def parse_complexity(run_paths, complexities, run_as_key=False):
# parse the AVRs for all the runs
tables = datalib.parse_all(map(lambda x: path_complexity(x), run_paths),
complexities,
datalib.REQUIRED,
keycolname='EPOCH-START')
if run_as_key:
# modify the map to use run dir as key, not Avr file
tables = dict([(path_run_from_complexity(x[0]), x[1])
for x in tables.items()])
return tables
def parse_avrs(run_paths, classification, recent_type, complexities, run_as_key=False):
# parse the AVRs for all the runs
avrs = datalib.parse_all(
map(lambda x: path_avr(x, classification, recent_type), run_paths),
complexities,
datalib.REQUIRED,
keycolname="Timestep",
)
if run_as_key:
# modify the map to use run dir as key, not Avr file
avrs = dict([(path_run_from_avr(x[0], classification, recent_type), x[1]) for x in avrs.items()])
return avrs
def parse_complexity(run_paths, complexities, run_as_key = False):
# parse the AVRs for all the runs
tables = datalib.parse_all( map(lambda x: path_complexity( x ),
run_paths),
complexities,
datalib.REQUIRED,
keycolname = 'EPOCH-START' )
if run_as_key:
# modify the map to use run dir as key, not Avr file
tables = dict( [(path_run_from_complexity( x[0] ),
x[1])
for x in tables.items()] )
return tables
def parse_stats(run_paths, types, run_as_key=False):
# make sure the datalib files exist
for path in run_paths:
__get_stats(path)
# parse the stats for all the runs
tables = datalib.parse_all(map(lambda x: path_stats(x), run_paths),
types,
datalib.REQUIRED,
keycolname='step')
if run_as_key:
# modify the map to use run dir as key, not Avr file
tables = dict([(path_run_from_stats(x[0]), x[1])
for x in tables.items()])
return tables
def parseValues(mode, run_paths, classification, dataset, values, run_as_key):
relpath = mode.relpath(classification, dataset)
tablename = mode.defaultValues
keycolname = mode.colName_timestep
# parse the AVRs for all the runs
tables = datalib.parse_all( map(lambda x: os.path.join(x, relpath),
run_paths),
[tablename],
datalib.REQUIRED,
keycolname = keycolname )
if run_as_key:
# modify the map to use run dir as key, not Avr file
avrs = dict( [(mode.pathRunFromValue( x[0], classification, dataset ),
x[1])
for x in tables.items()] )
return avrs
def parse_avrs(run_paths,
classification,
recent_type,
complexities,
run_as_key=False):
# parse the AVRs for all the runs
avrs = datalib.parse_all(map(
lambda x: path_avr(x, classification, recent_type), run_paths),
complexities,
datalib.REQUIRED,
keycolname='Timestep')
if run_as_key:
# modify the map to use run dir as key, not Avr file
avrs = dict([(path_run_from_avr(x[0], classification,
recent_type), x[1])
for x in avrs.items()])
return avrs
def parseValues(mode, run_paths, classification, dataset, values, run_as_key):
relpath = mode.relpath(classification, dataset)
tablename = mode.defaultValues
keycolname = mode.colName_timestep
# parse the AVRs for all the runs
tables = datalib.parse_all( map(lambda x: os.path.join(x, relpath),
run_paths),
[tablename],
datalib.REQUIRED,
keycolname = keycolname )
if run_as_key:
# modify the map to use run dir as key, not Avr file
avrs = dict( [(mode.pathRunFromValue( x[0], classification, dataset ),
x[1])
for x in tables.items()] )
return avrs
def parse_stats(run_paths, types, run_as_key = False):
# make sure the datalib files exist
for path in run_paths:
__get_stats( path )
# parse the stats for all the runs
tables = datalib.parse_all( map(lambda x: path_stats( x ),
run_paths),
types,
datalib.REQUIRED,
keycolname = 'step' )
if run_as_key:
# modify the map to use run dir as key, not Avr file
tables = dict( [(path_run_from_stats( x[0] ),
x[1])
for x in tables.items()] )
return tables
