def exclude( self ):
'''exclude data paths.
Only those data paths not matching the exclude term are accepted.
'''
if not self.exclude_paths: return
data_paths = DataTree.getPaths( self.data )
# currently enumerates - bfs more efficient
all_paths = list(itertools.product( *data_paths ))
for path in all_paths:
for s in self.exclude_paths:
if s in path:
self.debug( "%s: ignoring path %s because of :exclude:=%s" % (self.tracker, path, s))
try: DataTree.removeLeaf( self.data, path )
except KeyError: pass
elif s.startswith("r(") and s.endswith(")"):
# collect pattern matches:
# remove r()
s = s[2:-1]
# remove flanking quotation marks
if s[0] in ('"', "'") and s[-1] in ('"', "'"): s = s[1:-1]
rx = re.compile( s )
if any( ( rx.search( p ) for p in path ) ):
self.debug( "%s: ignoring path %s because of :exclude:=%s" % (self.tracker, path, s))
try: DataTree.removeLeaf( self.data, path )
except KeyError: pass
def __call__(self, data ):
if self.nlevels == None: raise NotImplementedError("incomplete implementation of %s" % str(self))
labels = DataTree.getPaths( data )
debug( "transform: started with paths: %s" % labels)
assert len(labels) >= self.nlevels, "expected at least %i levels - got %i" % (self.nlevels, len(labels))
paths = list(itertools.product( *labels[:-self.nlevels] ))
for path in paths:
work = DataTree.getLeaf( data, path )
if not work: continue
new_data = self.transform( work, path )
if new_data:
if path:
DataTree.setLeaf( data, path, new_data )
else:
# set new root
data = new_data
else:
warn( "no data at %s - removing branch" % str(path))
DataTree.removeLeaf( data, path )
debug( "transform: finished with paths: %s" % DataTree.getPaths( data ))
return data