24 from builtins
import object
30 """This module defines the Mapper base class."""
34 """Mapper is a base class for all mappers.
36 Subclasses may define the following methods:
38 map_{datasetType}(self, dataId, write)
39 Map a dataset id for the given dataset type into a ButlerLocation.
40 If write=True, this mapping is for an output dataset.
42 query_{datasetType}(self, key, format, dataId)
43 Return the possible values for the format fields that would produce
44 datasets at the granularity of key in combination with the provided
47 std_{datasetType}(self, item)
48 Standardize an object of the given data set type.
50 Methods that must be overridden:
53 Return a list of the keys that can be used in data ids.
61 map(self, datasetType, dataId, write=False)
63 queryMetadata(self, datasetType, key, format, dataId)
65 canStandardize(self, datasetType)
67 standardize(self, datasetType, item, dataId)
69 validate(self, dataId)
74 '''Instantiate a Mapper from a configuration.
75 In come cases the cfg may have already been instantiated into a Mapper, this is allowed and
76 the input var is simply returned.
78 :param cfg: the cfg for this mapper. It is recommended this be created by calling
80 :return: a Mapper instance
82 if isinstance(cfg, Policy):
83 return cfg[
'cls'](cfg)
87 """Create a new Mapper, saving arguments for pickling.
89 This is in __new__ instead of __init__ to save the user
90 from having to save the arguments themselves (either explicitly,
91 or by calling the super's __init__ with all their
92 *args,**kwargs. The resulting pickling system (of __new__,
93 __getstate__ and __setstate__ is similar to how __reduce__
94 is usually used, except that we save the user from any
95 responsibility (except when overriding __new__, but that
98 self = super(Mapper, cls).
__new__(cls)
106 return self._arguments
114 raise NotImplementedError(
"keys() unimplemented")
117 """Get possible values for keys given a partial data id.
119 :param datasetType: see documentation about the use of datasetType
120 :param key: this is used as the 'level' parameter
122 :param dataId: see documentation about the use of dataId
125 func = getattr(self,
'query_' + datasetType)
127 val = func(format, self.
validate(dataId))
131 """Return a list of the mappable dataset types."""
134 for attr
in dir(self):
135 if attr.startswith(
"map_"):
136 list.append(attr[4:])
139 def map(self, datasetType, dataId, write=False):
140 """Map a data id using the mapping method for its dataset type.
145 The datasetType to map
146 dataId : DataId instance
147 The dataId to use when mapping
148 write : bool, optional
149 Indicates if the map is being performed for a read operation
150 (False) or a write operation (True)
154 ButlerLocation or a list of ButlerLocation
155 The location(s) found for the map operation. If write is True, a
156 list is returned. If write is False a single ButlerLocation is
162 If no locaiton was found for this map operation, the derived mapper
163 class may raise a lsst.daf.persistence.NoResults exception. Butler
164 catches this and will look in the next Repository if there is one.
166 func = getattr(self,
'map_' + datasetType)
167 return func(self.
validate(dataId), write)
171 """Return true if this mapper can standardize an object of the given
174 return hasattr(self,
'std_' + datasetType)
177 """Standardize an object using the standardization method for its data
178 set type, if it exists."""
180 if hasattr(self,
'std_' + datasetType):
181 func = getattr(self,
'std_' + datasetType)
182 return func(item, self.
validate(dataId))
186 """Validate a dataId's contents.
188 If the dataId is valid, return it. If an invalid component can be
189 transformed into a valid one, copy the dataId, fix the component, and
190 return the copy. Otherwise, raise an exception."""
195 """Rename any existing object with the given type and dataId.
197 Not implemented in the base mapper.
199 raise NotImplementedError(
"Base-class Mapper does not implement backups")
202 """Get the registry"""