Marcin Stachniuk
25 października 2015
Marcin Stachniuk
Chorąży na chamberconf.pl
Recenzent książki: practicalunittesting.com
You Know, for Search…
Agenda
Instalacja ElasticSearch'a
1. Ściągnij i rozpakuj
curl -O https://download.elastic.co/elasticsearch/elasticsearch/elasticsearch-1.7.2.zip
2. Uruchom
bin/elasticsearch
bin/elasticsearch.bat
lub
3. I działaj
curl -X GET http://localhost:9200/
{
"status": 200,
"name": "Stiletto",
"cluster_name": "elasticsearch",
"version": {
...
"lucene_version": "4.10.4"
},
"tagline": "You Know, for Search"
}
4. Warto doinstalować Sense (Beta) Chrome plugin
Instalacja ElasticSearch'a
Podstawowe pojęcia
Elasticsearch | SQL |
---|---|
Index | Schema |
Type | Table |
Document | Row |
Operacje CRUD
Create
Tworzenie Indexu (opcjonalnie)
PUT /moviesindex
Tworzenie dokumentu
POST /moviesindex/movies
{
"movieid": 2761502,
"title": "Pulp Fiction",
"year": "1994",
"imdbid": null
}
PUT /moviesindex/movies/3129479
{
"movieid": 3129479,
"title": "Matrix",
"year": "1999",
"imdbid": "tt0133093"
}
Operacje CRUD
Read
Odczytywanie dokumentu
GET /moviesindex/movies/3129479
Wyszukiwanie wszystkich dokumentów
POST /moviesindex/movies/_search
Operacje CRUD
Update
Modyfikacja danego dokumentu (usunięcie i dodanie
na nowo)
PUT /moviesindex/movies/3129479
{
"movieid": 3129479,
"title": "Matrix Matrix Matrix",
"year": "1999",
"imdbid": "tt0133093"
}
Częściowa aktualizacja
POST /moviesindex/movies/3129479/_update
{
"doc" : {
"imdbid": "tt42"
}
}
Operacje CRUD
Delete
Usunięcie
DELETE /moviesindex/movies/3129479
Wyszukiwanie
URI Search
GET /esmd/movie/_search?q=title:matrix
Term Query
POST /esmd/movie/_search
{
"query" : {
"term" : { "title" : "matrix" }
}
}
Term Query, full text search
POST /esmd/movie/_search
{
"query" : {
"term" : { "_all" : "matrix" }
}
}
Wyszukiwanie
Match query
POST /esmd/movie/_search
{
"query": {
"match": {
"title": {
"query": "matrix revolution",
"operator": "or"
}
}
}
}
Wyszukiwanie
Multi Match Query
POST /esmd/movie/_search
{
"query": {
"multi_match" : {
"query": "2012",
"fields": [ "title", "year" ]
}
}
}
Multi Match Query
POST /esmd/movie/_search
{
"query": {
"multi_match" : {
"query": "2012",
"fields" : [ "title^3", "*id" ]
}
}
}
Wyszukiwanie
Bool query
POST /esmd/movie/_search
{
"query": {
"bool": {
"must": {
"match": {
"title": {
"query": "american pie",
"operator": "and"
}
}
},
"should": [
{
"term": {
"year": {
"value": "2010"
}
}
}
]
}
}
}
Wyszukiwanie
Fuzzy query
POST /esmd/movie/_search
{
"query" : {
"fuzzy" : { "title" : "matrixx" }
}
}
Domyślne odległości dla tekstów
Wyszukiwanie
Query string query
POST /esmd/movie/_search
{
"query": {
"query_string": {
"default_field": "title",
"query": "this AND that OR thus"
}
}
}
Query string query
POST /esmd/movie/_search
{
"query": {
"query_string": {
"query": "(title:this OR plot:this) AND (title:that OR plot:that)"
}
}
}
Wyszukiwanie
Reqexp query
POST /esmd/movie/_search
{
"query": {
"regexp": {
"title": "a.*q"
}
}
}
Wyszukiwanie
Poświetlanie znalezionego fragmentu
POST /esmd/movie/_search
{
"query": {
"term": {
"plot": "matrix"
}
},
"highlight": {
"fields": {
"title": {
"type": "plain",
"force_source" : true
},
"plot": {
"type": "plain",
"force_source" : true
}
}
}
}
Wyszukiwanie - wszystkie query
Match Query
Multi Match Query
Bool Query
Boosting Query
Common Terms Query
Constant Score Query
Dis Max Query
Filtered Query
Fuzzy Like This Query
Fuzzy Like This Field Query
Function Score Query
Fuzzy Query
GeoShape Query
Has Child Query
Has Parent Query
Ids Query
Indices Query
Match All Query
More Like This Query
Nested Query
Prefix Query
Query String Query
Simple Query String Query
Range Query
Regexp Query
Span First Query
Span Multi Term Query
Span Near Query
Span Not Query
Span Or Query
Span Term Query
Term Query
Terms Query
Top Children Query
Wildcard Query
Minimum Should Match
Multi Term Query Rewrite
Template Query
Filtrowanie
Kiedy używać?
Filtrowanie
Term Filter
POST /esmd/movie/_search
{
"filter": {
"term": {
"year": "2017"
}
}
}
Filtrowanie
Range filter
POST /esmd/movie/_search
{
"filter": {
"range": {
"year": {
"from": 1889,
"to": 1890
}
}
}
}
Filtrowanie
Exist + not + and filter
POST /esmd/movie/_search
{
"filter": {
"and": {
"filters": [
{
"not": {
"exists": {
"field": "rating"
}
}
},
{
"not": {
"exists": {
"field": "directors"
}
}
}
]
}
}
}
Wszystkie filtry
And Filter
Bool Filter
Exists Filter
Geo Bounding Box Filter
Geo Distance Filter
Geo Distance Range Filter
Geo Polygon Filter
GeoShape Filter
Geohash Cell Filter
Has Child Filter
Has Parent Filter
Ids Filter
Indices Filter
Limit Filter
Match All Filter
Missing Filter
Nested Filter
Not Filter
Or Filter
Prefix Filter
Query Filter
Range Filter
Regexp Filter
Script Filter
Term Filter
Terms Filter
Type Filter
Jak to wszystko działa?
Odwrócony index
1. The quick brown fox jumped over the lazy dog
2. Quick brown foxes leap over lazy dogs in summer
Analiza tekstu
1. Tokenizacja tekstu do pojedynczych słów:
The quick brown foxes → [The, quick, brown, foxes]
2. Zamiana liter na małe
The → the
3. Usuwanie "stopwords"
[The, quick, brown, foxes] → [quick, brown, foxes]
Domyślne Angielskie stopwords (link)
a, an, and, are, as, at, be, but, by, for, if, in, into, is, it, no,
not, of, on, or, such, that, the, their, then, there, these, they,
this, to, was, will, with
Jak szukać Szekspira? "To be or not to be"
Analiza tekstu
4. Sprowadzanie do form podstawowych
foxes → fox
5. Robienie słów bardziej wyszukiwalnych, np.
John's → john
l'église → eglis
äußerst → ausserst
Analiza tekstu
Wsparcie dla języków:
Arabic, Armenian, Basque, Brazilian, Bulgarian, Catalan, Chinese, Czech, Danish, Dutch, English, Finnish, French, Galician, German, Greek, Hindi, Hungarian, Indonesian, Irish, Italian, Japanese, Korean, Kurdish, Norwegian, Persian, Portuguese, Romanian, Russian, Spanish, Swedish, Turkish, and Thai.
Wsparcie do polskiego za pomocą plugina:
Odwrócony index
Query: quick brown
Obliczanie punktacji
Częstość słów (term frequency)
Odwrotna częstość w dokumentach (inverse document frequency)
Norma długości pola (field-length norm)
Ostatecznie:
Obliczanie punktacji
Podgląd punktacji
POST /esmd/movie/_search?explain
{
"query": {
"term": { "title": "matrix" }
}
}
"_explanation": {
"value": 4.9666376,
"description": "weight(title:matrix in 21100) [PerFieldSimilarity], result of:",
"details": [
{
"value": 4.9666376,
"description": "fieldWeight in 21100, product of:",
"details": [
{
"value": 1,
"description": "tf(freq=1.0), with freq of:",
"details": [
{
"value": 1,
"description": "termFreq=1.0"
}
]
},
{
"value": 11.352315,
"description": "idf(docFreq=1, maxDocs=62659)"
},
{
"value": 0.4375,
"description": "fieldNorm(doc=21100)"
}
]
}
]
}
Trochę o architekturze
Klaster (Cluster) - identyfikowany przez nazwę
Typy węzłów:
Typy shard'ów:
Architektua ElasticSearch'a
Architektua ElasticSearch'a
Zapis
Architektua ElasticSearch'a
Wyszukiwanie
Agregacje
2 typy agregacji:
Agregacje
Sum Aggregation
POST /esmd/movie/_search
{
"size": 0,
"aggregations" : {
"ammount of votes" : {
"sum" : {
"field": "rating.votes"
}
}
}
}
Agregacje
Stats Aggregation
POST /esmd/movie/_search
{
"size": 0,
"aggregations": {
"stats of votes": {
"stats": {
"field": "rating.votes"
}
}
}
}
Agregacje
Terms Aggregation
POST /esmd/movie/_search
{
"size": 0,
"aggregations": {
"possible years of movies": {
"terms": {
"field": "year"
}
}
}
}
POST /esmd/movie/_search
{
"size": 0,
"aggregations": {
"possible years of movies": {
"terms": {
"field": "year",
"size": 100
}
}
}
}
Agregacje
Zagnieżdżone aggregacje
POST /esmd/movie/_search
{
"size": 0,
"aggregations": {
"possible years of movies": {
"terms": {
"field": "year"
},
"aggregations": {
"rank": {
"terms": {
"field": "rating.rank"
}
}
}
}
}
}
Agregacje mogą kłamać
Jak sobie z tym radzić?
Błędy:
Zwiększenie parametru "size"
Czasem agregacje dają niespodziewane wyniki
"buckets": [
{
"key": "american",
"doc_count": 1
},
{
"key": "fiction",
"doc_count": 1
},
{
"key": "pie",
"doc_count": 1
},
{
"key": "pulp",
"doc_count": 1
}
]
POST /moviesindex/movies/_search
{
"size": 0,
"aggregations" : {
"by titles": {
"terms": {
"field": "title"
}
}
}
}
Konieczne jest wtedy zastosowanie mappingu
"buckets": [
{
"key": "American Pie",
"doc_count": 1
},
{
"key": "Pulp Fiction",
"doc_count": 1
}
]
PUT /moviesindex
{
"mappings": {
"movies": {
"properties": {
"title" : {
"type": "string",
"fields": {
"raw" : {
"type": "string",
"index": "not_analyzed"
}
}
}
}
}
}
}
POST /moviesindex/movies/_search
{
"size": 0,
"aggregations" : {
"by titles": {
"terms": {
"field": "title.raw"
}
}
}
}
Wszystkie agregacje
Min Aggregation
Max Aggregation
Sum Aggregation
Avg Aggregation
Stats Aggregation
Extended Stats Aggregation
Value Count Aggregation
Percentiles Aggregation
Percentile Ranks Aggregation
Cardinality Aggregation
Geo Bounds Aggregation
Top hits Aggregation
Scripted Metric Aggregation
Global Aggregation
Filter Aggregation
Filters Aggregation
Missing Aggregation
Nested Aggregation
Reverse nested Aggregation
Children Aggregation
Terms Aggregation
Significant Terms Aggregation
Range Aggregation
Date Range Aggregation
IPv4 Range Aggregation
Histogram Aggregation
Date Histogram Aggregation
Geo Distance Aggregation
GeoHash grid Aggregation
Sugestie
Sugestie
Term suggester
POST /esmd/_suggest
{
"sugest in title": {
"text": "matri",
"term": {
"field": "title"
}
}
}
Wszystkie podpowiadaczki:
"sugest in title": [
{
"text": "matri",
"offset": 0,
"length": 5,
"options": [
{
"text": "matrix",
"score": 0.8,
"freq": 26
},
{
"text": "materi",
"score": 0.8,
"freq": 18
},
{
"text": "matti",
"score": 0.8,
"freq": 11
}, ...
Ekosystem
Jak żyć z ElasticSearchem?
Monitoring Marvel
Monitoring Marvel
Monitoring Kopf
Monitoring Elastic HQ
Monitoring Bigdesk
Analytics and search dashboard for Elasticsearch
Elasticsearch as a Service
Linki
Marcin Stachniuk
25 października 2015
Dziękuję!