Coverage for tests/test_calculation.py: 19%
58 statements
« prev ^ index » next coverage.py v6.5.0, created at 2023-01-25 02:29 -0800
« prev ^ index » next coverage.py v6.5.0, created at 2023-01-25 02:29 -0800
1# This file is part of daf_relation.
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/>.
22from __future__ import annotations
24import unittest
26from lsst.daf.relation import (
27 Calculation,
28 ColumnError,
29 ColumnExpression,
30 EngineError,
31 SortTerm,
32 UnaryOperationRelation,
33 iteration,
34 tests,
35)
38class CalculationTestCase(tests.RelationTestCase):
39 """Tests for the Calculation operation and relations based on it."""
41 def setUp(self) -> None:
42 self.a = tests.ColumnTag("a")
43 self.b = tests.ColumnTag("b")
44 self.expression = ColumnExpression.function(
45 "__add__",
46 ColumnExpression.reference(self.a),
47 ColumnExpression.literal(10),
48 )
49 self.engine = iteration.Engine(name="preferred")
50 self.leaf = self.engine.make_leaf(
51 {self.a}, payload=iteration.RowSequence([{self.a: 0}, {self.a: 1}]), name="leaf"
52 )
54 def test_attributes(self) -> None:
55 """Check that all UnaryOperation and Relation attributes have the
56 expected values.
57 """
58 relation = self.leaf.with_calculated_column(self.b, self.expression)
59 assert isinstance(relation, UnaryOperationRelation)
60 self.assertEqual(relation.columns, {self.a, self.b})
61 self.assertEqual(relation.engine, self.engine)
62 self.assertEqual(relation.min_rows, self.leaf.min_rows)
63 self.assertEqual(relation.max_rows, self.leaf.max_rows)
64 operation = relation.operation
65 assert isinstance(operation, Calculation)
66 self.assertEqual(operation.expression, self.expression)
67 self.assertEqual(operation.columns_required, {self.a})
68 self.assertTrue(operation.is_empty_invariant)
69 self.assertTrue(operation.is_count_invariant)
70 self.assertFalse(operation.is_order_dependent)
71 self.assertFalse(operation.is_count_dependent)
73 def test_apply_failures(self) -> None:
74 """Test failure modes of constructing and applying Calculations."""
75 # Calculations must depend on existing columns.
76 with self.assertRaises(ColumnError):
77 self.leaf.with_calculated_column(self.b, ColumnExpression.literal(10))
78 # Required columns must be present.
79 with self.assertRaises(ColumnError):
80 self.leaf.with_calculated_column(self.b, ColumnExpression.reference(tests.ColumnTag("c")))
81 # New column must not already be present.
82 with self.assertRaises(ColumnError):
83 self.leaf.with_calculated_column(self.a, self.expression)
85 def test_backtracking_apply(self) -> None:
86 """Test apply logic that involves reordering operations in the existing
87 tree to perform the new operation in a preferred engine.
88 """
89 new_engine = iteration.Engine(name="downstream")
90 predicate = ColumnExpression.reference(self.b).lt(ColumnExpression.literal(20))
91 # Apply a bunch of operations in a new engine that a Calculation should
92 # commute with.
93 target = (
94 self.leaf.transferred_to(new_engine)
95 .with_calculated_column(self.b, self.expression)
96 .with_rows_satisfying(predicate)
97 .without_duplicates()
98 .with_only_columns({self.a})
99 .sorted([SortTerm(ColumnExpression.reference(self.a))])
100 )[:2]
101 # Apply a new Calculation with backtracking and see that it appears
102 # before the transfer to the new engine, with adjustments as needed
103 # downstream (to the Projection and Chain, in this case).
104 tag = tests.ColumnTag("tag")
105 relation = target.with_calculated_column(
106 tag, self.expression, preferred_engine=self.engine, require_preferred_engine=True
107 )
108 self.assert_relations_equal(
109 relation,
110 (
111 self.leaf.with_calculated_column(tag, self.expression)
112 .transferred_to(new_engine)
113 .with_calculated_column(self.b, self.expression)
114 .with_rows_satisfying(predicate)
115 .without_duplicates()
116 .with_only_columns({self.a, tag})
117 .sorted([SortTerm(ColumnExpression.reference(self.a))])
118 )[:2],
119 )
121 def test_no_backtracking(self) -> None:
122 """Test apply logic that handles preferred engines without reordering
123 operations in the existing tree.
124 """
125 new_engine = iteration.Engine(name="downstream")
126 # Construct a relation tree we can't reorder when inserting a
127 # Calculation, for various reasons.
128 leaf = self.engine.make_leaf(
129 {self.a},
130 payload=iteration.RowSequence([{self.a: 100}, {self.a: 200}]),
131 name="leaf",
132 )
133 # Can't insert after leaf because Materialization is locked.
134 target = leaf.transferred_to(new_engine).materialized("lock")
135 # Preferred engine is ignored if we can't backtrack and don't enable
136 # anything else.
137 self.assert_relations_equal(
138 target.with_calculated_column(self.b, self.expression, preferred_engine=self.engine),
139 target.with_calculated_column(self.b, self.expression),
140 )
141 # We can force this to be an error.
142 with self.assertRaises(EngineError):
143 target.with_calculated_column(
144 self.b, self.expression, preferred_engine=self.engine, require_preferred_engine=True
145 )
146 # We can also automatically transfer (back) to the preferred engine.
147 self.assert_relations_equal(
148 target.with_calculated_column(
149 self.b, self.expression, preferred_engine=self.engine, transfer=True
150 ),
151 target.transferred_to(self.engine).with_calculated_column(self.b, self.expression),
152 )
153 # Can't commute past an existing Calculation that provides a required
154 # column.
155 target = leaf.transferred_to(new_engine).with_calculated_column(self.b, self.expression)
156 with self.assertRaises(EngineError):
157 target.with_calculated_column(
158 tests.ColumnTag("c"),
159 ColumnExpression.reference(self.b),
160 preferred_engine=self.engine,
161 require_preferred_engine=True,
162 )
164 def test_iteration(self) -> None:
165 """Test Calculation execution in the iteration engine."""
166 relation = self.leaf.with_calculated_column(self.b, self.expression)
167 self.assertEqual(
168 list(self.engine.execute(relation)),
169 [{self.a: 0, self.b: 10}, {self.a: 1, self.b: 11}],
170 )
172 def test_str(self) -> None:
173 """Test str(Calculation) and
174 str(UnaryOperationRelation[Calculation]).
175 """
176 relation = self.leaf.with_calculated_column(self.b, self.expression)
177 self.assertEqual(str(relation), "+[b=__add__(a, 10)](leaf)")
180if __name__ == "__main__": 180 ↛ 181line 180 didn't jump to line 181, because the condition on line 180 was never true
181 unittest.main()