def process_galaxy(connection, galaxies, verbosity):
"""
Process the galaxies
"""
# Build the tables
build_dynamic_tables(connection)
for galaxy in galaxies:
galaxy_id = galaxy[GALAXY.c.galaxy_id]
steps_done = get_step_done_ids(connection, galaxy_id, STEP_DONE_ID_MEAN_STANDARD_DEVIATION, True)
if verbosity:
LOG.info(
'galaxy_id: {0}, galaxy_name: {1}'.format(
galaxy_id,
galaxy[GALAXY.c.name]
)
)
tables_required = get_tables_required(steps_done)
if len(tables_required) > 0:
for table in tables_required:
with connection.begin():
create_mean_standard_deviation = CreateMeanStandardDeviation(connection, table, galaxy_id)
create_mean_standard_deviation.calculate_details()
create_mean_standard_deviation.update_details()
add_step_done_id(
connection,
galaxy_id,
STEP_DONE_ID_MEAN_STANDARD_DEVIATION,
table.name
)
def process_galaxy(connection, galaxies, verbosity):
"""
Process the galaxies
"""
# Build the tables
build_dynamic_tables(connection)
for galaxy in galaxies:
galaxy_id = galaxy[GALAXY.c.galaxy_id]
steps_done = get_step_done_ids(connection, galaxy_id, STEP_DONE_ID_BUILD_MASK)
if verbosity:
LOG.info(
'galaxy_id: {0}, galaxy_name: {1}'.format(
galaxy_id,
galaxy[GALAXY.c.name]
)
)
if STEP_DONE_ID_BUILD_MASK not in steps_done:
with connection.begin():
build_mask = ProcessGalaxyMasks(connection, galaxy_id)
build_mask.run()
add_step_done_id(
connection,
galaxy_id,
STEP_DONE_ID_BUILD_MASK
)
def add_original_data(connection, galaxy_id, h5_file, data_required):
galaxy_group = h5_file['galaxy']
pixel_group = galaxy_group['pixel']
count = 0
total_time = 0
# Add the original values
for filter_name in data_required:
with connection.begin():
time_for_filter = insert_original_data(
connection,
filter_name,
pixel_group,
galaxy_group.attrs['output_format'],
galaxy_group.attrs['dimension_x'],
galaxy_group.attrs['dimension_y'],
galaxy_id
)
add_step_done_id(connection, galaxy_id, STEP_DONE_ID_ORIGINAL_VALUES, filter_name)
if time_for_filter > 0:
count += 1
total_time += time_for_filter
LOG.info(
'add_original_data, galaxy_id: {0}, total_time: {1:.3f}s, average_time: {2:.3f}s'.format(
galaxy_id,
total_time,
total_time / count if count > 0 else 0
)
)
def add_sed_data(connection, galaxy_id, h5_file, data_required):
galaxy_group = h5_file['galaxy']
pixel_group = galaxy_group['pixel']
count = 0
total_time = 0
# For each pair insert the data
for pair in data_required:
with connection.begin():
time_for_pair = insert_sed_data(
connection,
pair,
pixel_group,
galaxy_group.attrs['output_format'],
galaxy_group.attrs['dimension_x'],
galaxy_group.attrs['dimension_y'],
galaxy_id
)
add_step_done_id(connection, galaxy_id, STEP_DONE_ID_SED_DATA, pair[0], pair[1])
count += 1
total_time += time_for_pair
LOG.info(
'add_sed_data, galaxy_id: {0}, total_time: {1:.3f}s, average_time: {2:.3f}s'.format(
galaxy_id,
total_time,
total_time / count if count > 0 else 0
)
)
def process_galaxy(connection, galaxies, verbosity):
"""
Process the galaxies
"""
# Build the tables
build_dynamic_tables(connection)
for galaxy in galaxies:
galaxy_id = galaxy[GALAXY.c.galaxy_id]
steps_done = get_step_done_ids(
connection,
galaxy_id,
[STEP_DONE_ID_BUILD_MASK, STEP_DONE_ID_CENTROID_CENTRE_OF_MASS])
# have we built the mask
if STEP_DONE_ID_BUILD_MASK in steps_done:
if STEP_DONE_ID_CENTROID_CENTRE_OF_MASS not in steps_done:
with connection.begin():
centroid_center_of_mass = CentroidCentreOfMass(connection, galaxy, verbosity)
centroid_center_of_mass.calculate_results()
if len(centroid_center_of_mass.layers) > 0:
connection.execute(
IMAGE_DETAIL.insert(),
galaxy_id=galaxy_id,
x_image=centroid_center_of_mass.dimension_x / 2.0,
y_image=centroid_center_of_mass.dimension_y / 2.0,
x_centroid=numpy.asscalar(centroid_center_of_mass.centroid[0]),
y_centroid=numpy.asscalar(centroid_center_of_mass.centroid[1]),
x_centre_of_mass=numpy.asscalar(centroid_center_of_mass.centre_of_mass[0]),
y_centre_of_mass=numpy.asscalar(centroid_center_of_mass.centre_of_mass[1])
)
else:
connection.execute(
IMAGE_DETAIL.insert(),
galaxy_id=galaxy_id,
x_image=centroid_center_of_mass.dimension_x / 2.0,
y_image=centroid_center_of_mass.dimension_y / 2.0,
x_centroid=-1.0,
y_centroid=-1.0,
x_centre_of_mass=-1.0,
y_centre_of_mass=-1.0
)
add_step_done_id(connection, galaxy_id, STEP_DONE_ID_CENTROID_CENTRE_OF_MASS)
elif verbosity >= 1:
LOG.info('Nothing to do for the galaxy {0}'.format(galaxy_id))
elif verbosity >= 1:
LOG.info('The mask has not been built for the galaxy {0}'.format(galaxy_id))
def process_galaxy(connection, galaxies, verbosity):
"""
Process the galaxies
"""
# Build the tables
build_dynamic_tables(connection)
for galaxy in galaxies:
time_start = time.time()
galaxy_id = galaxy[GALAXY.c.galaxy_id]
steps_done = get_step_done_ids(connection, galaxy_id, STEP_DONE_ID_BUILD_GALAXY_DETAILS, True)
if verbosity:
LOG.info(
'galaxy_id: {0}, galaxy_name: {1}'.format(
galaxy_id,
galaxy[GALAXY.c.name]
)
)
tables_required = get_tables_required(steps_done)
if len(tables_required) > 0:
for table in tables_required:
with connection.begin():
time_start_update = time.time()
create_galaxy_detail = CreateGalaxyDetail(connection, table, galaxy_id)
create_galaxy_detail.calculate_details()
create_galaxy_detail.insert_details()
add_step_done_id(
connection,
galaxy_id,
STEP_DONE_ID_BUILD_GALAXY_DETAILS,
table.name
)
time_end_update = time.time()
LOG.info(
'galaxy_id: {0}, table: {1}, time: {2:.3f}s'.format(
galaxy_id,
table.name,
time_end_update - time_start_update
)
)
time_end = time.time()
LOG.info(
'galaxy_id: {0}, time: {1:.3f}s'.format(
galaxy_id,
time_end - time_start
)
)
def copy_hdf5_files(args):
s3_helper = S3Helper()
bucket_aws_pogs = s3_helper.get_bucket(BUCKET_NAME)
bucket_pogs = s3_helper.get_bucket(BUCKET_NAME_OLD_POGS)
engine = create_engine(DB_LOGIN)
connection = engine.connect()
subquery = select([STEP_DONE.c.galaxy_id]).distinct()
for galaxy in connection.execute(select([GALAXY]).where(~GALAXY.c.galaxy_id.in_(subquery)).order_by(GALAXY.c.galaxy_id)):
if args.verbose:
LOG.info('Checking galaxy_id: {0}, name: {1}'.format(galaxy[GALAXY.c.galaxy_id], galaxy[GALAXY.c.name]))
# noinspection PyBroadException
try:
block_dir = get_block_dir(galaxy[GALAXY.c.galaxy_id] / 1000)
name_aws_pogs = get_galaxy_file_name(galaxy[GALAXY.c.name], galaxy[GALAXY.c.run_id], galaxy[GALAXY.c.galaxy_id])
key_aws_pogs = Key(bucket_aws_pogs)
key_aws_pogs.key = '{0}/{1}.hdf5'.format(block_dir, name_aws_pogs)
if not key_aws_pogs.exists():
# Does it exist in POGS?
name_pogs = get_galaxy_file_name_pogs(galaxy[GALAXY.c.name], galaxy[GALAXY.c.run_id], galaxy[GALAXY.c.galaxy_id])
key_pogs = Key(bucket_pogs)
key_pogs.key = '{0}/{0}.hdf5'.format(name_pogs)
if key_pogs.exists():
check_and_copy_or_restore(galaxy[GALAXY.c.name], galaxy[GALAXY.c.run_id], galaxy[GALAXY.c.galaxy_id], args.verbose)
else:
add_step_done_id(connection, galaxy[GALAXY.c.galaxy_id], STEP_DONE_ID_NO_HDF5_FILE)
except BaseException:
LOG.exception('error during fetch, quitting')
break
if args.verbose:
LOG.info('Done')
def process_galaxy(connection, galaxies):
"""
Process the galaxies
"""
# Build the tables
build_dynamic_tables(connection)
for galaxy in galaxies:
galaxy_id = galaxy[GALAXY.c.galaxy_id]
steps_done = get_step_done_ids(
connection,
galaxy_id,
[STEP_DONE_ID_BUILD_GALAXY_DETAILS, STEP_DONE_ID_NORMALISE_Z_MIN_MAX],
True
)
tables_required = get_tables_required(steps_done)
if len(tables_required) > 0:
time_start = time.time()
for table in tables_required:
with connection.begin():
time_start_update = time.time()
mean_standard = connection.execute(
select([GALAXY_DETAIL]).where(
and_(
GALAXY_DETAIL.c.galaxy_id == galaxy_id,
GALAXY_DETAIL.c.table_name == table.name
)
)
).first()
insert_data = []
if mean_standard is not None:
mean = mean_standard[GALAXY_DETAIL.c.mean]
standard_deviation = mean_standard[GALAXY_DETAIL.c.standard_deviation]
min_value = mean_standard[GALAXY_DETAIL.c.min_value]
max_value = mean_standard[GALAXY_DETAIL.c.max_value]
max_minus_min = max_value - min_value
if max_minus_min is not None and max_minus_min != 0.0 \
and standard_deviation is not None and standard_deviation != 0.0:
if table.name.startswith('original_value__'):
select_statement = select(
[table]
).where(
and_(
table.c.galaxy_id == galaxy_id,
table.c.galaxy_id == MASK_POINT.c.galaxy_id,
table.c.x == MASK_POINT.c.x,
table.c.y == MASK_POINT.c.y,
table.c.value > 0.0
)
)
else:
select_statement = select(
[table]
).where(
and_(
table.c.galaxy_id == galaxy_id,
table.c.galaxy_id == MASK_POINT.c.galaxy_id,
table.c.x == MASK_POINT.c.x,
table.c.y == MASK_POINT.c.y
)
)
for table_data in connection.execute(select_statement):
value = table_data[table.c.value]
z_score = (value - mean) / standard_deviation
insert_data.append(
{
'original_id': table_data[table.c.id],
'value_zscore': z_score,
'value_min_max': (value - min_value) / max_minus_min,
'value_ln': math.log(value - min_value + 1),
'value_softmax': 1.0 / (1.0 + math.exp(-z_score))
}
)
if len(insert_data) > 0:
table_extended = DATA_TABLES_EXTENDED[table.name + '__extended']
connection.execute(table_extended.insert(), insert_data)
add_step_done_id(
connection,
galaxy_id,
STEP_DONE_ID_NORMALISE_Z_MIN_MAX,
table.name
)
time_end_update = time.time()
LOG.info(
'galaxy_id: {0}, table: {1}__extended, time: {2:.3f}s, inserts: {3}'.format(
galaxy_id,
table.name,
time_end_update - time_start_update,
len(insert_data)
)
)
time_end = time.time()
LOG.info('galaxy_id: {0}, time: {1:.3f}s'.format(galaxy_id, time_end - time_start))
