def clustering_process(self, records, zoom, pix_x, pix_y):
"""
Iterate records and create point clusters
We use a simple method that for every point, that is not within any
cluster, calculate it's 'catchment' area and add it to the cluster
If a point is within a cluster 'catchment' area increase point
count for that cluster and recalculate clusters minimum bbox
:param records: list of records.
:type records: list
:param zoom: zoom level of map
:type zoom: int
:param pix_x: pixel x of icon
:type pix_x: int
:param pix_y: pixel y of icon
:type pix_y: int
"""
cluster_points = []
for record in records:
# get x,y of site
centroid = record.get_centroid()
if not centroid:
continue
x = centroid.x
y = centroid.y
# check every point in cluster_points
for pt in cluster_points:
if 'minbbox' not in pt:
pt['minbbox'] = pt['bbox']
if within_bbox((x, y), pt['minbbox']):
# it's in the cluster 'catchment' area
pt['count'] += 1
pt['minbbox'] = update_min_bbox((x, y), pt['minbbox'])
break
else:
# point is not in the catchment area of any cluster
x_range, y_range = overlapping_area(zoom, pix_x, pix_y, y)
bbox = (x - x_range * 1.5, y - y_range * 1.5,
x + x_range * 1.5, y + y_range * 1.5)
serializer = LocationSiteClusterSerializer(record)
new_cluster = {
'count': 1,
'bbox': bbox,
'coordinates': [x, y],
'record': serializer.data
}
cluster_points.append(new_cluster)
return cluster_points
def update_cluster(boundary):
"""
Updating cluster in boundary. It will get all
biological collection children and generate clustering from that
:param boundary: boundary that will be checked
:type boundary: Boundary
"""
from bims.models.biological_collection_record import (
BiologicalCollectionRecord)
from bims.models.location_site import LocationSite
from bims.models.survey import Survey
from bims.serializers.location_site_serializer import (
LocationSiteClusterSerializer)
records = BiologicalCollectionRecord.objects.filter(validated=True).filter(
Q(site__geometry_point__intersects=boundary.geometry)
| Q(site__geometry_line__intersects=boundary.geometry)
| Q(site__geometry_polygon__intersects=boundary.geometry)
| Q(site__geometry_multipolygon__intersects=boundary.geometry))
sites = records.values('site').distinct()
# update boundary of site
LocationSite.objects.filter(id__in=sites).update(boundary=boundary)
# get all children model if no CollectionModel given
verbose_name = LocationSite._meta.verbose_name
surveys = Survey.objects.filter(sites__in=sites)
# create/update new cluster count
try:
cluster = Cluster.objects.get(boundary=boundary, module=verbose_name)
except Cluster.DoesNotExist:
cluster = Cluster()
cluster.boundary = boundary
cluster.module = verbose_name
cluster.site_count = sites.count()
details = {
'records': records.count(),
'sites': sites.count(),
'survey': surveys.count(),
}
if sites.count() == 1:
try:
site = LocationSite.objects.get(id=sites[0]['site'])
details['site_detail'] = LocationSiteClusterSerializer(site).data
except LocationSite.DoesNotExist:
pass
cluster.details = json.dumps(details)
cluster.save()
# logging
logger.info('CLUSTER : %s, %s' % (cluster.boundary, cluster.module))
logger.info('FOUND site_count: %s' % cluster.site_count)
logger.info('DETAILS: %s' % details)
def clustering_process(
collection_records,
site_records,
zoom,
pix_x,
pix_y,
cluster_points=[],
sites=[]):
"""
Iterate records and create point clusters
We use a simple method that for every point, that is not within any
cluster, calculate it's 'catchment' area and add it to the cluster
If a point is within a cluster 'catchment' area increase point
count for that cluster and recalculate clusters minimum bbox
:param collection_records: collection records.
:type collection_records: search query set
:param site_records: site records.
:type site_records: search query set
:param zoom: zoom level of map
:type zoom: int
:param pix_x: pixel x of icon
:type pix_x: int
:param pix_y: pixel y of icon
:type pix_y: int
"""
for collection in GetCollectionAbstract.queryset_gen(
collection_records):
# get x,y of site
try:
x = collection.site.geometry_point.x
y = collection.site.geometry_point.y
location_site = collection.site
if collection.site.id in sites:
continue
sites.append(collection.site_id)
except (ValueError, AttributeError):
continue
# check every point in cluster_points
for pt in cluster_points:
if 'minbbox' not in pt:
pt['minbbox'] = pt['bbox']
if within_bbox((x, y), pt['minbbox']):
# it's in the cluster 'catchment' area
pt['count'] += 1
pt['minbbox'] = update_min_bbox(
(x, y), pt['minbbox'])
break
else:
# point is not in the catchment area of any cluster
x_range, y_range = overlapping_area(
zoom, pix_x, pix_y, y)
bbox = (
x - x_range * 1.5, y - y_range * 1.5,
x + x_range * 1.5, y + y_range * 1.5
)
serializer = LocationSiteClusterSerializer(
location_site)
new_cluster = {
'count': 1,
'bbox': bbox,
'coordinates': [x, y],
'record': serializer.data
}
cluster_points.append(new_cluster)
return cluster_points, sites