def mk_full_id_dict(_path, _file, _type):
_dict = {}
if _type == 'subscriber':
fields = SUB_FIELDS
elif _type == 'peer':
fields = PEER_FIELDS
elif _type == 'tgid':
fields = TGID_FIELDS
try:
with open(_path + _file, 'r', encoding='latin1') as _handle:
ids = csv_dict_reader(_handle,
fieldnames=fields,
restkey='OTHER',
dialect='excel',
delimiter=',')
for row in ids:
for item in row:
try:
_dict[int(row['ID'])] = dict(row)
except:
pass
_handle.close
return (_dict)
except IOError:
return _dict
def _read_file(self, key):
resp = self.boto_client.get_object(Bucket=self.bucket, Key=key)
if key.endswith('.parquet'):
body = resp['Body'].read()
reader = parquet_dict_reader(io.BytesIO(body))
yield from reader
with THREAD_LOCK:
self.bytes_processed += len(body)
self.compressed_bytes_processed += resp['ContentLength']
else:
with gz_open(resp['Body'], mode='rt') as gz_f:
reader = csv_dict_reader(gz_f, delimiter=' ')
reader.fieldnames = [
f.replace('-', '_') for f in reader.fieldnames
]
yield from reader
with THREAD_LOCK:
self.bytes_processed += gz_f.tell()
self.compressed_bytes_processed += resp['ContentLength']
p.communicate()
except OSError as ex:
print_file(COMBINED_OUTPUT_FILE_NAME)
print '#### ERROR Caught OSError `{0}`.'.format(ex)
print '&&&& FAILED {0}'.format(BENCHMARK_NAME)
exit(-1)
print_file(COMBINED_OUTPUT_FILE_NAME)
if p.returncode != 0:
print '#### ERROR Process exited with code {0}.'.format(p.returncode)
print '&&&& FAILED {0}'.format(BENCHMARK_NAME)
exit(p.returncode)
with open(COMBINED_OUTPUT_FILE_NAME) as input_file:
reader = csv_dict_reader(input_file)
variable_units = reader.next() # Get units header row.
distinguishing_variables = reader.fieldnames
measured_variables = [("STL Average Throughput", "+"),
("Thrust Average Throughput", "+")]
for record in reader:
for variable, directionality in measured_variables:
# Don't monitor regressions for STL implementations, nvbug 28980890:
if "STL" in variable:
continue
print "&&&& PERF {0}_{1}_{2}bit_{3}mib_{4} {5} {6}{7}".format(
record["Algorithm"], record["Element Type"],
def __init__(self, input_files, output_file, preserve_whitespace=True):
"""Read input files and open the output file and construct a new `io_manager`
object.
If `preserve_whitespace` is `False`, leading and trailing whitespace is
stripped from each CSV cell.
Raises
AssertionError :
If `len(input_files) <= 0` or `type(preserve_whitespace) != bool`.
"""
assert len(input_files) > 0, "No input files provided."
assert type(preserve_whitespace) == bool
self.preserve_whitespace = preserve_whitespace
self.readers = deque()
self.variable_names = None
self.variable_units = None
self.input_files = deque()
for input_file in input_files:
input_file_object = open(input_file)
reader = csv_dict_reader(filter_comments(input_file_object))
if not self.preserve_whitespace:
strip_list(reader.fieldnames)
if self.variable_names is None:
self.variable_names = reader.fieldnames
else:
# Make sure all inputs have the same schema.
assert self.variable_names == reader.fieldnames, \
"Input file (`" + input_file + "`) variable schema `" + \
str(reader.fieldnames) + "` does not match the variable schema `" + \
str(self.variable_names) + "`."
# Consume the next row, which should be the second line of the header.
variable_units = reader.next()
if not self.preserve_whitespace:
strip_dict(variable_units)
if self.variable_units is None:
self.variable_units = variable_units
else:
# Make sure all inputs have the same units schema.
assert self.variable_units == variable_units, \
"Input file (`" + input_file + "`) units schema `" + \
str(variable_units) + "` does not match the units schema `" + \
str(self.variable_units) + "`."
self.readers.append(reader)
self.input_files.append(input_file_object)
if output_file == "-": # Output to stdout.
self.output_file = stdout
else: # Output to user-specified file.
self.output_file = open(output_file, "w")
self.writer = csv_dict_writer(self.output_file,
fieldnames=self.variable_names)
def __init__(self,
baseline_input_file, observed_input_file,
output_file,
preserve_whitespace = False):
"""Read input files and open the output file and construct a new `io_manager`
object.
If `preserve_whitespace` is `False`, leading and trailing whitespace is
stripped from each CSV cell.
Raises
AssertionError :
If `type(preserve_whitespace) != bool`.
"""
assert type(preserve_whitespace) == bool
self.preserve_whitespace = preserve_whitespace
# Open baseline results.
self.baseline_input_file = open(baseline_input_file)
self.baseline_reader = csv_dict_reader(
filter_comments(self.baseline_input_file)
)
if not self.preserve_whitespace:
strip_list(self.baseline_reader.fieldnames)
self.variable_names = list(self.baseline_reader.fieldnames) # Copy.
self.variable_units = self.baseline_reader.next()
if not self.preserve_whitespace:
strip_dict(self.variable_units)
# Open observed results.
self.observed_input_file = open(observed_input_file)
self.observed_reader = csv_dict_reader(
filter_comments(self.observed_input_file)
)
if not self.preserve_whitespace:
strip_list(self.observed_reader.fieldnames)
# Make sure all inputs have the same variables schema.
assert self.variable_names == self.observed_reader.fieldnames, \
"Observed results input file (`" + observed_input_file + "`) " + \
"variable schema `" + str(self.observed_reader.fieldnames) + "` does " + \
"not match the baseline results input file (`" + baseline_input_file + \
"`) variable schema `" + str(self.variable_names) + "`."
# Consume the next row, which should be the second line of the header.
observed_variable_units = self.observed_reader.next()
if not self.preserve_whitespace:
strip_dict(observed_variable_units)
# Make sure all inputs have the same units schema.
assert self.variable_units == observed_variable_units, \
"Observed results input file (`" + observed_input_file + "`) " + \
"units schema `" + str(observed_variable_units) + "` does not " + \
"match the baseline results input file (`" + baseline_input_file + \
"`) units schema `" + str(self.variable_units) + "`."
if output_file == "-": # Output to stdout.
self.output_file = stdout
else: # Output to user-specified file.
self.output_file = open(output_file, "w")
self.writer = csv_dict_writer(
self.output_file, fieldnames = self.variable_names
)
def __init__(self, input_files, output_file, preserve_whitespace = True):
"""Read input files and open the output file and construct a new `io_manager`
object.
If `preserve_whitespace` is `False`, leading and trailing whitespace is
stripped from each CSV cell.
Raises
AssertionError :
If `len(input_files) <= 0` or `type(preserve_whitespace) != bool`.
"""
assert len(input_files) > 0, "No input files provided."
assert type(preserve_whitespace) == bool
self.preserve_whitespace = preserve_whitespace
self.readers = deque()
self.variable_names = None
self.variable_units = None
self.input_files = deque()
for input_file in input_files:
input_file_object = open(input_file)
reader = csv_dict_reader(filter_comments(input_file_object))
if not self.preserve_whitespace:
strip_list(reader.fieldnames)
if self.variable_names is None:
self.variable_names = reader.fieldnames
else:
# Make sure all inputs have the same schema.
assert self.variable_names == reader.fieldnames, \
"Input file (`" + input_file + "`) variable schema `" + \
str(reader.fieldnames) + "` does not match the variable schema `" + \
str(self.variable_names) + "`."
# Consume the next row, which should be the second line of the header.
variable_units = reader.next()
if not self.preserve_whitespace:
strip_dict(variable_units)
if self.variable_units is None:
self.variable_units = variable_units
else:
# Make sure all inputs have the same units schema.
assert self.variable_units == variable_units, \
"Input file (`" + input_file + "`) units schema `" + \
str(variable_units) + "` does not match the units schema `" + \
str(self.variable_units) + "`."
self.readers.append(reader)
self.input_files.append(input_file_object)
if output_file == "-": # Output to stdout.
self.output_file = stdout
else: # Output to user-specified file.
self.output_file = open(output_file, "w")
self.writer = csv_dict_writer(
self.output_file, fieldnames = self.variable_names
)
p.communicate()
except OSError as ex:
print_file(COMBINED_OUTPUT_FILE_NAME)
print '#### ERROR Caught OSError `{0}`.'.format(ex)
print '&&&& FAILED {0}'.format(BENCHMARK_NAME)
exit(-1)
print_file(COMBINED_OUTPUT_FILE_NAME)
if p.returncode != 0:
print '#### ERROR Process exited with code {0}.'.format(p.returncode)
print '&&&& FAILED {0}'.format(BENCHMARK_NAME)
exit(p.returncode)
with open(COMBINED_OUTPUT_FILE_NAME) as input_file:
reader = csv_dict_reader(input_file)
variable_units = reader.next() # Get units header row.
distinguishing_variables = reader.fieldnames
measured_variables = [
("STL Average Throughput", "+"),
("Thrust Average Throughput", "+")
]
for record in reader:
for variable, directionality in measured_variables:
print "&&&& PERF {0}_{1}_{2}bit_{3}mib_{4} {5} {6}{7}".format(
record["Algorithm"],
record["Element Type"],
