Introduction#

Github this shows how to setup opensearch serverless to build a note bookmark application. Frontend updated later on.

  • setup aws opensearch
  • interact via curl
  • using python sdk
  • basic dsl language
OpenSearch Serverless

Architecture#

Configure Permission#

from opensearch aws console, need grant permission to user, role to access collection and index by polcies, example

[
  {
    "Rules": [
      {
        "Resource": ["collection/test"],
        "Permission": ["aoss:*"],
        "ResourceType": "collection"
      },
      {
        "Resource": ["index/test/*"],
        "Permission": ["aoss:*"],
        "ResourceType": "index"
      }
    ],
    "Principal": ["arn:aws:iam::$ACCOUNT_ID:role/RoleForLambdaIndexOpenSearch"],
    "Description": "lambdaindexopensearch"
  }
]

OpenSearch Client#

auth and create a client

# opensearch domain
if "OPENSEARCH_DOMAIN" in os.environ:
    pass
else:
    os.environ["OPENSEARCH_DOMAIN"] = ""
    os.environ["REGION"] = "us-east-1"


# host and opensearch client
host = os.environ["OPENSEARCH_DOMAIN"]
client = boto3.client("opensearchserverless")
service = "aoss"
region = os.environ["REGION"]
credentials = boto3.Session().get_credentials()


# auth
awsauth = AWS4Auth(
    credentials.access_key,
    credentials.secret_key,
    region,
    service,
    session_token=credentials.token,
)

create a client

client = OpenSearch(
    hosts=[{"host": host, "port": 443}],
    http_auth=awsauth,
    use_ssl=True,
    verify_certs=True,
    connection_class=RequestsHttpConnection,
    timeout=300,
)

Basic Operation#

create an index

resp = client.indices.create(index_name)

delete an index

resp = client.indices.delete(index=index_name)

index a document

def test_index_data(index_name):
    """
    test index a document into the opensearch serverless
    """
    #
    response = client.index(
        index=index_name,
        # body={"title": "pricing", "creator": "Larry David", "year": 1989},
       body={
            "DocumentTitle": "pricing",
            "DocumentExcerpt": "100",
            "DocumentURI":"URL"
        },
        id=str(uuid.uuid4()),
    )
    print(response)

query data

def test_query_index(query):
    """
    test query from opensearch serverless
    """
    response = client.search(
        index="cdk-entest", body={"query": {"query_string": {"query": query}}}
    )
    print(response)

or test another query

def test_query(query_request):
    """
    query opensearch
    """
    resp = client.search(
        index="cdk-entest",
        body={
            "query": {
                "multi_match": {
                    "query": query_request,
                    "fields": [
                        "DocumentTitle",
                        "DocumentExcerpt",
                        "DocumentURI",
                    ],
                }
            }
        },
    )
    print(resp)

delete a document given the document id

client.delete(index=index_name, id=id)

Troubleshooting#

update boto3

python3 -m pip install boto3 --upgrade