def search_specified_group(postname="报"):
query_body = {
"query":{
"bool": {
"must": [
{"wildcard": {
"uname": {
"wildcard": "*" + postname
}
}},
{"range": {
"fansnum": {
"gte": 100000
}
}}
]
}
},
"size": 10000
}
search_results = es_profile.search(index="user_portrait_1222", doc_type="user", body=query_body)["hits"]["hits"]
uid_list = []
for item in search_results:
uid_list.append(item['_id'])
print item['_id'], item['_source']['uname'], '\n'
print "该群体有：", len(uid_list)
return uid_list
def nickname2uid(nickname_list):
uids_list = set()
query_body = {
'query': {
'filtered': {
'filter': {
'terms': {
'nick_name': nickname_list
}
}
}
},
'size': MAX_SEARCH_SIZE
}
es_results = es_user_profile.search(index=profile_index_name,doc_type=profile_index_type,\
body=query_body)['hits']['hits']
#print 'es_results:::',es_results
if es_results:
for result in es_results:
result = result['_source']
uid = result['uid']
uids_list.add(uid)
uids_list = list(uids_list)
#print 'uids_list::',uids_list
return uids_list
def get_fansnum_max():
query_body = {
'query':{
'match_all':{}
},
'size': 1,
'sort': [{'fansnum': {'order': 'desc'}}]
}
try:
fansnum_max_results = es_user_profile.search(index=profile_index_name, doc_type=profile_index_type, body=query_body)['hits']['hits']
except Exception, e:
raise e
def sort_total_number(prefix, uid_list, time, key_search, number):
if prefix == "weibo_":
if int(time) == 1:
order = prefix + 'day_last'
elif int(time) == 7:
order = prefix + 'week_sum'
elif int(time) == 30:
order = prefix + 'month_sum'
else:
order = prefix
if uid_list:
query_body = {
"query": {
"filtered": {
"filter": {
"terms": {
"uid": uid_list
}
}
}
},
"sort": {
order: {
"order": "desc"
}
},
"size": number
}
else:
query_body = {
"query": {
"match_all": {}
},
"sort": {
order: {
"order": "desc"
}
},
"size": number
}
search_results = es_user_profile.search(index="bci_history",
doc_type="bci",
body=query_body,
_source=False)['hits']['hits']
uid_list = []
if search_results:
for item in search_results:
uid_list.append(item['_id'])
return uid_list
def search_specified_group(postname="律师"):
query_body = {
"query": {
"bool": {
"must": [{
"wildcard": {
"nick_name": {
"wildcard": "*" + postname
}
}
}, {
"range": {
"fansnum": {
"gte": 10000
}
}
}]
}
},
"size": 10000
}
search_results = es_profile.search(index="weibo_user",
doc_type="user",
body=query_body)["hits"]["hits"]
uid_list = []
for item in search_results:
uid_list.append(item['_id'])
print item['_id'], item['_source']['nick_name'], '\n'
print "该群体有：", len(uid_list)
result = dict()
result['social_sensors'] = json.dumps(uid_list)
result['stop_time'] = 1460086441
result['create_at'] = 1377964800
result['task_name'] = "top lawers"
result['remark'] = "粉丝10000以上的律师"
result["history_status"] = json.dumps([])
result['burst_reason'] = ''
result['processing_status'] = "1"
result["warning_status"] = '0'
result["finish"] = "0"
result["create_by"] = 'admin'
es_profile.index(index="manage_sensing_task",
doc_type='task',
id='admin-top lawers',
body=result)
print "1"
return uid_list
def get_evaluate_max(index_name,index_type,field):
query_body = {
'query':{
'match_all':{}
},
'size':1,
'sort':[{field: {'order': 'desc'}}]
}
try:
result = es_user_profile.search(index=index_name, doc_type=index_type, body=query_body)['hits']['hits']
max_evaluate = result[0]['_source'][field]
except Exception, e:
raise e
max_evaluate = 1
def search_specified_group(postname="律师"):
query_body = {
"query":{
"bool": {
"must": [
{"wildcard": {
"nick_name": {
"wildcard": "*" + postname
}
}},
{"range": {
"fansnum": {
"gte": 10000
}
}}
]
}
},
"size": 10000
}
search_results = es_profile.search(index="weibo_user", doc_type="user", body=query_body)["hits"]["hits"]
uid_list = []
for item in search_results:
uid_list.append(item['_id'])
print item['_id'], item['_source']['nick_name'], '\n'
print "该群体有：", len(uid_list)
result = dict()
result['social_sensors'] = json.dumps(uid_list)
result['stop_time'] = 1460086441
result['create_at'] = 1377964800
result['task_name'] = "top lawers"
result['remark'] = "粉丝10000以上的律师"
result["history_status"] = json.dumps([])
result['burst_reason'] = ''
result['processing_status'] = "1"
result["warning_status"] = '0'
result["finish"] = "0"
result["create_by"] = 'admin'
es_profile.index(index="manage_sensing_task", doc_type='task', id='admin-top lawers', body=result)
print "1"
return uid_list
def sort_total_number(prefix, uid_list, time, key_search, number):
if prefix == "weibo_":
if int(time) == 1:
order = prefix + 'day_last'
elif int(time) == 7:
order = prefix + 'week_sum'
elif int(time) == 30:
order = prefix + 'month_sum'
else:
order = prefix
if uid_list:
query_body = {
"query":{
"filtered": {
"filter": {
"terms":{
"uid": uid_list
}
}
}
},
"sort":{order: {"order": "desc"}},
"size": number
}
else:
query_body = {
"query":{
"match_all": {}
},
"sort": { order: {"order": "desc"} },
"size": number
}
search_results = es_user_profile.search(index="bci_history", doc_type="bci", body=query_body, _source=False)['hits']['hits']
uid_list = []
if search_results:
for item in search_results:
uid_list.append(item['_id'])
return uid_list
def count_minweibouser_influence(index_name):
query_body = {
'query': {
'match_all': {}
},
'size': 1,
'sort': {
'user_index': {
'order': 'asc'
}
}
}
try:
max_result = es_user_profile.search(index=index_name,
doc_type=weibo_bci_index_type,
body=query_body)['hits']['hits']
for item in max_result:
max_user_index = item['_source']['user_index']
except:
max_user_index = 0
return max_user_index
def lookup_active_weibouser(today_date_time):
weibo_active_user_index_name = weibo_active_user_index_name_pre + ts2datetime(
today_date_time)
weibo_active_user_mappings(weibo_active_user_index_name)
bci_index_name = weibo_bci_index_name_pre + ''.join(
ts2datetime(today_date_time).split('-'))
# userlist = lookup_weiboxnr_concernedusers(weiboxnr_id)
user_max_index = count_maxweibouser_influence(bci_index_name)
user_min_index = count_minweibouser_influence(bci_index_name)
results = []
query_body = {
'query': {
'match_all': {}
},
'size': 100, #查询影响力排名前50的用户即可
'sort': {
'user_index': {
'order': 'desc'
}
}
}
try:
flow_text_exist=es_user_portrait.search(index=bci_index_name,\
doc_type=weibo_bci_index_type,body=query_body)['hits']['hits']
search_uid_list = [item['_source']['user'] for item in flow_text_exist]
print len(search_uid_list)
weibo_user_exist = es_user_profile.search(index=profile_index_name,\
doc_type=profile_index_type,body={'query':{'terms':{'uid':search_uid_list}}})['hits']['hits']
weibo_user_dict = dict()
#user_dict = dict()
for item_i in weibo_user_exist:
uid = item_i['_source']['uid']
weibo_user_dict[uid] = item_i['_source']
for item in flow_text_exist:
user_dict = dict()
#print 'item:', item['_source']
user_dict['influence'] = (item['_source']['user_index'] -
user_min_index) / (user_max_index -
user_min_index)
user_dict['fans_num'] = item['_source']['user_fansnum']
user_dict['friends_num'] = item['_source']['user_friendsnum']
user_dict['total_number'] = item['_source']['total_number']
user_dict['uid'] = item['_source']['user']
try:
uid = user_dict['uid']
weibo_user_info = weibo_user_dict[uid]
user_dict['uname'] = weibo_user_info['nick_name']
user_dict['location'] = weibo_user_info['user_location']
user_dict['url'] = weibo_user_info['photo_url']
except:
user_dict['uname'] = ''
user_dict['location'] = ''
user_dict['url'] = ''
#es_xnr.index(index_name= weibo_active_user_index_name,doc_type= weibo_active_user_index_type,body=user_dict,id=user_dict['uid'])
results.append(user_dict)
except:
results = []
print len(results)
return results
