def built_form_sets(config_entity):
"""
Constructs and persists buildings, buildingtypes, and placetypes and their associates and then returns them all
as a persisted BuiltFormSet. One BuiltFormSet is returned in an array
:param test: if test is set to true, a much more limited set of built forms is created
"""
from footprint.client.configuration.fixture import BuiltFormFixture
from footprint.client.configuration.utils import resolve_fixture
json_fixture = os.path.join(settings.ROOT_PATH, 'built_form_fixture.json')
built_form_fixture = resolve_fixture("built_form", "built_form", BuiltFormFixture, settings.CLIENT, config_entity=config_entity)
if settings.IMPORT_BUILT_FORMS == 'CSV' or (not os.path.exists(json_fixture)):
logger.info('Importing built forms from csv source')
# Get the fixture scoped for the config_entity
# Create any built_form class sets that are configured for the client at the config_entity's class scope
built_forms_dict = built_form_fixture.built_forms()
built_form_fixture.tag_built_forms(built_forms_dict)
built_forms = flatten(built_forms_dict.values())
return map(
lambda built_form_set_config: update_or_create_built_form_set(built_form_set_config, built_forms),
built_form_fixture.built_form_sets())
elif settings.IMPORT_BUILT_FORMS == 'JSON' and not BuiltForm.objects.count():
logger.info('Importing built forms from json fixture at ' + json_fixture)
call_command('loaddata', json_fixture)
return {}
def built_form_sets(config_entity):
"""
Constructs and persists buildings, buildingtypes, and placetypes and their associates and then returns them all
as a persisted BuiltFormSet. One BuiltFormSet is returned in an array
:param test: if test is set to true, a much more limited set of built forms is created
"""
from footprint.client.configuration.fixture import BuiltFormFixture
from footprint.client.configuration.utils import resolve_fixture
json_fixture = os.path.join(settings.ROOT_PATH, 'built_form_fixture.json')
built_form_fixture = resolve_fixture("built_form", "built_form", BuiltFormFixture, settings.CLIENT, config_entity=config_entity)
if settings.IMPORT_BUILT_FORMS == 'CSV' or (not os.path.exists(json_fixture)):
logger.info('Importing built forms from csv source')
# Get the fixture scoped for the config_entity
# Create any built_form class sets that are configured for the client at the config_entity's class scope
built_forms_dict = built_form_fixture.built_forms()
built_form_fixture.tag_built_forms(built_forms_dict)
built_forms = flatten(built_forms_dict.values())
return map(
lambda built_form_set_config: update_or_create_built_form_set(built_form_set_config, built_forms),
built_form_fixture.built_form_sets())
elif settings.IMPORT_BUILT_FORMS == 'JSON' and not BuiltForm.objects.count():
logger.info('Importing built forms from json fixture at ' + json_fixture)
call_command('loaddata', json_fixture)
return {}
def test_config_entity_api__permissions(self):
"""
Make sure that users only get ConfigEntity's that match their permission settings
:return:
"""
permission_configuration = TestConfigEntityPermissions.config_entity_configuration()
resource_name = 'config_entity'
# Iterate through the test_configurstions and extract a user for each group_key
# Make a dict with the user as the key and all the instances from the test_config that the
# user corresponds to. This gives a lookup of a user to the config_entities that we expect
# the user to be able to view
# Create a user->instances dict
# Combine our {user1:instances, user2:instances,...} dicts
user_to_expected_instances = merge_dict_list_values(
*map(
lambda test_configuration:\
# Combine our [user, instance] pairs into {user1:instances, user2:instances,...}
# Remove null keys (owing to groups with no users)
compact_dict(map_to_dict_with_lists(
# Each test_configuration has several groups.
# For each group resolve a user and return [user, instance]
lambda group_key: [
get_first_value_or_none(Group.objects.get(name=group_key).user_set.all()),
test_configuration['instance']],
test_configuration['groups'].keys())),
permission_configuration.test_configuration)
)
all_instances = set(unique(flatten(user_to_expected_instances.values())))
for user, instances in user_to_expected_instances.items():
other_instances = all_instances - set(instances)
# Fetch all instances with this user and create a lookup so we can test
# that the resulting instances are present or not present as expected according to
# the permissions
response = self.get(resource_name, user=user)
result_instance_lookup = map_to_dict(
lambda instance_dict: [int(instance_dict['id']), instance_dict],
self.deserialize(response)['objects'])
for instance in instances:
matching_instance = result_instance_lookup.get(instance.id)
assert_is_not_none(matching_instance,
"User %s should have view permission to instance %s with id %s and key %s but does." % \
(user.username,
instance,
instance.id,
permission_configuration.key_class.Fab.remove(
permission_configuration.instance_key_lambda(instance))))
for instance in other_instances:
assert_is_none(matching_instance,
"User %s should not have view permission to instance %s with id %s and key %s but does." % \
(user.username,
instance,
instance.id,
permission_configuration.key_class.Fab.remove(
permission_configuration.instance_key_lambda(instance))))
def reduce_dict_to_difference(dct, comparison_dict, deep=True):
"""
Given a dict dct and a similar dict comparison dict, return a new dict that only contains the key/values of dct that are different than comparison dict, whether it's a key not in comparison_dict or a matching key with a different value. Specify deep=True to do a comparison of internal dicts
# TODO This could handle list comparison better for deep=True. Right now it just marks the lists as different if they are not equal
:param dct:
:param comparison_dict:
:param deep: Default True, compares embedded dictionaries by recursing
:return: A new dict containing the differences
"""
differ = DictDiffer(dct, comparison_dict)
return merge(
# Find keys and key values changed at the top level
map_to_dict(lambda key: [key, dct[key]], flatten([differ.added(), differ.changed()])),
# If deep==True recurse on dictionaries defined on the values
*map(lambda key: reduce_dict_to_difference(*map(lambda dictionary: dictionary[key], [dct, comparison_dict])),
# recurse on inner each dict pair
# Find just the keys with dict values
filter(lambda key: isinstance(dct[key], dict), differ.unchanged())) if deep else {}
)
def hydrate_presentations(self, bundle):
"""
Does the reverse of dehydrate_presentations. If the user actually wanted to create new presentations via the
API they'd simply save a presentation pointing to the correct configEntity, so we could probably just
disregard this list on post/patch/put.
:param bundle:
:return:
"""
if bundle.data.get('id', 0) == 0:
# We can't handle presentations on new config_entities yet.
# One problem is that tastypie/django doesn't like presentations that are actually
# layer_libraries and result_libraries that have layers and results, respectively
bundle.data['presentations'] = None
return bundle
if bundle.data.get('presentations', None):
bundle.data['presentations'] = flatten(bundle.data['presentations'].values()) if isinstance(bundle.data['presentations'], dict) else bundle.data['presentations']
else:
bundle.data['presentations'] = []
return bundle
def hydrate_presentations(self, bundle):
"""
Does the reverse of dehydrate_presentations. If the user actually wanted to create new presentations via the
API they'd simply save a presentation pointing to the correct configEntity, so we could probably just
disregard this list on post/patch/put.
:param bundle:
:return:
"""
if bundle.data.get('id', 0) == 0:
# We can't handle presentations on new config_entities yet.
# One problem is that tastypie/django doesn't like presentations that are actually
# layer_libraries and result_libraries that have layers and results, respectively
bundle.data['presentations'] = None
return bundle
if bundle.data.get('presentations', None):
bundle.data['presentations'] = flatten(bundle.data['presentations'].values()) if isinstance(bundle.data['presentations'], dict) else bundle.data['presentations']
else:
bundle.data['presentations'] = []
return bundle
def aggregate_within_variable_distance(distance_options):
thread_count = count_cores()
queue = queue_process()
sql_format = 'out {formatter} float'.format(formatter="{0}")
output_field_format = create_sql_calculations(distance_options['variable_field_list'], sql_format, ', ')
sql_format = 'cast({aggregation_type}({formatter}) as float) as {formatter}'.format(formatter="{0}", aggregation_type=distance_options['aggregation_type'])
sql_calculations_format = create_sql_calculations(distance_options['variable_field_list'], sql_format, ', ')
pSql = '''
drop function if exists aggregate_within_variable_distance_tool(
in_id int,
in_distance float,
in_geometry geometry,
out id int,
out wkb_geometry geometry,
{output_field_format}) cascade;'''.format(
output_field_format=output_field_format)
execute_sql(pSql)
pSql = '''
CREATE OR REPLACE FUNCTION aggregate_within_variable_distance_tool(
in_id int,
id_distance float,
in_geometry geometry,
out id int,
out wkb_geometry geometry,
{output_field_format})
AS
$$
select
$1 as id,
$3 as wkb_geometry,
{sql_calculations_format}
from {source_table} ref
WHERE st_dwithin($3, ref.wkb_geometry, $2) and (ref.{source_table_query});
$$
COST 10000
language SQL STABLE strict;
'''.format(source_table=distance_options['source_table'],
source_table_query=distance_options['source_table_query'],
output_field_format=output_field_format,
sql_calculations_format=sql_calculations_format)
execute_sql(pSql)
drop_table('{target_table_schema}.{target_table}_{suffix}'.format(
target_table_schema=distance_options['target_table_schema'],
target_table=distance_options['target_table'], suffix=distance_options['suffix']))
sql_format = '{formatter} float'.format(formatter="{0}", suffix=distance_options['suffix'])
output_table_field_format = create_sql_calculations(distance_options['variable_field_list'], sql_format, ', ')
pSql = '''create table {target_table_schema}.{target_table}_{suffix} (id int, wkb_geometry geometry, {output_table_field_format});'''.format(
target_table_schema=distance_options['target_table_schema'], target_table=distance_options['target_table'],
suffix=distance_options['suffix'], output_table_field_format=output_table_field_format)
execute_sql(pSql)
pSql = 'select cast(id as int) from {source_table} where id is not null order by id'.format(
source_table=distance_options['source_table'])
id_list = flatten(report_sql_values(pSql, 'fetchall'))
insert_sql = '''
insert into {target_table_schema}.{target_table}_{suffix}
select (f).* from (
select aggregate_within_variable_distance_tool(id, distance, wkb_geometry) as f
from {source_table}
where id >= {bottom_range_id} and id <= {top_range_id} and {source_table_query}
) s
where (f).id is not null;
'''.format(
target_table_schema=distance_options['target_table_schema'],
source_table_query=distance_options['source_table_query'],
target_table=distance_options['target_table'],
source_table=distance_options['source_table'],
suffix=distance_options['suffix'],
bottom_range_id="{start_id}",
top_range_id="{end_id}")
for i in range(thread_count):
t = MultithreadProcess(queue, insert_sql)
t.setDaemon(True)
t.start()
#populate queue with data
rows_per_thread = len(id_list) / thread_count
offset = 0
for i in range(thread_count):
if i == thread_count - 1:
## last bucket gets any remainder, too
last_thread = len(id_list) - 1
else:
last_thread = offset + rows_per_thread - 1
rows_to_process = {
'start_id': id_list[offset],
'end_id': id_list[last_thread]
}
offset += rows_per_thread
queue.put(rows_to_process)
#wait on the queue until everything has been processed
queue.join()
add_attribute_idx(distance_options['target_table_schema'],
'{target_table}_{suffix}'.format(target_table=distance_options['target_table'],
suffix=distance_options['suffix']), 'id')
update_table_field_format = create_sql_calculations(distance_options['variable_field_list'], '{0} = (case when b.{0}_var is null then 0 else b.{0}_var end)', ', ')
select_format = create_sql_calculations(distance_options['variable_field_list'], '{0} as {0}_var', ', ')
pSql = '''
update {target_table_schema}.{target_table} a set {update_table_field_format}
from (select id as {suffix}_id, wkb_geometry, {select_format} from {target_table_schema}.{target_table}_{suffix}) b
where st_intersects(st_centroid(a.analysis_geom), b.wkb_geometry) and {target_table_query};
'''.format(
target_table_schema=distance_options['target_table_schema'],
target_table=distance_options['target_table'],
target_table_query=distance_options['target_table_query'],
target_table_pk=distance_options['target_table_pk'],
update_table_field_format=update_table_field_format,
select_format=select_format,
suffix=distance_options['suffix']
)
execute_sql(pSql)
drop_table('{target_table_schema}.{target_table}_{suffix}'.format(
target_table_schema=distance_options['target_table_schema'],
target_table=distance_options['target_table'],
suffix=distance_options['suffix']))
def calculate_distance(distance_options):
print 'Calculating distance from target features areas'
##ST_DISTANCE returns distances in meters from geometries in WGS84 projection if set to false
thread_count = count_cores()
queue = queue_process()
#if the source table query has not results set all values to the max and break
zero_values_check = report_sql_values(
'''select
sum(*)
from {source_table} where {source_table_query};'''.format(**distance_options), 'fetchone')
if len(zero_values_check) == 0:
pSql = '''
update {target_table_schema}.{target_table} a set
{column} = {maximum_distance}
where {target_table_query} and {column} = 0
'''.format(**distance_options)
execute_sql(pSql)
return
pSql = '''drop function if exists distance_tool(
in_id int,
in_wkb_geometry geometry,
out id int,
out {column} float) cascade;'''.format(**distance_options)
execute_sql(pSql)
pSql = '''
CREATE OR REPLACE FUNCTION distance_tool(
in_id int,
in_wkb_geometry geometry,
out id int,
out {column} float)
AS
$$
select
$1 as id,
cast(st_distance(st_centroid($2), st_centroid(ref.geometry)) as float) as {column}
from
(select *, {source_geometry_column} as geometry from {source_table}) ref
where ST_DWITHIN($2, ref.geometry, {maximum_distance}) and ({source_table_query})
order by {column};
$$
COST 10000
language SQL STABLE strict;
'''.format(**distance_options)
execute_sql(pSql)
drop_table('{target_table_schema}.{target_table}_{column}'.format(**distance_options))
pSql = '''
create table {target_table_schema}.{target_table}_{column} (id int, {column} float);
'''.format(**distance_options)
execute_sql(pSql)
id_list = flatten(report_sql_values(
'''select cast({target_table_pk} as int) from {target_table_schema}.{target_table}
where {target_table_query} order by {target_table_pk}'''.format(
**distance_options), 'fetchall'))
insert_sql = '''
insert into {target_table_schema}.{target_table}_{column}
select (f).* from (
select distance_tool(a.id, a.wkb_geometry) as f
from (select {target_table_pk} as id, {target_geometry_column} as wkb_geometry
from {target_table_schema}.{target_table}
where
{target_table_pk} >= {bottom_range_id} and
{target_table_pk} <= {top_range_id} and
({target_table_query})
) a
) s;
'''.format(bottom_range_id="{start_id}", top_range_id="{end_id}", **distance_options)
for i in range(thread_count):
t = MultithreadProcess(queue, insert_sql)
t.setDaemon(True)
t.start()
#populate queue with data
rows_per_thread = len(id_list) / thread_count
offset = 0
for i in range(thread_count):
if i == thread_count - 1:
## last bucket gets any remainder, too
last_thread = len(id_list) - 1
else:
last_thread = offset + rows_per_thread - 1
rows_to_process = {
'start_id': id_list[offset],
'end_id': id_list[last_thread]
}
offset += rows_per_thread
queue.put(rows_to_process)
#wait on the queue until everything has been processed
queue.join()
add_attribute_idx(distance_options['target_table_schema'],
'{target_table}_{column}'.format(**distance_options), 'id')
pSql = '''
update {target_table_schema}.{target_table} a set
{column} = source_column
from (select id as source_id, {column} as source_column from {target_table_schema}.{target_table}_{column}) b
where cast(a.{target_table_pk} as int) = b.source_id and ({target_table_query})
'''.format(**distance_options)
execute_sql(pSql)
pSql = '''
update {target_table_schema}.{target_table} a set
{column} = {maximum_distance}
where {target_table_query} and {column} = 0
'''.format(**distance_options)
execute_sql(pSql)
drop_table('{target_table_schema}.{target_table}_{column}'.format(**distance_options))
def aggregate_within_distance(distance_options):
thread_count = count_cores()
queue = queue_process()
source_table_column_list = []
for key, value in distance_options['variable_field_dict'].items():
source_table_column_list += value
source_table_column_list = list(set(source_table_column_list))
sql_format = 'out {formatter} float'.format(formatter="{0}")
output_field_format = create_sql_calculations(source_table_column_list, sql_format, ', ')
sql_format = 'cast({aggregation_type}({formatter}) as float) as {formatter}_{suffix}'.format(formatter="{0}", **distance_options)
sql_calculations_format = create_sql_calculations(source_table_column_list, sql_format, ', ')
pSql = '''drop function if exists aggregate_within_distance_tool(
in_id int,
in_wkb_geometry geometry,
out id int,
{output_field_format}) cascade;'''.format(
output_field_format=output_field_format)
execute_sql(pSql)
pSql = '''
CREATE OR REPLACE FUNCTION aggregate_within_distance_tool(
in_id int,
in_wkb_geometry geometry,
out id int,
{output_field_format})
AS
$$
select
$1 as id,
{sql_calculations_format}
from (select *, {source_geometry_column} as geometry from {source_table}) ref
where ST_DWITHIN( $2, ref.geometry, {distance}) and (ref.{source_table_query});
$$
COST 10000
language SQL STABLE strict;
'''.format(source_table=distance_options['source_table'],
source_table_query=distance_options['source_table_query'],
distance=distance_options['distance'],
source_geometry_column=distance_options['source_geometry_column'],
output_field_format=output_field_format,
sql_calculations_format=sql_calculations_format)
execute_sql(pSql)
drop_table('{target_table_schema}.{target_table}_{suffix}'.format(
target_table_schema=distance_options['target_table_schema'],
target_table=distance_options['target_table'], suffix=distance_options['suffix']))
sql_format = '{formatter}_{suffix} float'.format(formatter="{0}", **distance_options)
output_table_field_format = create_sql_calculations(source_table_column_list, sql_format, ', ')
pSql = '''create table {target_table_schema}.{target_table}_{suffix} (id int, {output_table_field_format});'''.format(
target_table_schema=distance_options['target_table_schema'], target_table=distance_options['target_table'],
suffix=distance_options['suffix'], output_table_field_format=output_table_field_format)
execute_sql(pSql)
pSql = 'select cast({target_table_pk} as int) from {target_table_schema}.{target_table} where {target_table_query} order by {target_table_pk}'.format(
target_table_schema=distance_options['target_table_schema'],
target_table=distance_options['target_table'],
target_table_pk=distance_options['target_table_pk'],
target_table_query=distance_options['target_table_query'])
id_list = flatten(report_sql_values(pSql, 'fetchall'))
insert_sql = '''
insert into {target_table_schema}.{target_table}_{suffix}
select (f).* from (
select aggregate_within_distance_tool({target_table_pk}, {target_geometry_column}) as f
from {target_table_schema}.{target_table}
where
{target_table_pk} >= {bottom_range_id} and
{target_table_pk} <= {top_range_id} and
{target_table_query}
offset 0) s
where (f).id is not null;
'''.format(
target_table_schema=distance_options['target_table_schema'],
target_table=distance_options['target_table'],
target_table_query=distance_options['target_table_query'],
target_geometry_column=distance_options['target_geometry_column'],
source_table=distance_options['source_table'],
suffix=distance_options['suffix'],
target_table_pk=distance_options['target_table_pk'],
bottom_range_id="{start_id}",
top_range_id="{end_id}")
for i in range(thread_count):
t = MultithreadProcess(queue, insert_sql)
t.setDaemon(True)
t.start()
#populate queue with data
rows_per_thread = len(id_list) / thread_count
offset = 0
for i in range(thread_count):
if i == thread_count - 1:
## last bucket gets any remainder, too
last_thread = len(id_list) - 1
else:
last_thread = offset + rows_per_thread - 1
rows_to_process = {
'start_id': id_list[offset],
'end_id': id_list[last_thread]
}
offset += rows_per_thread
queue.put(rows_to_process)
#wait on the queue until everything has been processed
queue.join()
add_attribute_idx(distance_options['target_table_schema'],
'{target_table}_{suffix}'.format(target_table=distance_options['target_table'],
suffix=distance_options['suffix']), 'id')
count = 1
update_sql_format = ''
if len(distance_options['variable_field_dict']) > 0:
for key, value in distance_options['variable_field_dict'].items():
update_table_field_format = create_sql_calculations(value, '{formatter}_{suffix}'.format(formatter='b.{0}',
**distance_options), ' + ')
if count == 1:
update_sql_format += key + ' = ' + "(case when {0} is null then 0 else {0} end)".format(update_table_field_format)
else:
update_sql_format += ', ' + key + ' = ' + "(case when {0} is null then 0 else {0} end)".format(update_table_field_format)
count +=1
pSql = '''
update {target_table_schema}.{target_table} a set {update_sql_format}
from (select * from {target_table_schema}.{target_table}_{suffix}) b
where a.{target_table_pk} = b.id and {target_table_query}
'''.format(
target_table_schema=distance_options['target_table_schema'],
target_table=distance_options['target_table'],
target_table_query=distance_options['target_table_query'],
target_table_pk=distance_options['target_table_pk'],
update_sql_format=update_sql_format,
suffix=distance_options['suffix']
)
execute_sql(pSql)
drop_table('{target_table_schema}.{target_table}_{suffix}'.format(
target_table_schema=distance_options['target_table_schema'],
target_table=distance_options['target_table'],
suffix=distance_options['suffix']))
def test_config_entity_api__permissions(self):
"""
Make sure that users only get ConfigEntity's that match their permission settings
:return:
"""
permission_configuration = TestConfigEntityPermissions.config_entity_configuration(
)
resource_name = 'config_entity'
# Iterate through the test_configurstions and extract a user for each group_key
# Make a dict with the user as the key and all the instances from the test_config that the
# user corresponds to. This gives a lookup of a user to the config_entities that we expect
# the user to be able to view
# Create a user->instances dict
# Combine our {user1:instances, user2:instances,...} dicts
user_to_expected_instances = merge_dict_list_values(
*map(
lambda test_configuration:\
# Combine our [user, instance] pairs into {user1:instances, user2:instances,...}
# Remove null keys (owing to groups with no users)
compact_dict(map_to_dict_with_lists(
# Each test_configuration has several groups.
# For each group resolve a user and return [user, instance]
lambda group_key: [
get_first_value_or_none(Group.objects.get(name=group_key).user_set.all()),
test_configuration['instance']],
test_configuration['groups'].keys())),
permission_configuration.test_configuration)
)
all_instances = set(
unique(flatten(user_to_expected_instances.values())))
for user, instances in user_to_expected_instances.items():
other_instances = all_instances - set(instances)
# Fetch all instances with this user and create a lookup so we can test
# that the resulting instances are present or not present as expected according to
# the permissions
response = self.get(resource_name, user=user)
result_instance_lookup = map_to_dict(
lambda instance_dict:
[int(instance_dict['id']), instance_dict],
self.deserialize(response)['objects'])
for instance in instances:
matching_instance = result_instance_lookup.get(instance.id)
assert_is_not_none(matching_instance,
"User %s should have view permission to instance %s with id %s and key %s but does." % \
(user.username,
instance,
instance.id,
permission_configuration.key_class.Fab.remove(
permission_configuration.instance_key_lambda(instance))))
for instance in other_instances:
assert_is_none(matching_instance,
"User %s should not have view permission to instance %s with id %s and key %s but does." % \
(user.username,
instance,
instance.id,
permission_configuration.key_class.Fab.remove(
permission_configuration.instance_key_lambda(instance))))
def aggregate_within_variable_distance(distance_options):
thread_count = count_cores()
queue = queue_process()
sql_format = 'out {formatter} float'.format(formatter="{0}")
output_field_format = create_sql_calculations(distance_options['variable_field_list'], sql_format, ', ')
sql_format = 'cast({aggregation_type}({formatter}) as float) as {formatter}'.format(formatter="{0}", aggregation_type=distance_options['aggregation_type'])
sql_calculations_format = create_sql_calculations(distance_options['variable_field_list'], sql_format, ', ')
pSql = '''
drop function if exists aggregate_within_variable_distance_tool(
in_id int,
in_distance float,
in_geometry geometry,
out id int,
out wkb_geometry geometry,
{output_field_format}) cascade;'''.format(
output_field_format=output_field_format)
execute_sql(pSql)
pSql = '''
CREATE OR REPLACE FUNCTION aggregate_within_variable_distance_tool(
in_id int,
id_distance float,
in_geometry geometry,
out id int,
out wkb_geometry geometry,
{output_field_format})
AS
$$
select
$1 as id,
$3 as wkb_geometry,
{sql_calculations_format}
from {source_table} ref
WHERE st_dwithin($3, ref.wkb_geometry, $2) and (ref.{source_table_query});
$$
COST 10000
language SQL STABLE strict;
'''.format(source_table=distance_options['source_table'],
source_table_query=distance_options['source_table_query'],
output_field_format=output_field_format,
sql_calculations_format=sql_calculations_format)
execute_sql(pSql)
drop_table('{target_table_schema}.{target_table}_{suffix}'.format(
target_table_schema=distance_options['target_table_schema'],
target_table=distance_options['target_table'], suffix=distance_options['suffix']))
sql_format = '{formatter} float'.format(formatter="{0}", suffix=distance_options['suffix'])
output_table_field_format = create_sql_calculations(distance_options['variable_field_list'], sql_format, ', ')
pSql = '''create table {target_table_schema}.{target_table}_{suffix} (id int, wkb_geometry geometry, {output_table_field_format});'''.format(
target_table_schema=distance_options['target_table_schema'], target_table=distance_options['target_table'],
suffix=distance_options['suffix'], output_table_field_format=output_table_field_format)
execute_sql(pSql)
pSql = 'select cast(id as int) from {source_table} where id is not null order by id'.format(
source_table=distance_options['source_table'])
id_list = flatten(report_sql_values(pSql, 'fetchall'))
insert_sql = '''
insert into {target_table_schema}.{target_table}_{suffix}
select (f).* from (
select aggregate_within_variable_distance_tool(id, distance, wkb_geometry) as f
from {source_table}
where id >= {bottom_range_id} and id <= {top_range_id} and {source_table_query}
) s
where (f).id is not null;
'''.format(
target_table_schema=distance_options['target_table_schema'],
source_table_query=distance_options['source_table_query'],
target_table=distance_options['target_table'],
source_table=distance_options['source_table'],
suffix=distance_options['suffix'],
bottom_range_id="{start_id}",
top_range_id="{end_id}")
for i in range(thread_count):
t = MultithreadProcess(queue, insert_sql)
t.setDaemon(True)
t.start()
#populate queue with data
rows_per_thread = len(id_list) / thread_count
offset = 0
for i in range(thread_count):
if i == thread_count - 1:
## last bucket gets any remainder, too
last_thread = len(id_list) - 1
else:
last_thread = offset + rows_per_thread - 1
rows_to_process = {
'start_id': id_list[offset],
'end_id': id_list[last_thread]
}
offset += rows_per_thread
queue.put(rows_to_process)
#wait on the queue until everything has been processed
queue.join()
add_attribute_idx(distance_options['target_table_schema'],
'{target_table}_{suffix}'.format(target_table=distance_options['target_table'],
suffix=distance_options['suffix']), 'id')
update_table_field_format = create_sql_calculations(distance_options['variable_field_list'], '{0} = (case when b.{0}_var is null then 0 else b.{0}_var end)', ', ')
select_format = create_sql_calculations(distance_options['variable_field_list'], '{0} as {0}_var', ', ')
pSql = '''
update {target_table_schema}.{target_table} a set {update_table_field_format}
from (select id as {suffix}_id, wkb_geometry, {select_format} from {target_table_schema}.{target_table}_{suffix}) b
where st_intersects(st_centroid(a.analysis_geom), b.wkb_geometry) and {target_table_query};
'''.format(
target_table_schema=distance_options['target_table_schema'],
target_table=distance_options['target_table'],
target_table_query=distance_options['target_table_query'],
target_table_pk=distance_options['target_table_pk'],
update_table_field_format=update_table_field_format,
select_format=select_format,
suffix=distance_options['suffix']
)
execute_sql(pSql)
drop_table('{target_table_schema}.{target_table}_{suffix}'.format(
target_table_schema=distance_options['target_table_schema'],
target_table=distance_options['target_table'],
suffix=distance_options['suffix']))
def calculate_distance(distance_options):
print 'Calculating distance from target features areas'
##ST_DISTANCE returns distances in meters from geometries in WGS84 projection if set to false
thread_count = count_cores()
queue = queue_process()
#if the source table query has not results set all values to the max and break
zero_values_check = report_sql_values(
'''select
sum(*)
from {source_table} where {source_table_query};'''.format(**distance_options), 'fetchone')
if len(zero_values_check) == 0:
pSql = '''
update {target_table_schema}.{target_table} a set
{column} = {maximum_distance}
where {target_table_query} and {column} = 0
'''.format(**distance_options)
execute_sql(pSql)
return
pSql = '''drop function if exists distance_tool(
in_id int,
in_wkb_geometry geometry,
out id int,
out {column} float) cascade;'''.format(**distance_options)
execute_sql(pSql)
pSql = '''
CREATE OR REPLACE FUNCTION distance_tool(
in_id int,
in_wkb_geometry geometry,
out id int,
out {column} float)
AS
$$
select
$1 as id,
cast(st_distance(st_centroid($2), st_centroid(ref.geometry)) as float) as {column}
from
(select *, {source_geometry_column} as geometry from {source_table}) ref
where ST_DWITHIN($2, ref.geometry, {maximum_distance}) and ({source_table_query})
order by {column};
$$
COST 10000
language SQL STABLE strict;
'''.format(**distance_options)
execute_sql(pSql)
drop_table('{target_table_schema}.{target_table}_{column}'.format(**distance_options))
pSql = '''
create table {target_table_schema}.{target_table}_{column} (id int, {column} float);
'''.format(**distance_options)
execute_sql(pSql)
id_list = flatten(report_sql_values(
'''select cast({target_table_pk} as int) from {target_table_schema}.{target_table}
where {target_table_query} order by {target_table_pk}'''.format(
**distance_options), 'fetchall'))
insert_sql = '''
insert into {target_table_schema}.{target_table}_{column}
select (f).* from (
select distance_tool(a.id, a.wkb_geometry) as f
from (select {target_table_pk} as id, {target_geometry_column} as wkb_geometry
from {target_table_schema}.{target_table}
where
{target_table_pk} >= {bottom_range_id} and
{target_table_pk} <= {top_range_id} and
({target_table_query})
) a
) s;
'''.format(bottom_range_id="{start_id}", top_range_id="{end_id}", **distance_options)
for i in range(thread_count):
t = MultithreadProcess(queue, insert_sql)
t.setDaemon(True)
t.start()
#populate queue with data
rows_per_thread = len(id_list) / thread_count
offset = 0
for i in range(thread_count):
if i == thread_count - 1:
## last bucket gets any remainder, too
last_thread = len(id_list) - 1
else:
last_thread = offset + rows_per_thread - 1
rows_to_process = {
'start_id': id_list[offset],
'end_id': id_list[last_thread]
}
offset += rows_per_thread
queue.put(rows_to_process)
#wait on the queue until everything has been processed
queue.join()
add_attribute_idx(distance_options['target_table_schema'],
'{target_table}_{column}'.format(**distance_options), 'id')
pSql = '''
update {target_table_schema}.{target_table} a set
{column} = source_column
from (select id as source_id, {column} as source_column from {target_table_schema}.{target_table}_{column}) b
where cast(a.{target_table_pk} as int) = b.source_id and ({target_table_query})
'''.format(**distance_options)
execute_sql(pSql)
pSql = '''
update {target_table_schema}.{target_table} a set
{column} = {maximum_distance}
where {target_table_query} and {column} = 0
'''.format(**distance_options)
execute_sql(pSql)
drop_table('{target_table_schema}.{target_table}_{column}'.format(**distance_options))
def aggregate_within_distance(distance_options):
thread_count = count_cores()
queue = queue_process()
source_table_column_list = []
for key, value in distance_options['variable_field_dict'].items():
source_table_column_list += value
source_table_column_list = list(set(source_table_column_list))
sql_format = 'out {formatter} float'.format(formatter="{0}")
output_field_format = create_sql_calculations(source_table_column_list, sql_format, ', ')
sql_format = 'cast({aggregation_type}({formatter}) as float) as {formatter}_{suffix}'.format(formatter="{0}", **distance_options)
sql_calculations_format = create_sql_calculations(source_table_column_list, sql_format, ', ')
pSql = '''drop function if exists aggregate_within_distance_tool(
in_id int,
in_wkb_geometry geometry,
out id int,
{output_field_format}) cascade;'''.format(
output_field_format=output_field_format)
execute_sql(pSql)
pSql = '''
CREATE OR REPLACE FUNCTION aggregate_within_distance_tool(
in_id int,
in_wkb_geometry geometry,
out id int,
{output_field_format})
AS
$$
select
$1 as id,
{sql_calculations_format}
from (select *, {source_geometry_column} as geometry from {source_table}) ref
where ST_DWITHIN( $2, ref.geometry, {distance}) and (ref.{source_table_query});
$$
COST 10000
language SQL STABLE strict;
'''.format(source_table=distance_options['source_table'],
source_table_query=distance_options['source_table_query'],
distance=distance_options['distance'],
source_geometry_column=distance_options['source_geometry_column'],
output_field_format=output_field_format,
sql_calculations_format=sql_calculations_format)
execute_sql(pSql)
drop_table('{target_table_schema}.{target_table}_{suffix}'.format(
target_table_schema=distance_options['target_table_schema'],
target_table=distance_options['target_table'], suffix=distance_options['suffix']))
sql_format = '{formatter}_{suffix} float'.format(formatter="{0}", **distance_options)
output_table_field_format = create_sql_calculations(source_table_column_list, sql_format, ', ')
pSql = '''create table {target_table_schema}.{target_table}_{suffix} (id int, {output_table_field_format});'''.format(
target_table_schema=distance_options['target_table_schema'], target_table=distance_options['target_table'],
suffix=distance_options['suffix'], output_table_field_format=output_table_field_format)
execute_sql(pSql)
pSql = 'select cast({target_table_pk} as int) from {target_table_schema}.{target_table} where {target_table_query} order by {target_table_pk}'.format(
target_table_schema=distance_options['target_table_schema'],
target_table=distance_options['target_table'],
target_table_pk=distance_options['target_table_pk'],
target_table_query=distance_options['target_table_query'])
id_list = flatten(report_sql_values(pSql, 'fetchall'))
insert_sql = '''
insert into {target_table_schema}.{target_table}_{suffix}
select (f).* from (
select aggregate_within_distance_tool({target_table_pk}, {target_geometry_column}) as f
from {target_table_schema}.{target_table}
where
{target_table_pk} >= {bottom_range_id} and
{target_table_pk} <= {top_range_id} and
{target_table_query}
offset 0) s
where (f).id is not null;
'''.format(
target_table_schema=distance_options['target_table_schema'],
target_table=distance_options['target_table'],
target_table_query=distance_options['target_table_query'],
target_geometry_column=distance_options['target_geometry_column'],
source_table=distance_options['source_table'],
suffix=distance_options['suffix'],
target_table_pk=distance_options['target_table_pk'],
bottom_range_id="{start_id}",
top_range_id="{end_id}")
for i in range(thread_count):
t = MultithreadProcess(queue, insert_sql)
t.setDaemon(True)
t.start()
#populate queue with data
rows_per_thread = len(id_list) / thread_count
offset = 0
for i in range(thread_count):
if i == thread_count - 1:
## last bucket gets any remainder, too
last_thread = len(id_list) - 1
else:
last_thread = offset + rows_per_thread - 1
rows_to_process = {
'start_id': id_list[offset],
'end_id': id_list[last_thread]
}
offset += rows_per_thread
queue.put(rows_to_process)
#wait on the queue until everything has been processed
queue.join()
add_attribute_idx(distance_options['target_table_schema'],
'{target_table}_{suffix}'.format(target_table=distance_options['target_table'],
suffix=distance_options['suffix']), 'id')
count = 1
update_sql_format = ''
if len(distance_options['variable_field_dict']) > 0:
for key, value in distance_options['variable_field_dict'].items():
update_table_field_format = create_sql_calculations(value, '{formatter}_{suffix}'.format(formatter='b.{0}',
**distance_options), ' + ')
if count == 1:
update_sql_format += key + ' = ' + "(case when {0} is null then 0 else {0} end)".format(update_table_field_format)
else:
update_sql_format += ', ' + key + ' = ' + "(case when {0} is null then 0 else {0} end)".format(update_table_field_format)
count +=1
pSql = '''
update {target_table_schema}.{target_table} a set {update_sql_format}
from (select * from {target_table_schema}.{target_table}_{suffix}) b
where a.{target_table_pk} = b.id and {target_table_query}
'''.format(
target_table_schema=distance_options['target_table_schema'],
target_table=distance_options['target_table'],
target_table_query=distance_options['target_table_query'],
target_table_pk=distance_options['target_table_pk'],
update_sql_format=update_sql_format,
suffix=distance_options['suffix']
)
execute_sql(pSql)
drop_table('{target_table_schema}.{target_table}_{suffix}'.format(
target_table_schema=distance_options['target_table_schema'],
target_table=distance_options['target_table'],
suffix=distance_options['suffix']))
def __init__(self, db_entity, attribute):
self.db_entity = db_entity
self.attribute = attribute
self.feature_class = db_entity.feature_class
self.unique_values = flatten(map(lambda value: value.values(), self.get_unique_values))
def __init__(self, db_entity, attribute):
self.db_entity = db_entity
self.attribute = attribute
self.feature_class = db_entity.feature_class
self.unique_values = flatten(
map(lambda value: value.values(), self.get_unique_values))