Getting started with semantic_text on AWS Bedrock

beckrock-semantic_text.json

GET /

PUT starwars
{
  "settings": {
    "analysis": {
      "analyzer": {
        "my_analyzer": {
          "char_filter": [
            "html_strip"
          ],
          "tokenizer": "standard",
          "filter": [
            "lowercase",
            "stop",
            "snowball"
          ]
        }
      }
    }
  },
  "mappings": {
    "properties": {
      "quote": {
        "type": "text",
        "analyzer": "my_analyzer"
      }
    }
  }
}

GET starwars/_analyze
{
  "analyzer" : "my_analyzer",
  "text":  "These are <em>not</em> the droids you are looking for."
}

PUT starwars/_doc/1
{
  "quote": "These are <em>not</em> the droids you are looking for.",
  "speaker": "Obi-Wan Kenobi"
}
PUT starwars/_doc/2
{
  "quote": "<b>No.</b> I am your father.",
  "speaker": "Darth Vader"
}

GET starwars/_search
{
    "query": {
        "match": {
          "quote": "Droid"
        }
    }
}


PUT _inference/completion/titan_completion
{
  "service": "amazonbedrock",
  "service_settings": {
    "access_key": "AKI...",
    "secret_key": "...",
    "region": "us-west-2",
    "provider": "amazontitan",
    "model": "amazon.titan-text-express-v1"
  }
}

POST _inference/completion/titan_completion
{
  "input": "Who said 'these are not the droids you are looking for'?"
}

PUT _inference/text_embedding/titan_embeddings
{
  "service": "amazonbedrock",
  "service_settings": {
    "access_key": "AKI...",
    "secret_key": "...",
    "region": "us-west-2",
    "provider": "amazontitan",
    "model": "amazon.titan-embed-text-v2:0"
  }
}

GET _inference

PUT semantic-starwars
{
  "settings": {
    "analysis": {
      "analyzer": {
        "my_analyzer": {
          "char_filter": [
            "html_strip"
          ],
          "tokenizer": "standard",
          "filter": [
            "lowercase",
            "stop",
            "snowball"
          ]
        }
      }
    }
  },
  "mappings": {
    "properties": {
      "quote": {
        "type": "text",
        "analyzer": "my_analyzer",
        "copy_to": ["quote_titan"]
      },
      "quote_titan": {
        "type": "semantic_text",
        "inference_id": "titan_embeddings"
      }
    }
  }
}

POST _reindex
{
  "source": {
    "index": "starwars"
  },
  "dest": {
    "index": "semantic-starwars"
  }
}

GET semantic-starwars/_doc/1

GET semantic-starwars/_search
{
  "query": {
    "match": {
      "quote": "robot"
    }
  }
}

GET semantic-starwars/_search
{
  "query": {
    "semantic": {
      "field": "quote_titan",
      "query": "robot"
    }
  },
  "_source": "quote"
}

GET semantic-starwars/_search
{
  "query": {
    "semantic": {
      "field": "quote_titan",
      "query": "dad"
    }
  },
  "_source": "quote"
}

GET semantic-starwars/_search
{
  "query": {
    "bool": {
      "should": [
        {
          "semantic": {
            "field": "quote_titan",
            "query": "android",
            "boost": 1,
            "_name": "elser"
          }
        },
        {
          "match": {
            "quote": {
              "query": "android",
              "boost": 3,
              "_name": "keyword"
            }
          }
        }
      ]
    }
  },
  "_source": "quote"
}

DELETE starwars
DELETE semantic-starwars