def cluster(self, data, keys, save_path):
self._clust_len = len(self._cluster) - 1
self._basic_cluster(data, keys, self._cluster_tree, 0)
_file.op_file(file_path=save_path,
data=self._cluster_tree,
model='json',
method='save')
"""
def search_tencent(self,
dts,
root_path,
branchs=2,
candidate=1,
distance=3):
root = _file.op_file(root_path, method='read')
dist = {}
for idx, i in enumerate(root['center_point']):
val = data.point_distance(dts, i)
dist[round(val, 3)] = idx
keys = list(dist.keys())
keys.sort()
sel_point = len(root['center_point']) if len(
root['center_point']) < branchs[0] else branchs[0]
branchs.pop(0)
all_res = []
for j in range(sel_point):
k = 'file' + str(dist[keys[j]])
path = root['festival'][k]
sr = self.similirity(dts, path, branchs, candidate, distance)
all_res.extend(sr)
boult = [g for g in sorted(all_res, key=lambda k: k[0], reverse=False)]
save_len = candidate if candidate < len(boult) else len(boult)
return boult[0:save_len]
def take9_file(self, root_path):
file_tree = {
"tree0": 0,
"tree1": 0,
"tree2": 0,
"tree3": 0,
"tree4": 0,
"tree5": 0,
"tree6": 0,
"tree7": 0,
"tree8": 0
}
for f in range(3, 9):
key = 'tree' + str(f)
f_p = 'data/tencent/tree' + str(f) + '.json'
file_tree[key] = _file.op_file(f_p, method='read')
vals = list(file_tree[key].values())
ks = list(file_tree[key].keys())
sp = 'data/tencent/tc_tree' + str(f) + '.json'
self.cluster(vals, ks, sp)
del file_tree[key]
del self._cluster_tree
self._cluster_tree = {
"position": 'root',
"festival": [],
"center_point": None
}
print(ord)
def tencent(self, data, words, clusters=[5]):
_kmeans_tree = {"position": "root", "center_point": [], "festival": {}}
class_data = {}
one = clusters.pop(0)
km = KMeans(init="k-means++", n_clusters=one)
km.fit_predict(data)
points = []
for j, i in enumerate(km.cluster_centers_):
key = 'file' + str(j)
points.append(list(i))
class_data[key] = {}
_kmeans_tree['center_point'] = points
# 将所有数据按类分开,存成字典
for a, b in enumerate(km.labels_):
key2 = 'file' + str(b)
class_data[key2][words[a]] = data[a]
# 各类存到不同的文件
for idx in range(one):
key1 = 'file' + str(idx)
save_path = 'data/tree' + str(idx) + '.json'
_kmeans_tree['festival'][key1] = save_path
_file.op_file(file_path=save_path,
data=class_data[key1],
model='json',
method='save')
# 保存后删除
del class_data[key1]
# 存储根节点查找文件
_file.op_file(file_path='data/root.json',
data=_kmeans_tree,
model='json',
method='save')
def similirity(self,
data,
file_path,
branchs=[2, 2],
candidate=3,
distance=15):
self._max_dist = distance
self._search_branch = branchs
self._search_result = []
result = _file.op_file(file_path, model='json', method='read')
self.search_tree(data, result)
sr = [
c for c in sorted(
self._search_result, key=lambda k: k[0], reverse=False)
]
save_len = candidate if candidate < len(sr) else len(sr)
return sr[0:save_len]