https://github.com/GPflow/GPflow
Tip revision: 766be842991222eed3185111c5a552350c0a9d89 authored by Jesper Nielsen on 26 May 2022, 16:17:32 UTC
Add small utility for reshaping / transposing.
Add small utility for reshaping / transposing.
Tip revision: 766be84
test_grouping.py
# Copyright 2022 The GPflow Contributors. All Rights Reserved.
#
# Licensed under the Apache License, Version 2.0 (the "License");
# you may not use this file except in compliance with the License.
# You may obtain a copy of the License at
#
# http://www.apache.org/licenses/LICENSE-2.0
#
# Unless required by applicable law or agreed to in writing, software
# distributed under the License is distributed on an "AS IS" BASIS,
# WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
# See the License for the specific language governing permissions and
# limitations under the License.
from typing import Collection, Sequence, Tuple
from unittest.mock import Mock
import pandas as pd
import pytest
from benchmark.grouping import GroupingKey as GK
from benchmark.grouping import GroupingSpec, _iter_by, group
from benchmark.metadata import BenchmarkMetadata
from benchmark.plotter_api import Plotter
METRICS_COLUMNS = ["run_id", "dataset", "model", "metric", "value"]
METRICS_DF = pd.DataFrame(
[
("run_1", "ds1", "md1", "mt1", 1.0),
("run_2", "ds1", "md1", "mt1", 2.0),
("run_1", "ds2", "md1", "mt1", 3.0),
("run_1", "ds1", "md2", "mt1", 4.0),
("run_1", "ds1", "md1", "mt2", 5.0),
],
columns=METRICS_COLUMNS,
)
PL1 = Mock(Plotter)
PL1.name = "pl1"
PL2 = Mock(Plotter)
PL2.name = "pl2"
BM1 = Mock(BenchmarkMetadata)
BM1.run_id = "run_1"
BM1.git_branch_name = "branch_1"
BM1.user = "user_1"
BM1.timestamp = "20220805_111111.111111"
BM2 = Mock(BenchmarkMetadata)
BM2.run_id = "run_2"
BM2.git_branch_name = "branch_2"
BM2.user = "user_2"
BM2.timestamp = "20220805_222222.222222"
def test_iter_by() -> None:
assert [
[],
[GK.USER],
[GK.RAM],
[GK.HOSTNAME],
[GK.USER, GK.RAM],
[GK.TIMESTAMP],
[GK.USER, GK.HOSTNAME],
[GK.HOSTNAME, GK.RAM],
[GK.USER, GK.TIMESTAMP],
[GK.TIMESTAMP, GK.RAM],
[GK.USER, GK.HOSTNAME, GK.RAM],
[GK.HOSTNAME, GK.TIMESTAMP],
[GK.USER, GK.TIMESTAMP, GK.RAM],
[GK.USER, GK.HOSTNAME, GK.TIMESTAMP],
[GK.HOSTNAME, GK.TIMESTAMP, GK.RAM],
[GK.USER, GK.HOSTNAME, GK.TIMESTAMP, GK.RAM],
] == list(
_iter_by(
[
GK.USER,
GK.HOSTNAME,
GK.TIMESTAMP,
GK.RAM,
]
)
)
@pytest.mark.parametrize(
"spec,expected",
[
# Group by no key:
(
GroupingSpec(by=[], minimise=False),
[
(
(),
pd.DataFrame(
[
("run_1", "ds1", "md1", "mt1", 1.0),
("run_2", "ds1", "md1", "mt1", 2.0),
("run_1", "ds2", "md1", "mt1", 3.0),
("run_1", "ds1", "md2", "mt1", 4.0),
("run_1", "ds1", "md1", "mt2", 5.0),
],
columns=METRICS_COLUMNS,
),
[PL1, PL2],
),
],
),
# Group by a metadata key:
(
GroupingSpec(by=[GK.GIT_BRANCH_NAME], minimise=False),
[
(
("branch_1",),
pd.DataFrame(
[
("run_1", "ds1", "md1", "mt1", 1.0),
("run_1", "ds2", "md1", "mt1", 3.0),
("run_1", "ds1", "md2", "mt1", 4.0),
("run_1", "ds1", "md1", "mt2", 5.0),
],
columns=METRICS_COLUMNS,
),
[PL1, PL2],
),
(
("branch_2",),
pd.DataFrame(
[
("run_2", "ds1", "md1", "mt1", 2.0),
],
columns=METRICS_COLUMNS,
),
[PL1, PL2],
),
],
),
# Group by a metrics_df key:
(
GroupingSpec(by=[GK.DATASET], minimise=False),
[
(
("ds1",),
pd.DataFrame(
[
("run_1", "ds1", "md1", "mt1", 1.0),
("run_2", "ds1", "md1", "mt1", 2.0),
("run_1", "ds1", "md2", "mt1", 4.0),
("run_1", "ds1", "md1", "mt2", 5.0),
],
columns=METRICS_COLUMNS,
),
[PL1, PL2],
),
(
("ds2",),
pd.DataFrame(
[
("run_1", "ds2", "md1", "mt1", 3.0),
],
columns=METRICS_COLUMNS,
),
[PL1, PL2],
),
],
),
# Group by a (the) plotter key:
(
GroupingSpec(by=[GK.PLOTTER], minimise=False),
[
(
(),
pd.DataFrame(
[
("run_1", "ds1", "md1", "mt1", 1.0),
("run_2", "ds1", "md1", "mt1", 2.0),
("run_1", "ds2", "md1", "mt1", 3.0),
("run_1", "ds1", "md2", "mt1", 4.0),
("run_1", "ds1", "md1", "mt2", 5.0),
],
columns=METRICS_COLUMNS,
),
[PL1],
),
(
(),
pd.DataFrame(
[
("run_1", "ds1", "md1", "mt1", 1.0),
("run_2", "ds1", "md1", "mt1", 2.0),
("run_1", "ds2", "md1", "mt1", 3.0),
("run_1", "ds1", "md2", "mt1", 4.0),
("run_1", "ds1", "md1", "mt2", 5.0),
],
columns=METRICS_COLUMNS,
),
[PL2],
),
],
),
# Group by all the keys:
(
GroupingSpec(
by=[
GK.USER,
GK.GIT_BRANCH_NAME,
GK.TIMESTAMP,
GK.DATASET,
GK.MODEL,
GK.METRIC,
GK.PLOTTER,
],
minimise=False,
),
[
(
("user_1", "branch_1", "20220805_111111.111111", "ds1", "md1", "mt1"),
pd.DataFrame(
[
("run_1", "ds1", "md1", "mt1", 1.0),
],
columns=METRICS_COLUMNS,
),
[PL1],
),
(
("user_1", "branch_1", "20220805_111111.111111", "ds1", "md1", "mt1"),
pd.DataFrame(
[
("run_1", "ds1", "md1", "mt1", 1.0),
],
columns=METRICS_COLUMNS,
),
[PL2],
),
(
("user_1", "branch_1", "20220805_111111.111111", "ds1", "md1", "mt2"),
pd.DataFrame(
[
("run_1", "ds1", "md1", "mt2", 5.0),
],
columns=METRICS_COLUMNS,
),
[PL1],
),
(
("user_1", "branch_1", "20220805_111111.111111", "ds1", "md1", "mt2"),
pd.DataFrame(
[
("run_1", "ds1", "md1", "mt2", 5.0),
],
columns=METRICS_COLUMNS,
),
[PL2],
),
(
("user_1", "branch_1", "20220805_111111.111111", "ds1", "md2", "mt1"),
pd.DataFrame(
[
("run_1", "ds1", "md2", "mt1", 4.0),
],
columns=METRICS_COLUMNS,
),
[PL1],
),
(
("user_1", "branch_1", "20220805_111111.111111", "ds1", "md2", "mt1"),
pd.DataFrame(
[
("run_1", "ds1", "md2", "mt1", 4.0),
],
columns=METRICS_COLUMNS,
),
[PL2],
),
(
("user_1", "branch_1", "20220805_111111.111111", "ds2", "md1", "mt1"),
pd.DataFrame(
[
("run_1", "ds2", "md1", "mt1", 3.0),
],
columns=METRICS_COLUMNS,
),
[PL1],
),
(
("user_1", "branch_1", "20220805_111111.111111", "ds2", "md1", "mt1"),
pd.DataFrame(
[
("run_1", "ds2", "md1", "mt1", 3.0),
],
columns=METRICS_COLUMNS,
),
[PL2],
),
(
("user_2", "branch_2", "20220805_222222.222222", "ds1", "md1", "mt1"),
pd.DataFrame(
[
("run_2", "ds1", "md1", "mt1", 2.0),
],
columns=METRICS_COLUMNS,
),
[PL1],
),
(
("user_2", "branch_2", "20220805_222222.222222", "ds1", "md1", "mt1"),
pd.DataFrame(
[
("run_2", "ds1", "md1", "mt1", 2.0),
],
columns=METRICS_COLUMNS,
),
[PL2],
),
],
),
# Group by all the keys, and minimise:
(
GroupingSpec(
by=[
GK.USER,
GK.GIT_BRANCH_NAME,
GK.TIMESTAMP,
GK.DATASET,
GK.MODEL,
GK.METRIC,
GK.PLOTTER,
],
minimise=True,
),
[
(
("branch_1", "ds1", "md1", "mt1"),
pd.DataFrame(
[
("run_1", "ds1", "md1", "mt1", 1.0),
],
columns=METRICS_COLUMNS,
),
[PL1],
),
(
("branch_1", "ds1", "md1", "mt1"),
pd.DataFrame(
[
("run_1", "ds1", "md1", "mt1", 1.0),
],
columns=METRICS_COLUMNS,
),
[PL2],
),
(
("branch_1", "ds1", "md1", "mt2"),
pd.DataFrame(
[
("run_1", "ds1", "md1", "mt2", 5.0),
],
columns=METRICS_COLUMNS,
),
[PL1],
),
(
("branch_1", "ds1", "md1", "mt2"),
pd.DataFrame(
[
("run_1", "ds1", "md1", "mt2", 5.0),
],
columns=METRICS_COLUMNS,
),
[PL2],
),
(
("branch_1", "ds1", "md2", "mt1"),
pd.DataFrame(
[
("run_1", "ds1", "md2", "mt1", 4.0),
],
columns=METRICS_COLUMNS,
),
[PL1],
),
(
("branch_1", "ds1", "md2", "mt1"),
pd.DataFrame(
[
("run_1", "ds1", "md2", "mt1", 4.0),
],
columns=METRICS_COLUMNS,
),
[PL2],
),
(
("branch_1", "ds2", "md1", "mt1"),
pd.DataFrame(
[
("run_1", "ds2", "md1", "mt1", 3.0),
],
columns=METRICS_COLUMNS,
),
[PL1],
),
(
("branch_1", "ds2", "md1", "mt1"),
pd.DataFrame(
[
("run_1", "ds2", "md1", "mt1", 3.0),
],
columns=METRICS_COLUMNS,
),
[PL2],
),
(
("branch_2", "ds1", "md1", "mt1"),
pd.DataFrame(
[
("run_2", "ds1", "md1", "mt1", 2.0),
],
columns=METRICS_COLUMNS,
),
[PL1],
),
(
("branch_2", "ds1", "md1", "mt1"),
pd.DataFrame(
[
("run_2", "ds1", "md1", "mt1", 2.0),
],
columns=METRICS_COLUMNS,
),
[PL2],
),
],
),
],
)
def test_group(
spec: GroupingSpec,
expected: Sequence[Tuple[Tuple[str, ...], pd.DataFrame, Collection[Plotter]]],
) -> None:
actual = group(METRICS_DF, [PL1, PL2], [BM1, BM2], spec)
assert len(expected) == len(actual)
for expected_gr, actual_gr in zip(expected, actual):
expected_key, expected_df, expected_plotters = expected_gr
actual_key, actual_df, actual_plotters = actual_gr
assert expected_key == actual_key
pd.testing.assert_frame_equal(expected_df, actual_df.reset_index(drop=True))
assert expected_plotters == list(actual_plotters)
