113 lines
2.7 KiB
Bash
113 lines
2.7 KiB
Bash
INDEX_ROOT="lawsoc"
|
|
INDEX_DATE=$(date +%Y-%m-%d)
|
|
INDEX_NAME="${INDEX_ROOT}_${INDEX_DATE}"
|
|
|
|
curl -XPOST "http://localhost:9200/${INDEX_NAME}/_close"
|
|
curl -XPUT "http://localhost:9200/${INDEX_NAME}/_settings" -d '{
|
|
"index":{
|
|
"analysis":{
|
|
"filter":{
|
|
"code":{
|
|
"type":"pattern_capture",
|
|
"preserve_original":1,
|
|
"patterns":[
|
|
"(\\p{Ll}+|\\p{Lu}\\p{Ll}+|\\p{Lu}+)",
|
|
"(\\d+)"
|
|
]
|
|
},
|
|
"email":{
|
|
"type": "pattern_capture",
|
|
"preserve_original" : 1,
|
|
"patterns" : [
|
|
"([^@]+)",
|
|
"(\\p{L}+)",
|
|
"(\\d+)",
|
|
"@(.+)"
|
|
]
|
|
},
|
|
"phonetic_metaphone" : {
|
|
"encoder" : "metaphone",
|
|
"replace" : false,
|
|
"type" : "phonetic"
|
|
},
|
|
"default_delimiter" : {
|
|
"type": "word_delimiter",
|
|
"generate_word_parts": true,
|
|
"generate_number_parts": true,
|
|
"catenate_words": false,
|
|
"catenate_numbers": false,
|
|
"catenate_all": false,
|
|
"split_on_case_change": true,
|
|
"preserve_original": false,
|
|
"split_on_numerics": true,
|
|
"stem_english_possessive": true
|
|
},
|
|
"my_delimiter" : {
|
|
"type": "word_delimiter",
|
|
"generate_word_parts": true,
|
|
"generate_number_parts": true,
|
|
"catenate_words": false,
|
|
"catenate_numbers": false,
|
|
"catenate_all": false,
|
|
"split_on_case_change": false,
|
|
"preserve_original": true,
|
|
"split_on_numerics": false,
|
|
"stem_english_possessive": false
|
|
}
|
|
},
|
|
"analyzer" : {
|
|
"index_analyzer" : {
|
|
"tokenizer" : "standard",
|
|
"filter" : [
|
|
"standard",
|
|
"lowercase",
|
|
"stop",
|
|
"asciifolding",
|
|
"phonetic_metaphone",
|
|
"my_delimiter"
|
|
]
|
|
},
|
|
"search_analyzer" : {
|
|
"tokenizer" : "standard",
|
|
"filter" : [
|
|
"standard",
|
|
"lowercase",
|
|
"stop",
|
|
"asciifolding",
|
|
"phonetic_metaphone"
|
|
]
|
|
},
|
|
"name_analyzer" : {
|
|
"type" : "custom",
|
|
"tokenizer" : "standard",
|
|
"filter" : [
|
|
"asciifolding",
|
|
"lowercase"
|
|
]
|
|
},
|
|
"email_analyzer" : {
|
|
"type" : "custom",
|
|
"tokenizer" : "uax_url_email",
|
|
"filter" : [
|
|
"email",
|
|
"lowercase",
|
|
"unique"
|
|
]
|
|
},
|
|
"url_analyzer" : {
|
|
"type" : "custom",
|
|
"tokenizer" : "uax_url_email"
|
|
},
|
|
"metaphone_analyzer" : {
|
|
"type" : "custom",
|
|
"tokenizer" : "standard",
|
|
"filter" : [
|
|
"phonetic_metaphone"
|
|
]
|
|
}
|
|
}
|
|
}
|
|
}
|
|
}'
|
|
curl -XPOST "http://localhost:9200/${INDEX_NAME}/_open"
|