Coverage for python / lsst / analysis / tools / tasks / makeMetricTable.py: 17%
94 statements
« prev ^ index » next coverage.py v7.13.5, created at 2026-05-06 09:07 +0000
1# This file is part of analysis_tools.
2#
3# Developed for the LSST Data Management System.
4# This product includes software developed by the LSST Project
5# (https://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 <https://www.gnu.org/licenses/>.
21from __future__ import annotations
23__all__ = (
24 "MakeMetricTableConfig",
25 "MakeMetricTableTask",
26)
28import numpy as np
29from astropy import units as u
30from astropy.table import Table
32import lsst.pipe.base as pipeBase
33from lsst.pex.config import ListField
34from lsst.pipe.base import connectionTypes as ct
35from lsst.skymap import BaseSkyMap
37from ..utils import getTractCorners
40class MakeMetricTableConnections(
41 pipeBase.PipelineTaskConnections,
42 dimensions=(),
43 defaultTemplates={"metricBundleName": "", "outputTableName": ""},
44):
45 data = ct.Input(
46 doc="Metric bundle to read from the butler",
47 name="{metricBundleName}",
48 storageClass="MetricMeasurementBundle",
49 deferLoad=True,
50 dimensions=(),
51 multiple=True,
52 )
54 skymap = ct.Input(
55 doc="The skymap that covers the tract that the data is from.",
56 name=BaseSkyMap.SKYMAP_DATASET_TYPE_NAME,
57 storageClass="SkyMap",
58 dimensions=("skymap",),
59 )
61 metricTable = ct.Output(
62 doc="Table containing metrics, one row per input metric bundle",
63 name="{outputTableName}",
64 storageClass="ArrowAstropy",
65 dimensions=(),
66 )
68 def __init__(self, *, config=None):
69 super().__init__(config=config)
70 self.data = ct.Input(
71 doc=self.data.doc,
72 name=self.data.name,
73 storageClass=self.data.storageClass,
74 deferLoad=self.data.deferLoad,
75 dimensions=frozenset(sorted(config.inputDataDimensions)),
76 multiple=self.data.multiple,
77 )
78 self.metricTable = ct.Output(
79 doc=self.metricTable.doc,
80 name=self.metricTable.name,
81 storageClass=self.metricTable.storageClass,
82 dimensions=frozenset(sorted(config.outputTableDimensions)),
83 )
85 assert config is not None, "Missing required config object."
87 if "tract" not in config.inputDataDimensions:
88 del self.skymap
90 self.dimensions.update(frozenset(sorted(config.outputTableDimensions)))
93class MakeMetricTableConfig(
94 pipeBase.PipelineTaskConfig,
95 pipelineConnections=MakeMetricTableConnections,
96):
97 inputDataDimensions = ListField[str](
98 doc="Dimensions of the input data.",
99 default=("skymap", "tract"),
100 optional=False,
101 )
102 outputTableDimensions = ListField[str](
103 doc="Dimensions of the output data.",
104 default=("skymap",),
105 optional=False,
106 )
107 dataIdFieldsToIncludeAsColumns = ListField[str](
108 doc="DataId fields to include as columns in the table. "
109 "These are added in addition to the Metric names. "
110 "At least one field must be specified.",
111 default=("tract",),
112 optional=False,
113 )
116class MakeMetricTableTask(pipeBase.PipelineTask):
117 """Turn metric bundles and combine them into a metric table."""
119 ConfigClass = MakeMetricTableConfig
120 _DefaultName = "makeMetricTable"
122 def runQuantum(self, butlerQC, inputRefs, outputRefs):
123 """Take a set of metric bundles, seperate each into its different
124 metrics, then put the values into a table with the metric names as
125 column headers.
127 Parameters
128 ----------
129 butlerQC : `lsst.pipe.base.QuantumContext`
130 inputRefs : `lsst.pipe.base.InputQuantizedConnection`
131 outputRefs : `lsst.pipe.base.OutputQuantizedConnection`
132 """
134 inputs = butlerQC.get(inputRefs)
135 if "skymap" in inputs:
136 skymap = inputs["skymap"]
137 else:
138 skymap = None
140 # Extract the info from the dataIds that is needed
141 # to populate the requested columns.
142 fields = self.config.dataIdFieldsToIncludeAsColumns
143 dataIdInfo = []
144 for data in inputRefs.data:
145 dataIdInfo.append({field: data.dataId[field] for field in fields})
147 metricBundles = []
148 for inputHandle in inputs["data"]:
149 metricBundles.append(inputHandle.get())
151 outputs = self.run(dataIdInfo, metricBundles, skymap)
152 butlerQC.put(outputs, outputRefs)
154 def run(self, dataIdInfo, metricBundles, skymap):
155 """Take the metric bundles and expand them out, then make a table of
156 the information. Add tract corner information if the bundles are
157 tract-level.
159 Parameters
160 ----------
161 dataIdInfo : `list`
162 A list of dicts that hold information extracted from the metric
163 bundle dataIds.
164 metricBundles : `list` of
165 `lsst.analysis.tools.interfaces._metricMeasurementBundle.MetricMeasurementBundle`
166 skymap : `lsst.skymap`
168 Returns
169 -------
170 metricTableStruct : `pipe.base.Struct` containing `astropy.table.Table`
171 """
173 if len(dataIdInfo) == 0:
174 raise pipeBase.NoWorkFound("dataIdInfo list is empty")
175 if len(metricBundles) == 0:
176 raise pipeBase.NoWorkFound("metricBundles list is empty")
178 metricsDict = {}
179 metricUnits = {}
181 # Add requested info from the first dataId to the metrics dict.
182 for key, value in dataIdInfo[0].items():
183 metricsDict[key] = [value]
185 # Add tract corners if inputs are at the tract-level
186 if "tract" in self.config.inputDataDimensions:
187 corners = getTractCorners(skymap, dataIdInfo[0]["tract"])
188 metricsDict["corners"] = [corners]
190 # Add the metrics and units from the first bundle to the dicts
191 for name, metrics in metricBundles[0].items():
192 percentUnitUsed = False
193 for metric in metrics:
194 fullName = f"{name}_{metric.metric_name}"
195 metricsDict[fullName] = [metric.quantity.value]
196 # "Dimensionless" and percent units not allowed in Tables:
197 if metric.quantity.unit is u.dimensionless_unscaled:
198 continue
199 elif metric.quantity.unit is u.pct:
200 percentUnitUsed = True
201 self.log.debug(
202 "Unable to propagate astropy percent unit for metric %s "
203 "in bundle %s to astropy table.",
204 metric.metric_name,
205 name,
206 )
207 continue
208 else:
209 metricUnits[fullName] = metric.quantity.unit
210 if percentUnitUsed:
211 self.log.warn(
212 "One or more metrics in the %s metric bundle uses the percent unit, "
213 "which is not supported in astropy Tables. "
214 "The value(s) have been propagated without the percent unit. "
215 "Use --log-level debug to list all affected metrics.",
216 name,
217 )
219 # Check if any additional columns are needed; add to dict if needed.
220 for i, metricBundle in enumerate(metricBundles[1:]):
221 metricRecord = []
222 for key, value in dataIdInfo[i + 1].items():
223 metricsDict[key].append(value)
224 metricRecord.append(key)
226 if "tract" in self.config.inputDataDimensions:
227 corners = getTractCorners(skymap, dataIdInfo[i + 1]["tract"])
228 metricsDict["corners"].append(corners)
229 metricRecord.append("corners")
231 for name, metrics in metricBundle.items():
232 for metric in metrics:
233 fullName = f"{name}_{metric.metric_name}"
234 # Check if the metric already exists in the output
235 if fullName in metricsDict.keys():
236 metricsDict[fullName].append(metric.quantity.value)
237 else:
238 values = [np.nan] * (len(metricsDict[metricRecord[0]]) - 1)
239 values.append(metric.quantity.value)
240 if (
241 metric.quantity.unit is not u.dimensionless_unscaled
242 and metric.quantity.unit is not u.pct
243 ):
244 metricUnits[fullName] = metric.quantity.unit
245 metricsDict[fullName] = values
246 metricRecord.append(fullName)
248 # If a metric that existed in a previous bundle does
249 # not exist for this one then add a nan
250 for metricName in metricsDict.keys():
251 if metricName not in metricRecord:
252 metricsDict[metricName].append(np.nan)
254 metricTableStruct = pipeBase.Struct(metricTable=Table(metricsDict, units=metricUnits))
255 return metricTableStruct