OpenSearch 2.x-3.x: New Features and Ecosystem Evolution

# Create k-NN index
PUT /vectors
{
  "settings": {
    "index": {
      "knn": true,
      "knn.algo_param.ef_search": 100
    }
  },
  "mappings": {
    "properties": {
      "embedding": {
        "type": "knn_vector",
        "dimension": 128,
        "method": {
          "name": "hnsw",
          "space_type": "cosinesimil",
          "engine": "faiss"
        }
      }
    }
  }
}

# Search vectors
POST /vectors/_search
{
  "size": 10,
  "query": {
    "knn": {
      "embedding": {
        "vector": [0.1, 0.2, ...],
        "k": 10
      }
    }
  }
}

# Field-level security
PUT /_opendistro/_security/api/roles/custom-role
{
  "index_permissions": [{
    "index_patterns": [" sensitive-*"],
    "field_security": {
      "except": ["password", "ssn"]
    }
  }]
}

# Segments merging optimization
# Improved memory management
# Better caching strategies

# Binary vectors support
PUT /bin-vectors
{
  "mappings": {
    "properties": {
      "embedding": {
        "type": "knn_vector",
        "dimension": 128,
        "space_type": "hamming"
      }
    }
  }
}

# Hybrid search
POST /_search
{
  "query": {
    "bool": {
      "must": [
        { "knn": { "embedding": { "vector": [...], "k": 10 } } }
      ],
      "should": [
        { "match": { "content": "search query" } }
      ]
    }
  }
}

# Pipeline aggregations
POST /sales/_search
{
  "aggs": {
    "max_price": {
      "max": { "field": "price" }
    }
  }
}

# Create visualization
# Bar charts, line charts, pie charts
# Maps, heat maps
# Saved objects

# Dashboard JSON
{
  "title": "Sales Dashboard",
  "panels": [
    {"id": "chart1", "type": "line"},
    {"id": "chart2", "type": "bar"}
  ]
}

# pipeline.yaml
pipeline:
  source:
    file:
      path: "/path/to/logs.log"
  processor:
    - grok:
        match:
          message: '%{IP:client} %{WORD:method} %{URIPATHPARAM:request} %{NUMBER:status}'
  sink:
    - opensearch:
        hosts: ["https://localhost:9200"]
        index: "logs"

# Ingest logs
POST /_ingest/pipeline/logs
{
  "description": "Parse logs",
  "processors": [
    {
      "grok": {
        "field": "message",
        "patterns": ["%{TIMESTAMP_ISO8601:timestamp} %{LOGLEVEL:level} %{GREEDYDATA:message}"]
      }
    }
  ]
}