Ein schöner Tag in Köln im Café an der Straßenecke

imcafe.sh

# required plugins: analysis-icu, analysis-combo
# ./bin/plugin -install elasticsearch/elasticsearch-analysis-icu/1.7.0
# ./bin/plugin -install yakaz/elasticsearch-analysis-combo/1.1.0
# 
curl -XDELETE 'localhost:9200/ling'

echo '{
  "settings" : {
    "index" : {
        "analysis" : {
            "analyzer" : {
                "snow_icu" : {
                    "type" : "custom",
                    "tokenizer" : "icu_tokenizer",
                    "filter" : [ "snowball", "icu_folding" ]
                },
                "default" : {
                    "type" : "combo",
                    "sub_analyzers" : [ "standard", "snow_icu" ],
                    "filter" : "unique"
                }
            },
            "filter" : {
                "snowball" : {
                    "type" : "snowball",
                    "language" : "German2"
                }
            }
        }
    }
  }
}' > settings.json


curl -XPUT 'localhost:9200/ling' --data-binary @settings.json

curl -XGET 'localhost:9200/ling/_settings?pretty'

echo '{
    "sentence" : "Ein schöner Tag in Köln im Café an der Straßenecke"
}' > ling-1.json

echo '{
    "sentence" : "Ein schoener Tag in Koeln im Café an der Straßenecke"
}' > ling-2.json


echo '{
    "sentence" : "Ein schoner Tag in Koln im Cafe an der Strassenecke"
}' > ling-3.json


curl -XPUT 'localhost:9200/ling/test/1' --data-binary @ling-1.json
curl -XPUT 'localhost:9200/ling/test/2' --data-binary @ling-2.json
curl -XPUT 'localhost:9200/ling/test/3' --data-binary @ling-3.json

curl -XGET 'localhost:9200/_refresh'

# german umlauts and accents

echo '{
  "query": {
    "match": {
      "sentence": "Ein schöner Tag in Köln im Café an der Straßenecke"
    }
  }
}' > query-1.json


# german umlaut expansion

echo '{
  "query": {
    "match": {
      "sentence": "Ein schoener Tag in Koeln im Café an der Straßenecke"
    }
  }
}' > query-2.json


# base form reduction, sharp s folding

echo '{
  "query": {
    "match": {
      "sentence": "Ein schoner Tag in Koln im Cafe an der Strassenecke"
    }
  }
}' > query-3.json

# compensate snowball overstemming. With snow_icu 1 hit => with combo(standard+snow_icu) 3 hits

echo '{
  "query": {
    "match": {
      "sentence": "cafe"
    }
  }
}' > query-4.json

# 4x3 hits 

curl -XPOST 'localhost:9200/ling/test/_search?pretty' --data-binary @query-1.json
curl -XPOST 'localhost:9200/ling/test/_search?pretty' --data-binary @query-2.json
curl -XPOST 'localhost:9200/ling/test/_search?pretty' --data-binary @query-3.json
curl -XPOST 'localhost:9200/ling/test/_search?pretty' --data-binary @query-4.json

exit