Docker と curl コマンドで試すベクトル検索 (Solr9)

用意するもの

• Docker
• curlコマンド

Docker で Solr の起動

command

docker run -d -p 8983:8983 --name my_solr solr

Docker で Solr コレクションの作成

command

docker exec -it my_solr bin/solr create -c sandbox

output

WARNING: Using _default configset with data driven schema functionality. NOT RECOMMENDED for production use.
         To turn off: bin/solr config -c sandbox -p 8983 -action set-user-property -property update.autoCreateFields -value false

Created new core 'sandbox'

>

フィールドタイプの定義

ここでは例としてベクトルのサイズが4のフィールドタイプを定義しています。

curl -X POST -H 'Content-type:application/json' --data-binary '{"add-field-type":{"name":"vector4","class":"solr.DenseVectorField","vectorDimension":4,"similarityFunction":"cosine"}}' http://localhost:8983/solr/sandbox/schema

フィールドの定義

curl -X POST -H 'Content-type:application/json' --data-binary '{  "add-field":{ "name":"v4", "type":"vector4", "indexed":true, "stored":true, } }' http://localhost:8983/solr/sandbox/schema

検索対象データの追加

curl -X POST -H 'Content-Type:application/json' --data-binary '[ { "id": "1", "v4": [1.0, 0.0, 0.0, 0.0] } ]' http://localhost:8983/solr/sandbox/update/json/docs?commit=true

curl -X POST -H 'Content-Type:application/json' --data-binary '[ { "id": "2", "v4": [0.0, 1.0, 0.0, 0.0] } ]' http://localhost:8983/solr/sandbox/update/json/docs?commit=true

curl -X POST -H 'Content-Type:application/json' --data-binary '[ { "id": "3", "v4": [0.0, 0.0, 1.0, 0.0] } ]' http://localhost:8983/solr/sandbox/update/json/docs?commit=true

curl -X POST -H 'Content-Type:application/json' --data-binary '[ { "id": "4", "v4": [0.0, 0.0, 0.0, 1.0] } ]' http://localhost:8983/solr/sandbox/update/json/docs?commit=true

検索対象データの追加（複数文書をまとめて追加したい場合）

curl -X POST -H 'Content-Type:application/json' --data-binary '[ { "id": "1", "v4": [1.0, 0.0, 0.0, 0.0] }, { "id": "2", "v4": [0.0, 1.0, 0.0, 0.0] }, { "id": "3", "v4": [0.0, 0.0, 1.0, 0.0] }, { "id": "4", "v4": [0.0, 0.0, 0.0, 1.0] } ]' http://localhost:8983/solr/sandbox/update/json/docs?commit=true

検索

curl -X GET 'http://localhost:8983/solr/sandbox/select?q=%7B!knn%20f=v4%20topK=10%7D%5B1.0%2C1.0%2C0.0%2C0.0%5D'

検索結果

{
  "responseHeader":{
    "status":0,
    "QTime":91,
    "params":{
      "q":"{!knn f=v4 topK=10}[1.0,1.0,0.0,0.0]"
    }
  },
  "response":{
    "numFound":4,
    "start":0,
    "numFoundExact":true,
    "docs":[{
      "id":"1",
      "v4":[1.0,0.0,0.0,0.0],
      "_version_":1804607180982190080
    },{
      "id":"2",
      "v4":[0.0,1.0,0.0,0.0],
      "_version_":1804607188756332544
    },{
      "id":"3",
      "v4":[0.0,0.0,1.0,0.0],
      "_version_":1804607196582903808
    },{
      "id":"4",
      "v4":[0.0,0.0,0.0,1.0],
      "_version_":1804607205283987456
    }]
  }
}

---

以上.