2022-11-17 19:53:58 +00:00
|
|
|
import json
|
2023-12-19 12:18:58 +00:00
|
|
|
import logging
|
2024-01-29 00:27:30 +00:00
|
|
|
import os
|
2023-12-17 20:30:20 +00:00
|
|
|
|
2024-01-29 00:27:30 +00:00
|
|
|
from elasticsearch import Elasticsearch
|
2023-12-17 20:30:20 +00:00
|
|
|
|
2024-01-29 01:09:54 +00:00
|
|
|
from services.rediscache import redis
|
2022-10-04 00:32:29 +00:00
|
|
|
|
|
|
|
|
2024-01-29 00:27:30 +00:00
|
|
|
logger = logging.getLogger('[services.search] ')
|
|
|
|
logger.setLevel(logging.DEBUG)
|
|
|
|
|
|
|
|
ELASTIC_HOST = os.environ.get('ELASTIC_HOST', 'localhost').replace('https://', '').replace('http://', '')
|
|
|
|
ELASTIC_USER = os.environ.get('ELASTIC_USER', '')
|
|
|
|
ELASTIC_PASSWORD = os.environ.get('ELASTIC_PASSWORD', '')
|
|
|
|
ELASTIC_PORT = os.environ.get('ELASTIC_PORT', 9200)
|
|
|
|
ELASTIC_AUTH = f'{ELASTIC_USER}:{ELASTIC_PASSWORD}' if ELASTIC_USER else ''
|
2024-01-29 01:09:54 +00:00
|
|
|
ELASTIC_URL = os.environ.get('ELASTIC_URL', f'https://{ELASTIC_AUTH}@{ELASTIC_HOST}:{ELASTIC_PORT}')
|
|
|
|
ELASTIC_REINDEX = os.environ.get('ELASTIC_REINDEX', '')
|
|
|
|
REDIS_TTL = 86400 # 1 day in seconds
|
2024-01-29 00:27:30 +00:00
|
|
|
|
|
|
|
|
2024-01-29 01:09:54 +00:00
|
|
|
class SearchService:
|
2024-01-29 01:41:46 +00:00
|
|
|
async def __init__(self, index_name='search_index'):
|
|
|
|
logging.info('Initializing SearchService')
|
2024-01-29 00:27:30 +00:00
|
|
|
self.index_name = index_name
|
2024-01-29 01:21:28 +00:00
|
|
|
self.elasticsearch_client = Elasticsearch(f'{ELASTIC_URL}', verify_certs=False)
|
2024-01-29 00:27:30 +00:00
|
|
|
self.check_index()
|
|
|
|
|
2024-01-29 01:09:54 +00:00
|
|
|
if ELASTIC_REINDEX:
|
|
|
|
self.recreate_index()
|
|
|
|
|
2024-01-29 01:41:46 +00:00
|
|
|
def info(self):
|
|
|
|
logging.info(f'Initializing SearchService {self.elasticsearch_client}')
|
|
|
|
|
2024-01-29 00:27:30 +00:00
|
|
|
def delete_index(self):
|
|
|
|
self.elasticsearch_client.indices.delete(index=self.index_name, ignore_unavailable=True)
|
|
|
|
|
|
|
|
def create_index(self):
|
|
|
|
index_settings = {
|
|
|
|
'settings': {
|
|
|
|
'index': {
|
|
|
|
'number_of_shards': 1,
|
|
|
|
'auto_expand_replicas': '0-all',
|
|
|
|
},
|
|
|
|
'analysis': {
|
|
|
|
'analyzer': {
|
|
|
|
'ru': {
|
|
|
|
'tokenizer': 'standard',
|
|
|
|
'filter': ['lowercase', 'ru_stop', 'ru_stemmer'],
|
|
|
|
}
|
|
|
|
},
|
|
|
|
'filter': {
|
|
|
|
'ru_stemmer': {
|
|
|
|
'type': 'stemmer',
|
|
|
|
'language': 'russian',
|
|
|
|
},
|
|
|
|
'ru_stop': {
|
|
|
|
'type': 'stop',
|
|
|
|
'stopwords': '_russian_',
|
|
|
|
},
|
|
|
|
},
|
|
|
|
},
|
|
|
|
},
|
|
|
|
'mappings': {
|
|
|
|
'properties': {
|
2024-01-29 01:09:54 +00:00
|
|
|
'body': {'type': 'text', 'analyzer': 'ru'},
|
2024-01-29 00:27:30 +00:00
|
|
|
'text': {'type': 'text'},
|
|
|
|
'author': {'type': 'text'},
|
|
|
|
}
|
|
|
|
},
|
|
|
|
}
|
|
|
|
|
|
|
|
self.elasticsearch_client.indices.create(index=self.index_name, body=index_settings)
|
|
|
|
self.elasticsearch_client.indices.close(index=self.index_name)
|
|
|
|
self.elasticsearch_client.indices.open(index=self.index_name)
|
|
|
|
|
|
|
|
def put_mapping(self):
|
|
|
|
mapping = {
|
|
|
|
'properties': {
|
2024-01-29 01:09:54 +00:00
|
|
|
'body': {'type': 'text', 'analyzer': 'ru'},
|
2024-01-29 00:27:30 +00:00
|
|
|
'text': {'type': 'text'},
|
|
|
|
'author': {'type': 'text'},
|
|
|
|
}
|
|
|
|
}
|
|
|
|
|
|
|
|
self.elasticsearch_client.indices.put_mapping(index=self.index_name, body=mapping)
|
|
|
|
|
|
|
|
def check_index(self):
|
|
|
|
if not self.elasticsearch_client.indices.exists(index=self.index_name):
|
|
|
|
logger.debug(f'Creating {self.index_name} index')
|
|
|
|
self.create_index()
|
|
|
|
self.put_mapping()
|
|
|
|
|
2024-01-29 01:09:54 +00:00
|
|
|
def recreate_index(self):
|
|
|
|
self.delete_index()
|
|
|
|
self.check_index()
|
|
|
|
|
2024-01-29 00:27:30 +00:00
|
|
|
def index_post(self, shout):
|
|
|
|
id_ = str(shout.id)
|
|
|
|
logger.debug(f'Indexing post id {id_}')
|
|
|
|
self.elasticsearch_client.index(index=self.index_name, id=id_, body=shout)
|
|
|
|
|
|
|
|
def search_post(self, query, limit, offset):
|
2024-01-29 01:09:54 +00:00
|
|
|
logger.debug(f'query: {query}')
|
2024-01-29 00:27:30 +00:00
|
|
|
search_body = {
|
2024-01-29 01:09:54 +00:00
|
|
|
'query': {'match': {'_all': query}},
|
2024-01-29 00:27:30 +00:00
|
|
|
}
|
|
|
|
|
|
|
|
search_response = self.elasticsearch_client.search(
|
|
|
|
index=self.index_name, body=search_body, size=limit, from_=offset
|
|
|
|
)
|
|
|
|
hits = search_response['hits']['hits']
|
|
|
|
|
|
|
|
return [
|
|
|
|
{
|
|
|
|
**hit['_source'],
|
|
|
|
'score': hit['_score'],
|
|
|
|
}
|
|
|
|
for hit in hits
|
|
|
|
]
|
|
|
|
|
2024-01-29 01:41:46 +00:00
|
|
|
|
|
|
|
search = SearchService()
|
|
|
|
|
|
|
|
|
|
|
|
async def search_text(text: str, limit: int = 50, offset: int = 0):
|
|
|
|
payload = []
|
|
|
|
try:
|
|
|
|
# Use a key with a prefix to differentiate search results from other Redis data
|
|
|
|
redis_key = f'search:{text}'
|
|
|
|
# Use OpenSearchService.search_post method
|
|
|
|
payload = search.search_post(text, limit, offset)
|
|
|
|
# Use Redis as cache with TTL
|
|
|
|
await redis.execute('SETEX', redis_key, REDIS_TTL, json.dumps(payload))
|
|
|
|
except Exception as e:
|
|
|
|
logging.error(f'Error during search: {e}')
|
|
|
|
return payload
|