def load_multidataset(filename, cache=False):
"""
Load a MultiDataSet from a file. First tries to load file as a
saved MultiDataSet object, then as a standard text-formatted MultiDataSet.
Parameters
----------
filename : string
The name of the file
cache : bool, optional
When set to True, a pickle file with the name filename + ".cache"
is searched for and loaded instead of filename if it exists
and is newer than filename. If no cache file exists or one
exists but it is older than filename, a cache file will be
written after loading from filename.
Returns
-------
MultiDataSet
"""
try:
# a saved MultiDataset object is ok
mds = _objs.MultiDataSet(fileToLoadFrom=filename)
except:
if cache:
bReadCache = False
cache_filename = filename + ".cache"
if _os.path.exists( cache_filename ) and \
_os.path.getmtime(filename) < _os.path.getmtime(cache_filename):
try:
print "Loading from cache file: ", cache_filename
mds = _objs.MultiDataSet(fileToLoadFrom=cache_filename)
return mds
except:
print "Failed to load from cache file"
else:
print "Cache file not found or is tool old -- one will be created after loading is completed"
# otherwise must use standard dataset file format
parser = _stdinput.StdInputParser()
mds = parser.parse_multidatafile(filename)
print "Writing cache file (to speed future loads): ", cache_filename
mds.save(cache_filename)
else:
# otherwise must use standard dataset file format
parser = _stdinput.StdInputParser()
mds = parser.parse_multidatafile(filename)
return mds
def load_gatestring_list(filename, readRawStrings=False):
"""
Load a gate string list from a file, formatted
using the standard text-format.
Parameters
----------
filename : string
The name of the file
readRawStrings : boolean
If True, gate strings are not converted
to tuples of gate labels.
Returns
-------
list of GateString objects
"""
if readRawStrings:
rawList = []
for line in open(filename, 'r'):
if len(line.strip()) == 0: continue
if len(line) == 0 or line[0] == '#': continue
rawList.append(line.strip())
return rawList
else:
std = _stdinput.StdInputParser()
return std.parse_stringfile(filename)
def write_multidataset(filename,
multidataset,
gatestring_list=None,
spamLabelOrder=None):
"""
Write a text-formatted multi-dataset file.
Parameters
----------
filename : string
The filename to write.
multidataset : MultiDataSet
The multi data set from which counts are obtained.
gatestring_list : list of GateStrings
The list of gate strings to include in the written dataset.
If None, all gate strings are output.
spamLabelOrder : list, optional
A list of the SPAM labels in multidataset which specifies
the column order in the output file.
"""
if gatestring_list is not None:
if len(gatestring_list) > 0 and not isinstance(gatestring_list[0],
_objs.GateString):
raise ValueError(
"Argument gatestring_list must be a list of GateString objects!"
)
else:
gatestring_list = multidataset.gsIndex.keys(
) #TODO: make access function for gatestrings?
spamLabels = multidataset.get_spam_labels()
if spamLabelOrder is not None:
assert (len(spamLabelOrder) == len(spamLabels))
assert (all([sl in spamLabels for sl in spamLabelOrder]))
assert (all([sl in spamLabelOrder for sl in spamLabels]))
spamLabels = spamLabelOrder
dsLabels = multidataset.keys()
headerString = '## Columns = ' + ", ".join(
["%s %s count" % (dsl, sl) for dsl in dsLabels for sl in spamLabels])
parser = _stdinput.StdInputParser()
f = open(filename, 'w')
print >> f, headerString
for gateString in gatestring_list: #gateString should be a GateString object here
gs = gateString.tup #gatestring tuple
print >> f, gateString.str + " " + " ".join(
[("%g" % multidataset[dsl][gs][sl]) for dsl in dsLabels
for sl in spamLabels])
f.close()
def write_dataset(filename,
dataset,
gatestring_list=None,
spamLabelOrder=None):
"""
Write a text-formatted dataset file.
Parameters
----------
filename : string
The filename to write.
dataset : DataSet
The data set from which counts are obtained.
gatestring_list : list of GateStrings, optional
The list of gate strings to include in the written dataset.
If None, all gate strings are output.
spamLabelOrder : list, optional
A list of the SPAM labels in dataset which specifies
the column order in the output file.
"""
if gatestring_list is not None:
if len(gatestring_list) > 0 and not isinstance(gatestring_list[0],
_objs.GateString):
raise ValueError(
"Argument gatestring_list must be a list of GateString objects!"
)
else:
gatestring_list = dataset.keys()
spamLabels = dataset.get_spam_labels()
if spamLabelOrder is not None:
assert (len(spamLabelOrder) == len(spamLabels))
assert (all([sl in spamLabels for sl in spamLabelOrder]))
assert (all([sl in spamLabelOrder for sl in spamLabels]))
spamLabels = spamLabelOrder
headerString = '## Columns = ' + ", ".join(
["%s count" % sl for sl in spamLabels])
parser = _stdinput.StdInputParser()
f = open(filename, 'w')
print >> f, headerString
for gateString in gatestring_list: #gateString should be a GateString object here
dataRow = dataset[gateString.tup]
print >> f, gateString.str + " " + " ".join([("%g" % dataRow[sl])
for sl in spamLabels])
f.close()
def load_gatestring_dict(filename):
"""
Load a gate string dictionary from a file, formatted
using the standard text-format.
Parameters
----------
filename : string
The name of the file.
Returns
-------
Dictionary with keys = gate string labels and
values = GateString objects.
"""
std = _stdinput.StdInputParser()
return std.parse_dictfile(filename)