25from .transforms
import PassThroughTransform
27__all__ = (
"BasePluginConfig",
"BasePlugin")
32 Base class measurement plugin config classes.
36 Most derived classes will want to set defaults that make sense
for the
44 Base class for measurement plugins.
46 This
is the base
class for `SingleFramePlugin`
and `ForcedPlugin`; derived
47 classes should inherit
from one of those.
51 config : `BasePluginConfig`
60 Relative execution orders are defined by a series of named constants
61 defined
in this
class: plugins
with a lower execution number are run
64 This approach was chosen instead of a full graph-based analysis of
65 dependencies because algorithm dependencies are usually both quite simple
66 and entirely substitutable: an algorithm that requires a centroid can
67 typically make use of any centroid algorithms outputs. That makes it
68 relatively easy to figure out the correct value to use
for any particular
73 """Order for algorithms which require only Footprint and Peaks (`float`).
77 Algorithms with this execution order include centroids.
81 """Order for algorithms which require a centroid (`float`).
85 These algorithms may refer assume that `getCentroid` will return a good
86 centroid,
and that a Footprint
and its Peaks are available.
90 """Order for algorithms which require a shape and a centroid (`float`).
94 These algorithms may assume that both `getCentroid` and `getShape` will
95 return good values,
and that a Footprint
and its Peaks are available.
99 """Order for algorithms which require shape, centroid and flux (`float`).
103 These algorithms may assume that `getCentroid` and `getShape` will
return
104 good values, that flux has been measured,
and that
and that a Footprint
105 and its Peaks are available.
108 DEFAULT_CATALOGCALCULATION = 4.0
109 """Order for catalog calculation plugins.
113 These plugins only operate on catalogs; they may not access pixel values.
116 ConfigClass = BasePluginConfig
117 """Plugin configuration information (`lsst.pex.config.Config`).
122 """Get the relative execution order of this plugin.
124 Must be reimplemented as a
class method by concrete derived classes.
126 raise NotImplementedError(
"All plugins must implement getExecutionOrder()")
129 object.__init__(self)
137 def fail(self, measRecord, error=None):
138 """Record a failure of the `measure` or `measureN` method.
143 Table record describing the source being measured.
144 error : `MeasurementError`, optional
145 Only provided if the measurement failed due to a
146 `MeasurementError` being raised; otherwise, will be `
None`.
150 When the plugin raises an exception, framework will call
151 `BasePlugin.fail` to allow the plugin to set its failure flag
152 field(s). When `BasePlugin.measureN` raises an exception,
153 `BasePlugin.fail` will be called repeatedly
with all the records that
156 If the exception
is an `MeasurementError`, it will be passed
as the
157 error argument;
in all other cases the error argument will be `
None`,
158 and the failure will be logged by the measurement framework
as a
162 traceback.print_exc()
163 message = ("The algorithm '%s' thinks it cannot fail, but it did; "
164 "please report this as a bug (the full traceback is above)."
165 % (self.__class__.__name__,))
166 raise NotImplementedError(message)
170 """Get the measurement transformation appropriate to this plugin.
173 may be instantiated with details of the algorithm configuration
and
174 then called
with information about calibration
and WCS to convert
from
175 raw measurement quantities to calibrated units. Calibrated data
is
176 then provided
in a separate output table.
180 By default, we copy everything
from the input to the output without
183 return PassThroughTransform
def getExecutionOrder(cls)
def __init__(self, config, name, logName=None)
def fail(self, measRecord, error=None)