# SPDX-License-Identifier: Apache-2.0 # # The OpenSearch Contributors require contributions made to # this file be licensed under the Apache-2.0 license or a # compatible open source license. # # Modifications Copyright OpenSearch Contributors. See # GitHub history for details. # ------------------------------------------------------------------------------------------ # THIS CODE IS AUTOMATICALLY GENERATED AND MANUAL EDITS WILL BE LOST # # To contribute, kindly make modifications in the opensearch-py client generator # or in the OpenSearch API specification, and run `nox -rs generate`. See DEVELOPER_GUIDE.md # and https://github.com/opensearch-project/opensearch-api-specification for details. # -----------------------------------------------------------------------------------------+ from typing import Any from .utils import NamespacedClient, _make_path, query_params class ListClient(NamespacedClient): @query_params("error_trace", "filter_path", "human", "pretty", "source") def help( self, params: Any = None, headers: Any = None, ) -> Any: """ Returns help for the List APIs. :arg error_trace: Whether to include the stack trace of returned errors. Default is false. :arg filter_path: Used to reduce the response. This parameter takes a comma-separated list of filters. It supports using wildcards to match any field or part of a field’s name. You can also exclude fields with "-". :arg human: Whether to return human readable values for statistics. Default is True. :arg pretty: Whether to pretty format the returned JSON response. Default is false. :arg source: The URL-encoded request definition. Useful for libraries that do not accept a request body for non-POST requests. """ return self.transport.perform_request( "GET", "/_list", params=params, headers=headers ) @query_params( "bytes", "cluster_manager_timeout", "error_trace", "expand_wildcards", "filter_path", "format", "h", "health", "help", "human", "include_unloaded_segments", "local", "master_timeout", "next_token", "pretty", "pri", "s", "size", "sort", "source", "time", "v", ) def indices( self, index: Any = None, params: Any = None, headers: Any = None, ) -> Any: """ Returns paginated information about indexes including number of primaries and replicas, document counts, disk size. :arg index: Comma-separated list of data streams, indexes, and aliases used to limit the request. Supports wildcards (`*`). To target all data streams and indexes, omit this parameter or use `*` or `_all`. :arg bytes: The unit used to display byte values. Valid choices are b, g, gb, k, kb, m, mb, p, pb, t, tb. :arg cluster_manager_timeout: Operation timeout for connection to cluster-manager node. :arg error_trace: Whether to include the stack trace of returned errors. Default is false. :arg expand_wildcards: The type of index that wildcard patterns can match. :arg filter_path: Used to reduce the response. This parameter takes a comma-separated list of filters. It supports using wildcards to match any field or part of a field’s name. You can also exclude fields with "-". :arg format: A short version of the Accept header, such as `JSON`, `YAML`. :arg h: Comma-separated list of column names to display. :arg health: The health status used to limit returned indexes. By default, the response includes indexes of any health status. Valid choices are green, red, yellow. :arg help: Return help information. Default is false. :arg human: Whether to return human readable values for statistics. Default is True. :arg include_unloaded_segments: If `true`, the response includes information from segments that are not loaded into memory. Default is false. :arg local: Return local information, do not retrieve the state from cluster-manager node. Default is false. :arg master_timeout (Deprecated: To promote inclusive language, use `cluster_manager_timeout` instead.): Operation timeout for connection to cluster-manager node. :arg next_token: Token to retrieve next page of indexes. :arg pretty: Whether to pretty format the returned JSON response. Default is false. :arg pri: If `true`, the response only includes information from primary shards. Default is false. :arg s: Comma-separated list of column names or column aliases to sort by. :arg size: Maximum number of indexes to be displayed in a page. :arg sort: Defines order in which indexes will be displayed. Accepted values are `asc` and `desc`. If `desc`, most recently created indexes would be displayed first. Valid choices are asc, desc. :arg source: The URL-encoded request definition. Useful for libraries that do not accept a request body for non-POST requests. :arg time: The unit used to display time values. Valid choices are d, h, m, micros, ms, nanos, s. :arg v: Verbose mode. Display column headers. Default is false. """ return self.transport.perform_request( "GET", _make_path("_list", "indices", index), params=params, headers=headers ) @query_params( "bytes", "cluster_manager_timeout", "error_trace", "filter_path", "format", "h", "help", "human", "local", "master_timeout", "next_token", "pretty", "s", "size", "sort", "source", "time", "v", ) def shards( self, index: Any = None, params: Any = None, headers: Any = None, ) -> Any: """ Returns paginated details of shard allocation on nodes. :arg index: A comma-separated list of data streams, indexes, and aliases used to limit the request. Supports wildcards (`*`). To target all data streams and indexes, omit this parameter or use `*` or `_all`. :arg bytes: The unit used to display byte values. Valid choices are b, g, gb, k, kb, m, mb, p, pb, t, tb. :arg cluster_manager_timeout: Operation timeout for connection to cluster-manager node. :arg error_trace: Whether to include the stack trace of returned errors. Default is false. :arg filter_path: Used to reduce the response. This parameter takes a comma-separated list of filters. It supports using wildcards to match any field or part of a field’s name. You can also exclude fields with "-". :arg format: A short version of the Accept header, such as `JSON`, `YAML`. :arg h: Comma-separated list of column names to display. :arg help: Return help information. Default is false. :arg human: Whether to return human readable values for statistics. Default is True. :arg local: Return local information, do not retrieve the state from cluster-manager node. Default is false. :arg master_timeout (Deprecated: To promote inclusive language, use `cluster_manager_timeout` instead.): Operation timeout for connection to cluster-manager node. :arg next_token: Token to retrieve next page of shards. :arg pretty: Whether to pretty format the returned JSON response. Default is false. :arg s: Comma-separated list of column names or column aliases to sort by. :arg size: Maximum number of shards to be displayed in a page. :arg sort: Defines order in which shards will be displayed. Accepted values are `asc` and `desc`. If `desc`, most recently created shards would be displayed first. Valid choices are asc, desc. :arg source: The URL-encoded request definition. Useful for libraries that do not accept a request body for non-POST requests. :arg time: The unit in which to display time values. Valid choices are d, h, m, micros, ms, nanos, s. :arg v: Verbose mode. Display column headers. Default is false. """ return self.transport.perform_request( "GET", _make_path("_list", "shards", index), params=params, headers=headers )
Memory