def __init__(self, seed=0,
which_engine_binary=hs.default_engine_binary,
hdfs_dir=hs.default_hdfs_dir,
jobtracker_uri=hs.default_jobtracker_uri,
hdfs_uri=hs.default_hdfs_uri,
which_hadoop_jar=hs.default_hadoop_jar,
which_hadoop_binary=hs.default_hadoop_binary,
output_path=hs.default_output_path,
input_filename=hs.default_input_filename,
table_data_filename=hs.default_table_data_filename,
command_dict_filename=hs.default_command_dict_filename,
one_map_task_per_line=True,
):
xu.assert_vpn_is_connected()
#
self.which_hadoop_binary = which_hadoop_binary
#
self.seed_generator = gu.int_generator(seed)
self.which_engine_binary = which_engine_binary
self.hdfs_dir = hdfs_dir
self.jobtracker_uri = jobtracker_uri
self.hdfs_uri = hdfs_uri
self.which_hadoop_jar = which_hadoop_jar
self.output_path = output_path
self.input_filename = input_filename
self.table_data_filename = table_data_filename
self.one_map_task_per_line = one_map_task_per_line
self.command_dict_filename = command_dict_filename
return
def initialize(self,
M_c,
M_r,
T,
initialization='from_the_prior',
n_chains=1):
"""Sample a latent state from prior
:param M_c: The column metadata
:type M_c: dict
:param M_r: The row metadata
:type M_r: dict
:param T: The data table in mapped representation (all floats, generated
by data_utils.read_data_objects)
:type T: list of lists
:returns: X_L, X_D -- the latent state
"""
output_path = self.output_path
input_filename = self.input_filename
table_data_filename = self.table_data_filename
intialize_args_dict_filename = self.command_dict_filename
xu.assert_vpn_is_connected()
#
table_data = dict(M_c=M_c, M_r=M_r, T=T)
initialize_args_dict = dict(command='initialize',
initialization=initialization)
xu.write_initialization_files(input_filename, table_data,
table_data_filename,
initialize_args_dict,
intialize_args_dict_filename, n_chains)
os.system('cp %s initialize_input' % input_filename)
self.send_hadoop_command(n_tasks=n_chains)
was_successful = self.get_hadoop_results()
hadoop_output = None
if was_successful:
hu.copy_hadoop_output(output_path, 'initialize_output')
X_L_list, X_D_list = hu.read_hadoop_output(output_path)
hadoop_output = X_L_list, X_D_list
return hadoop_output
def initialize(self, M_c, M_r, T, initialization='from_the_prior',
n_chains=1):
"""Sample a latent state from prior
:param M_c: The column metadata
:type M_c: dict
:param M_r: The row metadata
:type M_r: dict
:param T: The data table in mapped representation (all floats, generated
by data_utils.read_data_objects)
:type T: list of lists
:returns: X_L, X_D -- the latent state
"""
output_path = self.output_path
input_filename = self.input_filename
table_data_filename = self.table_data_filename
intialize_args_dict_filename = self.command_dict_filename
xu.assert_vpn_is_connected()
#
table_data = dict(M_c=M_c, M_r=M_r, T=T)
initialize_args_dict = dict(command='initialize',
initialization=initialization)
xu.write_initialization_files(input_filename,
table_data, table_data_filename,
initialize_args_dict,
intialize_args_dict_filename,
n_chains)
os.system('cp %s initialize_input' % input_filename)
self.send_hadoop_command(n_tasks=n_chains)
was_successful = self.get_hadoop_results()
hadoop_output = None
if was_successful:
hu.copy_hadoop_output(output_path, 'initialize_output')
X_L_list, X_D_list = hu.read_hadoop_output(output_path)
hadoop_output = X_L_list, X_D_list
return hadoop_output
def analyze(self, M_c, T, X_L, X_D, kernel_list=(), n_steps=1, c=(), r=(),
max_iterations=-1, max_time=-1, **kwargs):
"""Evolve the latent state by running MCMC transition kernels
:param M_c: The column metadata
:type M_c: dict
:param T: The data table in mapped representation (all floats, generated
by data_utils.read_data_objects)
:type T: list of lists
:param X_L: the latent variables associated with the latent state
:type X_L: dict
:param X_D: the particular cluster assignments of each row in each view
:type X_D: list of lists
:param kernel_list: names of the MCMC transition kernels to run
:type kernel_list: list of strings
:param n_steps: the number of times to run each MCMC transition kernel
:type n_steps: int
:param c: the (global) column indices to run MCMC transition kernels on
:type c: list of ints
:param r: the (global) row indices to run MCMC transition kernels on
:type r: list of ints
:param max_iterations: the maximum number of times ot run each MCMC
transition kernel. Applicable only if
max_time != -1.
:type max_iterations: int
:param max_time: the maximum amount of time (seconds) to run MCMC
transition kernels for before stopping to return
progress
:type max_time: float
:param kwargs: optional arguments to pass to hadoop_line_processor.jar.
Currently, presence of a 'chunk_size' kwarg causes
different behavior.
:returns: X_L, X_D -- the evolved latent state
"""
output_path = self.output_path
input_filename = self.input_filename
table_data_filename = self.table_data_filename
analyze_args_dict_filename = self.command_dict_filename
xu.assert_vpn_is_connected()
#
table_data = dict(M_c=M_c, T=T)
analyze_args_dict = dict(command='analyze', kernel_list=kernel_list,
n_steps=n_steps, c=c, r=r, max_time=max_time)
# chunk_analyze is a special case of analyze
if 'chunk_size' in kwargs:
chunk_size = kwargs['chunk_size']
chunk_filename_prefix = kwargs['chunk_filename_prefix']
chunk_dest_dir = kwargs['chunk_dest_dir']
analyze_args_dict['command'] = 'chunk_analyze'
analyze_args_dict['chunk_size'] = chunk_size
analyze_args_dict['chunk_filename_prefix'] = chunk_filename_prefix
# WARNING: chunk_dest_dir MUST be writeable by hadoop user mapred
analyze_args_dict['chunk_dest_dir'] = chunk_dest_dir
if not su.get_is_multistate(X_L, X_D):
X_L = [X_L]
X_D = [X_D]
#
SEEDS = kwargs.get('SEEDS', None)
xu.write_analyze_files(input_filename, X_L, X_D,
table_data, table_data_filename,
analyze_args_dict, analyze_args_dict_filename,
SEEDS)
os.system('cp %s analyze_input' % input_filename)
n_tasks = len(X_L)
self.send_hadoop_command(n_tasks)
was_successful = self.get_hadoop_results()
hadoop_output = None
if was_successful:
hu.copy_hadoop_output(output_path, 'analyze_output')
X_L_list, X_D_list = hu.read_hadoop_output(output_path)
hadoop_output = X_L_list, X_D_list
return hadoop_output
