187 self._tables.dataset_location_trash.columns.datastore_name == self.datastoreName
188 )
189 # Run query, transform results into a list of dicts that we can later
190 # use to delete.
191 rows = [{"dataset_id": row["dataset_id"], "datastore_name": self.datastoreName}
192 for row in self._db.query(sql).fetchall()]
193 # Start contextmanager, returning generator expression to iterate over.
194 yield (FakeDatasetRef(row["dataset_id"]) for row in rows)
195 # No exception raised in context manager block. Delete those rows
196 # from the trash table.
197 self._db.delete(self._tables.dataset_location_trash, ["dataset_id", "datastore_name"], *rows)
200class MonolithicDatastoreRegistryBridgeManager(DatastoreRegistryBridgeManager):
201 """An implementation of `DatastoreRegistryBridgeManager` that uses the same
202 two tables for all non-ephemeral datastores.
204 Parameters
205 ----------
206 db : `Database`
207 Object providing a database connection and generic distractions.
208 tables : `_TablesTuple`
209 Named tuple containing `sqlalchemy.schema.Table` instances.
210 opaque : `OpaqueTableStorageManager`
211 Manager object for opaque table storage in the `Registry`.
212 universe : `DimensionUniverse`
213 All dimensions know to the `Registry`.
214 """
215 def __init__(self, *, db: Database, tables: _TablesTuple,
216 opaque: OpaqueTableStorageManager, universe: DimensionUniverse):
217 super().__init__(opaque=opaque, universe=universe)
218 self._db = db
219 self._tables = tables
220 self._ephemeral: Dict[str, EphemeralDatastoreRegistryBridge] = {}
222 @classmethod
223 def initialize(cls, db: Database, context: StaticTablesContext, *,
224 opaque: OpaqueTableStorageManager,
225 datasets: Type[DatasetRecordStorageManager],
226 universe: DimensionUniverse,
227 ) -> DatastoreRegistryBridgeManager:
228 # Docstring inherited from DatastoreRegistryBridge
229 tables = context.addTableTuple(_makeTableSpecs(datasets))
230 return cls(db=db, tables=cast(_TablesTuple, tables), opaque=opaque, universe=universe)
232 def refresh(self) -> None:
233 # Docstring inherited from DatastoreRegistryBridge
234 # This implementation has no in-Python state that depends on which
235 # datastores exist, so there's nothing to do.
236 pass
238 def register(self, name: str, *, ephemeral: bool = False) -> DatastoreRegistryBridge:
239 # Docstring inherited from DatastoreRegistryBridge
240 if ephemeral:
241 return self._ephemeral.setdefault(name, EphemeralDatastoreRegistryBridge(name))
242 return MonolithicDatastoreRegistryBridge(name, db=self._db, tables=self._tables)
244 def findDatastores(self, ref: DatasetRef) -> Iterable[str]:
245 # Docstring inherited from DatastoreRegistryBridge
246 sql = sqlalchemy.sql.select(
247 [self._tables.dataset_location.columns.datastore_name]
248 ).select_from(
249 self._tables.dataset_location
250 ).where(
251 self._tables.dataset_location.columns.dataset_id == ref.getCheckedId()
252 )
253 for row in self._db.query(sql).fetchall():
254 yield row[self._tables.dataset_location.columns.datastore_name]
255 for name, bridge in self._ephemeral.items():
256 if ref in bridge:
257 yield name