def conflate(cdau_address, cat_address, cod_mun_cat):
"""Conflate CDAU over Cadastre addresses datasets"""
cod_mun = cod_mun_cat2ine(cod_mun_cat)
q = "ine_mun='{}' and (tipo_portal_pk='{}' or tipo_portal_pk='{}')"
exp = q.format(cod_mun, 'PORTAL', 'ACCESORIO')
c = 0
addresses = defaultdict(list)
index = cat_address.get_index()
to_add = []
to_change = {}
to_change_g = {}
for feat in cat_address.getFeatures():
g = feat['localId'].split('.')
ref = '.'.join(g[:3] + g[4:])
addresses[ref].append(feat)
for ad in cdau_address.search(exp):
c += 1
attr = get_cat_address(ad, cod_mun_cat)
ref = attr['localId']
pt = layer.Point(float(ad['x']), float(ad['y']))
if len(addresses[ref]) == 0: # can't resolve cadastral reference
area_of_candidates = layer.Point(pt).boundingBox(cdau_thr)
fids = index.intersects(area_of_candidates)
if len(fids) == 0: # no close cadastre address
feat = QgsFeature(cat_address.fields())
for key, value in list(attr.items()):
feat[key] = value
feat.setGeometry(layer.Geometry.fromPointXY(pt))
to_add.append(feat) # add new
else: # get nearest
min_dist = 100
candidate = None
for feat in addresses[ref]:
dist = feat.geometry().asPoint().sqrDist(pt)
if dist < min_dist:
min_dist = dist
candidate = feat
if candidate is not None: # update existing
to_change_g[candidate.id()] = layer.Geometry.fromPointXY(pt)
for key, value in list(attr.items()):
candidate[key] = value
to_change[candidate.id()] = layer.get_attributes(candidate)
log.info(_("Parsed %d addresses from '%s'"), c, 'CDAU')
report.inp_address_cdau = c
if to_change:
cat_address.writer.changeAttributeValues(to_change)
cat_address.writer.changeGeometryValues(to_change_g)
log.info(_("Replaced %d addresses from '%s'"), len(to_change), 'CDAU')
report.rep_address_cdau = len(to_change)
cat_address.source_date = cdau_address.source_date
report.address_date = cdau_address.source_date
if to_add:
cat_address.writer.addFeatures(to_add)
log.info(_("Added %d addresses from '%s'"), len(to_add), 'CDAU')
report.add_address_cdau = len(to_add)
report.inp_address += len(to_add)
report.inp_address_entrance += len(to_add)
cat_address.source_date = cdau_address.source_date
report.address_date = cdau_address.source_date
def _test_duplicates_mem(self):
dup_thr = 0.012
duplicates = defaultdict(list)
index = self.obj.get_index()
vertices_by_fid = {feat.id(): feat for feat in self.obj.getFeatures()}
for vertex in self.obj.getFeatures():
point = layer.Point(vertex.geometry().asPoint())
area_of_candidates = point.boundingBox(dup_thr)
fids = index.intersects(area_of_candidates)
fids.remove(vertex.id())
for fid in fids:
dup = vertices_by_fid[fid].geometry().asPoint()
dist = point.sqrDist(dup)
if dist < dup_thr**2:
duplicates[point].append(dup)
print("duplicados %d" % len(duplicates))
def _test_duplicates_shp1(self):
"""3x slower than test_duplicates_mem, 3x less memory"""
dup_thr = 0.012
duplicates = defaultdict(list)
index = self.obj_shp.get_index()
request = QgsFeatureRequest()
for vertex in self.obj_shp.getFeatures():
point = layer.Point(vertex.geometry().asPoint())
area_of_candidates = point.boundingBox(dup_thr)
fids = index.intersects(area_of_candidates)
fids.remove(vertex.id())
if fids:
request.setFilterFids(fids)
for v in self.obj_shp.getFeatures(request):
dup = v.geometry().asPoint()
dist = point.sqrDist(dup)
if dist < dup_thr**2:
duplicates[point].append(dup)
print("duplicados %d" % len(duplicates))