simon987/od-database
1
0
Fork 0
mirror of https://github.com/simon987/od-database.git synced 2025-04-19 10:26:44 +00:00
Code Issues Projects Releases Wiki Activity

Fixed indentation error

Browse Source
This commit is contained in:
Simon 2018-06-28 22:47:14 -04:00
parent 4c9d79fdbf
commit 9ff21e7943
1 changed files with 30 additions and 30 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

60
search/search.py
View File

@ -406,39 +406,39 @@ class ElasticSearchEngine(SearchEngine):
return stats return stats
def stream_all_docs(self): def stream_all_docs(self):
return helpers.scan(query={ return helpers.scan(query={
"query": { "query": {
"match_all": {} "match_all": {}
} }
}, scroll="5m", client=self.es, index=self.index_name) }, scroll="5m", client=self.es, index=self.index_name)
def are_empty(self, websites): def are_empty(self, websites):
result = self.es.search(body={ result = self.es.search(body={
"query": { "query": {
"bool": { "bool": {
"filter": { "filter": {
"terms": {
"website_id": websites
},
}
}
},
"aggs": {
"websites": {
"terms": { "terms": {
"website_id": websites "field": "website_id",
}, "size": 100000,
"min_doc_count": 1
}
} }
} },
}, "size": 0
"aggs": { }, index=self.index_name, request_timeout=30)
"websites": {
"terms": {
"field": "website_id",
"size": 100000,
"min_doc_count": 1
}
}
},
"size": 0
}, index=self.index_name, request_timeout=30)
non_empty_websites = [bucket["key"] for bucket in result["aggregations"]["websites"]["buckets"]] non_empty_websites = [bucket["key"] for bucket in result["aggregations"]["websites"]["buckets"]]
for website in websites: for website in websites:
if website not in non_empty_websites: if website not in non_empty_websites:
yield website yield website