def search_mention(uid, sensitive):
date = ts2datetime(time.time()).replace('-','')
stat_results = dict()
results = dict()
test_ts = time.time()
test_ts = datetime2ts('2013-09-07')
for i in range(0,7):
ts = test_ts -i*24*3600
date = ts2datetime(ts).replace('-', '')
if not sensitive:
at_temp = r_cluster.hget('at_' + str(date), str(uid))
else:
at_temp = r_cluster.hget('sensitive_at_' + str(date), str(uid))
if not at_temp:
continue
else:
result_dict = json.loads(at_temp)
for at_uid in result_dict:
if stat_results.has_key(at_uid):
stat_results[uid] += result_dict[at_uid]
else:
stat_results[uid] = result_dict[at_uid]
if not stat_results:
return [None, 0]
in_status = identify_uid_list_in(result_dict.keys())
for at_uid in result_dict:
if at_uid in in_status:
results[at_uid] = [result_dict[at_uid], '1']
else:
results[at_uid] = [result_dict[at_uid], '0']
sorted_results = sorted(results.items(), key=lambda x:x[1][0], reverse=True)
return [sorted_results[0:20], len(results)]
def get_sensitive_user_detail(uid_list, date, sensitive):
es_cluster = es_user_profile
ts = datetime2ts(date)
results = []
index_name = pre_influence_index + str(date).replace(
'-', '') # index_name:20130901
user_bci_results = es_bci.mget(index=index_name,
doc_type='bci',
body={'ids': uid_list},
_source=False,
fields=['user_index'])['docs']
user_profile_results = es_user_profile.mget(index="weibo_user",
doc_type="user",
body={"ids": uid_list},
_source=True)['docs']
top_influnce_value = get_top_value("user_index", es_bci, index_name, "bci")
for i in range(0, len(uid_list)):
personal_info = [''] * 6
uid = uid_list[i]
personal_info[0] = uid_list[i]
personal_info[1] = uid_list[i]
if user_profile_results[i]['found']:
profile_dict = user_profile_results[i]['_source']
uname = profile_dict['nick_name']
if uname:
personal_info[1] = uname
personal_info[2] = profile_dict['user_location']
personal_info[3] = profile_dict['fansnum']
personal_info[4] = profile_dict['statusnum']
if user_bci_results[i]['found']:
try:
tmp_bci = user_bci_results[i]['fields']['user_index'][0]
influence = math.log(
tmp_bci / float(top_influnce_value) * 9 + 1, 10) * 100
personal_info[5] = influence
except:
personal_info[5] = 0
else:
personal_info[5] = 0
if sensitive:
sensitive_words = redis_cluster.hget('sensitive_' + str(ts),
str(uid))
if sensitive_words:
sensitive_dict = json.loads(sensitive_words)
personal_info.append(sensitive_dict.keys())
else:
personal_info.append([])
else:
personal_info.append([])
results.append(personal_info)
return results
def search_mention(uid, sensitive):
date = ts2datetime(time.time()).replace('-', '')
stat_results = dict()
results = dict()
test_ts = time.time()
test_ts = datetime2ts('2013-09-07')
for i in range(0, 7):
ts = test_ts - i * 24 * 3600
date = ts2datetime(ts).replace('-', '')
if not sensitive:
at_temp = r_cluster.hget('at_' + str(date), str(uid))
else:
at_temp = r_cluster.hget('sensitive_at_' + str(date), str(uid))
if not at_temp:
continue
else:
result_dict = json.loads(at_temp)
for at_uid in result_dict:
if stat_results.has_key(at_uid):
stat_results[uid] += result_dict[at_uid]
else:
stat_results[uid] = result_dict[at_uid]
if not stat_results:
return [None, 0]
in_status = identify_uid_list_in(result_dict.keys())
for at_uid in result_dict:
if at_uid in in_status:
results[at_uid] = [result_dict[at_uid], '1']
else:
results[at_uid] = [result_dict[at_uid], '0']
sorted_results = sorted(results.items(),
key=lambda x: x[1][0],
reverse=True)
return [sorted_results[0:20], len(results)]
def get_sensitive_user_detail(uid_list, date, sensitive):
es_cluster = es_user_profile
ts = datetime2ts(date)
results = []
index_name = pre_influence_index + str(date).replace('-','') # index_name:20130901
user_bci_results = es_bci.mget(index=index_name, doc_type='bci', body={'ids':uid_list}, _source=False, fields=['user_index'])['docs']
user_profile_results = es_user_profile.mget(index="weibo_user", doc_type="user", body={"ids":uid_list}, _source=True)['docs']
top_influnce_value = get_top_value("user_index", es_bci, index_name, "bci")
for i in range(0, len(uid_list)):
personal_info = ['']*6
uid = uid_list[i]
personal_info[0] = uid_list[i]
personal_info[1] = uid_list[i]
if user_profile_results[i]['found']:
profile_dict = user_profile_results[i]['_source']
uname = profile_dict['nick_name']
if uname:
personal_info[1] = uname
personal_info[2] = profile_dict['user_location']
personal_info[3] = profile_dict['fansnum']
personal_info[4] = profile_dict['statusnum']
if user_bci_results[i]['found']:
try:
tmp_bci = user_bci_results[i]['fields']['user_index'][0]
influence = math.log(tmp_bci/float(top_influnce_value)*9+1, 10)*100
personal_info[5] = influence
except:
personal_info[5] = 0
else:
personal_info[5] = 0
if sensitive:
sensitive_words = redis_cluster.hget('sensitive_' + str(ts), str(uid))
if sensitive_words:
sensitive_dict = json.loads(sensitive_words)
personal_info.append(sensitive_dict.keys())
else:
personal_info.append([])
else:
personal_info.append([])
results.append(personal_info)
return results