Question
'Did you mean' functionality with AEM OOTB Search
I am trying to do a POC on did you mean functionality similar to google one using OOTB spellcheck feature. I have gone through the documentation mentioned below but I am not getting results. I have reindexed the content after these steps as well.
https://jackrabbit.apache.org/oak/docs/query/lucene.html#spellchecking
My query -> "SELECT [rep:suggest()] FROM [cq:Page] WHERE SUGGEST('tes') AND (ISDESCENDANTNODE('/content/myproject'))"
{
"statement": "SELECT [rep:suggest()] FROM [cq:Page] WHERE SUGGEST('tes') AND (ISDESCENDANTNODE('/content/myproject'))",
"language": "JCR-SQL2",
"explain": {
"logs": [
"Parsing JCR-SQL2 statement: explain SELECT [rep:suggest()] FROM [cq:Page] WHERE SUGGEST('tes') AND (ISDESCENDANTNODE('/content/myproject'))\r\n",
"cost using filter Filter(query=explain SELECT [rep:suggest()] FROM [cq:Page] WHERE SUGGEST('tes') AND (ISDESCENDANTNODE('/content/myproject')), path=/content/myproject//*, property=[native*lucene=[suggest?term=tes]])\r\n",
"cost for reference is Infinity\r\n",
"cost for property is Infinity\r\n",
"cost for nodeType is Infinity\r\n",
"Applicable IndexingRule found IndexRule: nt:base\r\n",
"Applicable IndexingRule found IndexRule: nt:base\r\n",
"Applicable IndexingRule found IndexRule: cq:Page\r\n",
"Applicable IndexingRule found IndexRule: cq:Page\r\n",
"[/oak:index/cqPageLucene-custom-3] Copied 0 files totaling 0 B [took 8ms]\r\n",
"[/oak:index/cqPageLucene-custom-3] Copied 0 files totaling 0 B [took 2ms]\r\n",
"[/oak:index/cqPageLucene-custom-3] Closing NRTIndex [UNKNOWN]\r\n",
"Applicable IndexingRule found IndexRule: nt:base\r\n",
"Applicable IndexingRule found IndexRule: cq:Page\r\n",
"[/oak:index/cqPageLucene] Copied 0 files totaling 0 B [took 7ms]\r\n",
"[/oak:index/cqPageLucene] Copied 0 files totaling 0 B [took 3ms]\r\n",
"[/oak:index/cqPageLucene] Closing NRTIndex [UNKNOWN]\r\n",
"Applicable IndexingRule found IndexRule: cq:Page\r\n",
"Applicable IndexingRule found IndexRule: nt:base\r\n",
"Applicable IndexingRule found IndexRule: nt:base\r\n",
"Applicable IndexingRule found IndexRule: nt:base\r\n",
"Lucene46Codec is loaded: oakCodec\r\n",
"Codec is loaded: Lucene46\r\n",
"Applicable IndexingRule found IndexRule: nt:base\r\n",
"cost for [/oak:index/cqPageLucene-custom-3] of type (lucene-property) with plan [lucene:cqPageLucene-custom-3(/oak:index/cqPageLucene-custom-3) *:*] is 2.00\r\n",
"cost for [/oak:index/cqPageLucene] of type (lucene-property) with plan [lucene:cqPageLucene(/oak:index/cqPageLucene) *:*] is 2.00\r\n",
"cost for lucene-property is Infinity\r\n",
"minCost: 2.2 of index :aggregate lucene < best Cost: 2.0 from index: lucene-property. Further index evaluation will be skipped\r\n",
"selected index org.apache.jackrabbit.oak.plugins.index.lucene.LucenePropertyIndex@33dfe2ac with plan /oak:index/cqPageLucene-custom-3 and org.apache.jackrabbit.oak.plugins.index.lucene.LucenePropertyIndex@33dfe2ac with plan /oak:index/cqPageLucene have similar costs 2.0 and 2.0 for query Filter(query=explain SELECT [rep:suggest()] FROM [cq:Page] WHERE SUGGEST('tes') AND (ISDESCENDANTNODE('/content/myproject')), path=/content/myproject//*, property=[native*lucene=[suggest?term=tes]]) - check query explanation / index definitions\r\n",
"cost for traverse is Infinity\r\n",
"count:\t1\tquery:\texplain SELECT [rep:suggest()] FROM [cq:Page] WHERE SUGGEST('x') AND (ISDESCENDANTNODE('x'))\r\n"
],
"plan": "[cq:Page] as [cq:Page] /* lucene:cqPageLucene-custom-3(/oak:index/cqPageLucene-custom-3) *:* where (suggest([cq:Page], 'tes')) and (isdescendantnode([cq:Page], [/content/myproject])) */",
"propertyIndexes": [
"cqPageLucene-custom-3(/oak:index/cqPageLucene-custom-3)"
]
},
"heuristics": {
"results": [
"/",
"/"
],
"count": 2,
"readPageTime": 5,
"countTime": 5,
"executionTime": 131,
"getNodesTime": 5,
"totalTime": 141
}
}
Attaching the query explanation JSON. If anyone has tried spellcheck for similar use case please let me know.