文档：https://www.elastic.co/guide/en/machine-learning/current/ml-nlp-deploy-model.html

部署刚刚下载好的内置模型

部署内存不够用

还得花钱，拉几把倒吧。就用自带的吧。

测试模型

POST _ml/trained_models/lang_ident_model_1/_infer
{
  "docs":[{"text": "The fool doth think he is wise, but the wise man knows himself to be a fool."}]
}

{
  "inference_results": [
    {
      "predicted_value": "en",
      "prediction_probability": 0.9999658805366392,
      "prediction_score": 0.9999658805366392
    }
  ]
}

创建管道

添加处理器

reference 推理

# Field map
{
  "message": "text"
}
# Inference configuration
{
  "classification":{
    "num_top_classes":5
  }
}

set 设置

#  field 
event.ingested
# value 
{{{_ingest.timestamp}}}

失败处理器

测试

[
  {
    "_source": {
    "text_field":"Hello, my name is Josh and I live in Berlin."
    }
  }
]

[
 {
   "_source":{
     "message":"Sziasztok! Ez egy rövid magyar szöveg. Nézzük, vajon sikerül-e azonosítania a language identification funkciónak? Annak ellenére is sikerülni fog, hogy a szöveg két angol szót is tartalmaz."
     }
  }
]

测试没问题，创建管道

使用

安装插件

注意版本号与es版本一直，都是8.9.1。安装完会自行重启。
下载mapper-annotated-text安装包

映射索引

注意message字段别写错

PUT ner-test
{
  "mappings": {
    "properties": {
      "ml.inference.predicted_value": {"type": "annotated_text"},
      "ml.inference.model_id": {"type": "keyword"},
      "message": {"type": "text"},
      "event.ingested": {"type": "date"}
    }
  }
}

索引文档

通过管道 lang_ident_model_1 索引一批文档

POST /_bulk?pipeline=lang_ident_model_1
{"create":{"_index":"ner-test","_id":"1"}}
{"message":"Hello, my name is Josh and I live in Berlin."}
{"create":{"_index":"ner-test","_id":"2"}}
{"message":"I work for Elastic which was founded in Amsterdam."}
{"create":{"_index":"ner-test","_id":"3"}}
{"message":"Elastic has headquarters in Mountain View, California."}
{"create":{"_index":"ner-test","_id":"4"}}
{"message":"Elastic's founder, Shay Banon, created Elasticsearch to solve a simple need: finding recipes!"}
{"create":{"_index":"ner-test","_id":"5"}}
{"message":"Elasticsearch is built using Lucene, an open source search library."}

或者用query

POST lang-test/_doc?pipeline=ner-test
{
  "message": "Mon pays ce n'est pas un pays, c'est l'hiver"
}

查看数据

"hits": [
      {
        "_index": "ner-test",
        "_id": "1",
        "_score": 1,
        "_source": {
          "message": "Hello, my name is Josh and I live in Berlin.",
          "event": {
            "ingested": "2024-04-13T20:31:48.855089336Z"
          },
          "ml": {
            "inference": {
              "predicted_value": "en",
              "top_classes": [
                {
                  "class_name": "en",
                  "class_probability": 0.9854748734614491,
                  "class_score": 0.9854748734614491
                },
                {
                  "class_name": "tg",
                  "class_probability": 0.003855695585908385,
                  "class_score": 0.003855695585908385
                },
                {
                  "class_name": "ig",
                  "class_probability": 0.0036940515396614113,
                  "class_score": 0.0036940515396614113
                },
                {
                  "class_name": "sw",
                  "class_probability": 0.0021393582129747924,
                  "class_score": 0.0021393582129747924
                },
                {
                  "class_name": "it",
                  "class_probability": 0.0011839650697029283,
                  "class_score": 0.0011839650697029283
                }
              ],
              "prediction_probability": 0.9854748734614491,
              "prediction_score": 0.9854748734614491,
              "model_id": "lang_ident_model_1"
            }
          }
        }
      },......

文档重新索引到新目标

POST _reindex
{
  "source": {
    "index": "ner-test-new",
    "size": 50
  },
  "dest": {
    "index": "ner-test",
    "pipeline": "lang_ident_model_1"
  }
}