Skip to content

S
search_tips
Commit 3b6b6fbb authored by lixiaofang's avatar lixiaofang
Browse files
Options

add

parent 6854ba01
Show whitespace changes
Inline Side-by-side
Showing with 112 additions and 155 deletions
search/utils/auto_tips.py
View file @ 3b6b6fbb
...@@ -17,79 +17,114 @@ from pypinyin import pinyin, lazy_pinyin ...@@ -17,79 +17,114 @@ from pypinyin import pinyin, lazy_pinyin
def get_suggest_tips(query, lat, lng, offset=0, size=50): def get_suggest_tips(query, lat, lng, offset=0, size=50):
try: try:
###先获取完全命中的query have_read_tips_set, ret_list, result_dict = get_query_by_es(query=str(query).lower(), lat=lat, lng=lng,
# qq = { offset=offset,
# "query": { size=size)
# "bool": { if len(ret_list) >= 50:
# "must": [{ return ret_list
# "term": {
# "ori_name": query else:
# } query_ret_list = []
# }, { wordresemble_ret_list = []
# "term": { value_data = []
# "is_online": True QUERY_KEY = "query:search_tip"
# } query_base64 = base64.b64encode(query.encode('utf8')).decode('utf8')
# }] if redis_client.hget(QUERY_KEY, query_base64) is not None:
# } value_data = json.loads(str(redis_client.hget(QUERY_KEY, query_base64), encoding='utf-8'))
# }
# } if len(value_data) > 0:
# logging.info("get qqqqqqqqq:%s" % qq) for i in value_data:
# whole_query = ESPerform.get_search_results(ESPerform.get_cli(), sub_index_name="suggest", query_body=qq, key = list(i.keys())[0]
# offset=0, size=1, is_suggest_request=True) ori_name = str(base64.b64decode(key), "utf-8")
# if ori_name not in have_read_tips_set:
# logging.info("get whole_query:%s" % whole_query) have_read_tips_set.add(ori_name)
result_num = i.get(key, 0)
describe = "约" + str(result_num) + "个结果" if result_num else ""
logging.info("get result_num:%s" % result_num)
highlight_marks = u'<ems>%s</ems>' % query
# highlight_name = ori_name.replace(query, highlight_marks)
highlight_name = set_highlihgt(query, ori_name)
if ori_name == query:
query_ret_list.append(
{"results_num": result_num, "ori_name": ori_name, "id": None, "is_online": True,
"offline_score": 0,
"type_flag": get_tips_word_type(ori_name), "highlight_name": highlight_name,
"describe": describe})
else:
wordresemble_ret_list.append(
{"results_num": result_num, "ori_name": ori_name, "id": None, "is_online": True,
"offline_score": 0,
"type_flag": get_tips_word_type(ori_name), "highlight_name": highlight_name,
"describe": describe})
ret_list.extend(query_ret_list)
ret_list.extend(wordresemble_ret_list)
###繁体字删掉，把搜索结果加到简体字上边
fanti_query = [{'痩脸针': "瘦脸针"}]
for item in ret_list:
result_num = [[item['results_num'], list(ret.values())[0], list(ret.keys())[0]] for ret in fanti_query
if list(ret.keys())[0] == item['ori_name']]
if len(result_num) > 0:
ret_list.remove(item)
for item in ret_list:
if item['ori_name'] == result_num[0][1]:
item['results_num'] += result_num[0][0]
####
if len(ret_list) >= 50:
return ret_list[0:50]
elif len(ret_list) < 50 and len(ret_list) > 0:
return ret_list
else:
##无结果的时候把汉字转成拼音再搜一次
ss = lazy_pinyin(query)
str_query = ''
for item in ss:
str_query += str(item)
have_read_tips_set, ret_list, result_dict = get_query_by_es(query=str_query, lat=lat, lng=lng,
offset=offset,
size=size)
logging.info("get ret_list:%s" % ret_list)
return ret_list
except:
logging.error("catch exception,err_msg:%s" % traceback.format_exc())
return list()
def set_highlihgt(query=None, ori_name=None):
###高亮调整
all_word = set()
query2 = ori_name
for item in range(0, len(query)):
all_word.add(query[item])
for item in all_word:
is_find = query2.find(item)
high_query = None
if is_find >= 0:
highlight_marks = u'<>%s</>' % item
high_query = query2.replace(item, highlight_marks)
query2 = high_query
if high_query == None:
highlight_name = query2
else:
highlight_name = high_query
highlight_name = query2.replace('<>', '<ems>').replace('</>', '</ems>')
# hits = whole_query["hits"]["hits"] return highlight_name
#
# for item in hits:
# have_read_tips_set.add(item["_source"]["ori_name"]) def get_query_by_es(query='', lat=0, lng=0, size=0, offset=0):
# id = item["_source"]["id"] try:
# ori_name = item["_source"]["ori_name"]
# results_num = item["_source"]["results_num"]
# is_online = item["_source"]["is_online"]
# offline_score = item["_source"]["offline_score"]
# type_flag = item["_source"]["type_flag"]
#
# highlight_marks = u'<ems>%s</ems>' % query
# item["_source"]["highlight_name"] = item["_source"]["ori_name"].replace(query, highlight_marks)
# highlight_name = item["_source"]["highlight_name"]
# if item["_source"]["type_flag"] == "hospital":
# if lat is not None and lng is not None and lat != 0.0 and lng != 0.0:
#
# logging.info("get g_hospital_pos_dict:%s" % g_hospital_pos_dict)
# if item["_source"]["ori_name"] in g_hospital_pos_dict:
# distance = point_distance(lng, lat,
# g_hospital_pos_dict[item["_source"]["ori_name"]][0],
# g_hospital_pos_dict[item["_source"]["ori_name"]][1])
# if distance < 1000 * 50:
# if distance < 1000:
# if distance < 100:
# item["_source"]["describe"] = "<100" + "米"
# else:
# item["_source"]["describe"] = "约" + str(int(distance)) + "米"
# else:
# item["_source"]["describe"] = "约" + str(
# round(1.0 * distance / 1000, 1)) + "km"
# else:
# item["_source"]["describe"] = ">50km"
# else:
# item["_source"]["describe"] = ""
#
# else:
# item["_source"]["describe"] = ""
# else:
# if item["_source"]["type_flag"] == "doctor":
# item["_source"]["describe"] = ""
# else:
# item["_source"]["describe"] = "约" + str(item["_source"]["results_num"]) + "个结果" if \
# item["_source"]["results_num"] else ""
#
# ret_list.append({"offline_score": offline_score, "ori_name": ori_name, "results_num": results_num, "id": id,
# "highlight_name": highlight_name, "type_flag": type_flag,
# "is_online": is_online})
###获取联想到的数据
# ios输入法在某些情况下会携带\\u2006
query = query.replace("\u2006", '') query = query.replace("\u2006", '')
q = { q = {
"suggest": { "suggest": {
...@@ -178,88 +213,7 @@ def get_suggest_tips(query, lat, lng, offset=0, size=50): ...@@ -178,88 +213,7 @@ def get_suggest_tips(query, lat, lng, offset=0, size=50):
ret_list.extend(get_tag_wiki_data) ret_list.extend(get_tag_wiki_data)
ret_list.extend(get_doctor_hospital_data) ret_list.extend(get_doctor_hospital_data)
if len(result_dict["suggest"]["tips-suggest"]) >= 50: return have_read_tips_set, ret_list, result_dict
return ret_list
else:
query_ret_list = []
wordresemble_ret_list = []
value_data = []
QUERY_KEY = "query:search_tip"
query_base64 = base64.b64encode(query.encode('utf8')).decode('utf8')
if redis_client.hget(QUERY_KEY, query_base64) is not None:
value_data = json.loads(str(redis_client.hget(QUERY_KEY, query_base64), encoding='utf-8'))
if len(value_data) > 0:
for i in value_data:
key = list(i.keys())[0]
ori_name = str(base64.b64decode(key), "utf-8")
if ori_name not in have_read_tips_set:
have_read_tips_set.add(ori_name)
result_num = i.get(key, 0)
describe = "约" + str(result_num) + "个结果" if result_num else ""
logging.info("get result_num:%s" % result_num)
highlight_marks = u'<ems>%s</ems>' % query
# highlight_name = ori_name.replace(query, highlight_marks)
highlight_name = set_highlihgt(query, ori_name)
if ori_name == query:
query_ret_list.append(
{"results_num": result_num, "ori_name": ori_name, "id": None, "is_online": True,
"offline_score": 0,
"type_flag": get_tips_word_type(ori_name), "highlight_name": highlight_name,
"describe": describe})
else:
wordresemble_ret_list.append(
{"results_num": result_num, "ori_name": ori_name, "id": None, "is_online": True,
"offline_score": 0,
"type_flag": get_tips_word_type(ori_name), "highlight_name": highlight_name,
"describe": describe})
ret_list.extend(query_ret_list)
ret_list.extend(wordresemble_ret_list)
###繁体字删掉，把搜索结果加到简体字上边
fanti_query = [{'痩脸针': "瘦脸针"}]
for item in ret_list:
result_num = [[item['results_num'], list(ret.values())[0], list(ret.keys())[0]] for ret in fanti_query
if list(ret.keys())[0] == item['ori_name']]
if len(result_num) > 0:
ret_list.remove(item)
for item in ret_list:
if item['ori_name'] == result_num[0][1]:
item['results_num'] += result_num[0][0]
####
if len(ret_list) >= 50:
return ret_list[0:50]
else:
return ret_list
except: except:
logging.error("catch exception,err_msg:%s" % traceback.format_exc())
return list()
return set(), list(), list()
def set_highlihgt(query=None, ori_name=None):
###高亮调整
all_word = set()
query2 = ori_name
for item in range(0, len(query)):
all_word.add(query[item])
for item in all_word:
is_find = query2.find(item)
high_query = None
if is_find >= 0:
highlight_marks = u'<ems>%s</ems>' % item
high_query = query2.replace(item, highlight_marks)
query2 = high_query
if high_query == None:
highlight_name = query2
else:
highlight_name = high_query
return highlight_name
trans2es/commons/commons.py
View file @ 3b6b6fbb
...@@ -17,8 +17,10 @@ def uuid4(): ...@@ -17,8 +17,10 @@ def uuid4():
""" """
return uuid.uuid4().hex return uuid.uuid4().hex
def get_tips_suggest_list(instance_cn_name): def get_tips_suggest_list(instance_cn_name):
try: try:
print(instance_cn_name)
# ch_full_weight = 6.0 * 1000 # ch_full_weight = 6.0 * 1000
# py_full_weight = 3.0 * 1000 # py_full_weight = 3.0 * 1000
full_weight = 3.0 * 1000 full_weight = 3.0 * 1000
......
trans2es/utils/itemwiki_transfer.py
View file @ 3b6b6fbb
...@@ -97,8 +97,9 @@ class ItemWikiTransfer(object): ...@@ -97,8 +97,9 @@ class ItemWikiTransfer(object):
ret_list.append(item_dict) ret_list.append(item_dict)
suggest_list = get_tips_suggest_list(instance.name) suggest_list = get_tips_suggest_list(str(instance.name).lower())
logging.info("get suggest_list:%s" % suggest_list)
return (item_dict, suggest_list) return (item_dict, suggest_list)
except: except:
logging.error("catch exception,err_msg:%s" % traceback.format_exc()) logging.error("catch exception,err_msg:%s" % traceback.format_exc())
......
Markdown is supported
0% or
You are about to add 0 people to the discussion. Proceed with caution.
Finish editing this message first!
Please register or to comment