def update_standard_tags():
stdtag = StandardTags()
total = db.tags.find().count()
logging.debug("remaining %d tags without standard." % total)
for i, t in enumerate(db.tags.find(timeout=False)): # {"$where":"this.standard == null"}, timeout=False
taglst = stdtag.transform(t["name"])
if not taglst or taglst[0][0] == t["name"]:
logging.debug("skip tag: %s", t["name"])
t["standard"] = []
else:
t["standard"] = taglst
logging.debug("%d, tag update standard %s --> %s" % (i, t["name"], taglst[0][0]))
ret = db.tags.update({"_id": t["_id"]}, t)
if not ret["ok"]:
logging.warn("tag update failed. tag: %s" % t["name"])
prog_d("tag update", i, total)
def update_standard_tags():
stdtag = StandardTags()
total = db.tags.find().count()
logging.debug('remaining %d tags without standard.' % total)
for i, t in enumerate(
db.tags.find(timeout=False)
): # {"$where":"this.standard == null"}, timeout=False
taglst = stdtag.transform(t['name'])
if not taglst or taglst[0][0] == t['name']:
logging.debug('skip tag: %s', t['name'])
t['standard'] = []
else:
t['standard'] = taglst
logging.debug('%d, tag update standard %s --> %s' %
(i, t['name'], taglst[0][0]))
ret = db.tags.update({"_id": t['_id']}, t)
if not ret['ok']:
logging.warn('tag update failed. tag: %s' % t['name'])
prog_d('tag update', i, total)
# prog_d('solve Matrix row', c, total)
# c += 1
return MImatrix
def _calMIvalue(self, a_set, b_set):
pab = math.fabs(float(len(a_set&b_set)) / self.root)
pa = math.fabs(float(len(a_set)) / self.root)
pb = math.fabs(float(len(b_set)) / self.root)
Iab = pab * math.log((pab+1) / (pa*pb))
Ha = -pa * math.log(pa)
Hb = -pb * math.log(pb)
return float(Iab) / (float(Ha + Hb) / 2)
rsdb = RecsysDatabase()
stdtag = StandardTags()
# set PROG before using this function
def prog_d(dstr, line=-1, total=100):
global PROG_REC
if line >= 0:
progress = int(float(line)/float(total) * 100 + 1)
if progress not in PROG_SCALE or progress == PROG_REC:
return
# print progress
PROG_REC = progress
dstr += ' %d%%(%d/%d) -=-=-' % (progress, line, total)
logging.info('-=-=- Processing ' + dstr)
else:
logging.info('-=-=- Finishing ' + dstr)
# -*- coding: utf-8 -*-
from book_recsys import *
from stdtag import StandardTags
# word2vec = pickle.load(open('dump/Word2VecMtrx.dmp'))
stdtag = StandardTags()
stdtag._loadStart()
def _accWeight(node_dict, level, fno, a=0.0):
up_lev = {}
for nd in node_dict.items():
p_mat = level[fno][2]
for upnd in p_mat[nd[0]].items(): # 遍历连接着的上层节点
if upnd[1] > 0:
if upnd[0] not in up_lev:
up_lev[upnd[0]] = 0.0
up_lev[upnd[0]] += nd[1]*upnd[1]*(1+a) # 权重叠加
return up_lev
def findLevel(tag, level):
for i, lev in enumerate(level):
if tag in lev[0]:
return i
def getWacWeight(tag, level, iter_level=-1):
#print '\r\n\r\n\r\n\r\n', tag
if iter_level == -1:
iter_level = len(level)-1
ret = {}
num = findLevel(tag, level)
# -*- coding: utf-8 -*-
from book_recsys import *
from stdtag import StandardTags
# word2vec = pickle.load(open('dump/Word2VecMtrx.dmp'))
stdtag = StandardTags()
stdtag._loadStart()
def _accWeight(node_dict, level, fno, a=0.0):
up_lev = {}
for nd in node_dict.items():
p_mat = level[fno][2]
for upnd in p_mat[nd[0]].items(): # 遍历连接着的上层节点
if upnd[1] > 0:
if upnd[0] not in up_lev:
up_lev[upnd[0]] = 0.0
up_lev[upnd[0]] += nd[1] * upnd[1] * (1 + a) # 权重叠加
return up_lev
def findLevel(tag, level):
for i, lev in enumerate(level):
if tag in lev[0]:
return i
def getWacWeight(tag, level, iter_level=-1):
#print '\r\n\r\n\r\n\r\n', tag
if iter_level == -1:
iter_level = len(level) - 1