def __or__(self, x):
return self.__class__(
weightedUnion(self._dict, x._dict)[1],
union(self._words, x._words),
self._index,
)
return self.__class__(result, self._words+x._words, self._index)
def _trivial(L):
# L is empty or has only one (mapping, weight) pair. If there is a
# pair, we may still need to multiply the mapping by its weight.
assert len(L) <= 1
if len(L) == 0:
return IIBucket()
[(result, weight)] = L
if weight != 1:
dummy, result = weightedUnion(IIBucket(), result, 0, weight)
return result
def _trivial(l_):
# l is empty or has only one (mapping, weight) pair. If there is a
# pair, we may still need to multiply the mapping by its weight.
assert len(l_) <= 1
if len(l_) == 0:
return IIBucket()
[(result, weight)] = l_
if weight != 1:
dummy, result = weightedUnion(IIBucket(), result, 0, weight)
return result
def mass_weightedUnion(L):
"A list of (mapping, weight) pairs -> their weightedUnion IIBucket."
if len(L) < 2:
return _trivial(L)
# Balance unions as closely as possible, smallest to largest.
merge = NBest(len(L))
for x, weight in L:
merge.add((x, weight), len(x))
while len(merge) > 1:
# Merge the two smallest so far, and add back to the queue.
(x, wx), dummy = merge.pop_smallest()
(y, wy), dummy = merge.pop_smallest()
dummy, z = weightedUnion(x, y, wx, wy)
merge.add((z, 1), len(z))
(result, weight), dummy = merge.pop_smallest()
return result
def mass_weightedUnion(l_):
"A list of (mapping, weight) pairs -> their weightedUnion IIBucket."
if len(l_) < 2:
return _trivial(l_)
# Balance unions as closely as possible, smallest to largest.
merge = NBest(len(l_))
for x, weight in l_:
merge.add((x, weight), len(x))
while len(merge) > 1:
# Merge the two smallest so far, and add back to the queue.
(x, wx), dummy = merge.pop_smallest()
(y, wy), dummy = merge.pop_smallest()
dummy, z = weightedUnion(x, y, wx, wy)
merge.add((z, 1), len(z))
(result, weight), dummy = merge.pop_smallest()
return result
def __or__(self, x):
return self.__class__(
weightedUnion(self._dict, x._dict)[1],
union(self._words, x._words),
self._index,
)
def __or__(self, x):
return self.__class__(
weightedUnion(self._dict, x._dict),
union(self._words, x._words),
self._index,
)
def getClusters(catalog_tool, filters):
# the objects are searched for in the tile limits (to get the same clusters every time)
grid_size = 16 # geopoints' and clusters' density on map / also depends on map frame size
# unpack map limits
if filters:
lat_min = float(filters[0]['geo_latitude']['query'][0])
lat_max = float(filters[0]['geo_latitude']['query'][1])
lon_min = float(filters[0]['geo_longitude']['query'][0])
lon_max = float(filters[0]['geo_longitude']['query'][1])
else: # this should not happen
return [], []
tlat_min, tlat_max, tlon_min, tlon_max = clusters.get_discretized_limits(
lat_min, lat_max, lon_min, lon_max, grid_size)
catalog = catalog_tool._catalog
# getting the inner indexes for lat and lon
lat_index = catalog.getIndex('geo_latitude')._index
lon_index = catalog.getIndex('geo_longitude')._index
# adjust to cover results outside frame, but very close to margins
# trying to fix cluster flickering near margins
# applying the lat and lon indexes to get the rids
rs = None
lat_set, lat_dict = _apply_index_with_range_dict_results(
lat_index, Decimal(str(tlat_min)), Decimal(str(tlat_max)))
w, rs = weightedIntersection(rs, lat_set)
lon_set, lon_dict = _apply_index_with_range_dict_results(
lon_index, Decimal(str(tlon_min)), Decimal(str(tlon_max)))
w, rs = weightedIntersection(rs, lon_set)
rs_final = None
# OR the filters and apply the index for each one
for f in filters:
rs_f = rs
#adjust geo limits in filters to be consistent with discretized tile limits
f['geo_longitude']['query'] = (Decimal(str(tlon_min)),
Decimal(str(tlon_max)))
f['geo_latitude']['query'] = (Decimal(str(tlat_min)),
Decimal(str(tlat_max)))
#this code is from the search function in the catalog implementation in Zope
for i in catalog.indexes.keys():
index = catalog.getIndex(i)
_apply_index = getattr(index, "_apply_index", None)
if _apply_index is None:
continue
r = _apply_index(f)
if r is not None:
r, u = r
w, rs_f = weightedIntersection(rs_f, r)
w, rs_final = weightedUnion(rs_f, rs_final)
r_list = list(rs_final)
# transform objects to points
points = []
for i in range(len(r_list)):
points.append(
clusters.Point(i, float(lat_dict[r_list[i]]),
float(lon_dict[r_list[i]])))
centers, groups = clusters.kmeans(tlat_min, tlat_max, tlon_min, tlon_max,
points, grid_size)
# transform group points to rids
for i in range(len(groups)):
groups[i] = map(lambda p: r_list[p.id], groups[i])
return centers, groups
def getClusters(catalog_tool, filters):
# the objects are searched for in the tile limits (to get the same clusters every time)
grid_size = 12 # geopoints' and clusters' density on map / also depends on map frame size
# unpack map limits
if filters:
lat_min = float(filters[0]['geo_latitude']['query'][0])
lat_max = float(filters[0]['geo_latitude']['query'][1])
lon_min = float(filters[0]['geo_longitude']['query'][0])
lon_max = float(filters[0]['geo_longitude']['query'][1])
else: # this should not happen
return [], []
tlat_min, tlat_max, tlon_min, tlon_max = clusters.get_discretized_limits(lat_min, lat_max, lon_min, lon_max, grid_size)
catalog = catalog_tool._catalog
# getting the inner indexes for lat and lon
lat_index = catalog.getIndex('geo_latitude')._index
lon_index = catalog.getIndex('geo_longitude')._index
# adjust to cover results outside frame, but very close to margins
# trying to fix cluster flickering near margins
# applying the lat and lon indexes to get the rids
rs = None
lat_set, lat_dict = _apply_index_with_range_dict_results(lat_index, Decimal(str(tlat_min)), Decimal(str(tlat_max)))
w, rs = weightedIntersection(rs, lat_set)
lon_set, lon_dict = _apply_index_with_range_dict_results(lon_index, Decimal(str(tlon_min)), Decimal(str(tlon_max)))
w, rs = weightedIntersection(rs, lon_set)
rs_final = None
# OR the filters and apply the index for each one
for f in filters:
rs_f = rs
#adjust geo limits in filters to be consistent with discretized tile limits
f['geo_longitude']['query'] = (Decimal(str(tlon_min)), Decimal(str(tlon_max)))
f['geo_latitude']['query'] = (Decimal(str(tlat_min)), Decimal(str(tlat_max)))
#this code is from the search function in the catalog implementation in Zope
for i in catalog.indexes.keys():
index = catalog.getIndex(i)
_apply_index = getattr(index, "_apply_index", None)
if _apply_index is None:
continue
r = _apply_index(f)
if r is not None:
r, u = r
w, rs_f = weightedIntersection(rs_f, r)
w, rs_final = weightedUnion(rs_f, rs_final)
r_list = list(rs_final)
# transform objects to points
points = []
for i in range(len(r_list)):
points.append(clusters.Point(i, float(lat_dict[r_list[i]]), float(lon_dict[r_list[i]])))
centers, groups = clusters.kmeans(tlat_min, tlat_max, tlon_min, tlon_max, points, grid_size)
# transform group points to rids
for i in range(len(groups)):
groups[i] = map(lambda p: r_list[p.id], groups[i])
return centers, groups
