Elasticsearch Python Client release notes

Review the changes, fixes, and more in each version of Elasticsearch Python Client.

To check for security updates, go to Security announcements for the Elastic stack.

• ES|QL query builder (technical preview) (#2997)
• Update OpenTelemetry conventions (#2999)
• Add option to disable accurate reporting of file and line location in warnings (Fixes #3003) (#3006)

• Remove if_primary_term, if_seq_no and op_type from Create API
• Remove master_timeout from Ingest Get Ip Location Database API
• Remove application, priviledge and username from the Security Get User API
• Rename type_query_string to type in License Post Start Trial API
• Add require_data_stream to Index API
• Add settings_filter to Cluster Get Component Template API
• Add cause to Cluster Put Component Template API
• Add master_timeout to Cluster State API
• Add ccs_minimize_roundtrips to EQL Search API
• Add keep_alive and keep_on_completion to ES|QL Async Query API
• Add format to ES|QL Async Query Get API
• Add ES|QL Get Query and List Queries APIs
• Add Indices Delete Data Stream Options API
• Add Indices Get Data Stream Options and Put Data Stream Options APIS
• Add Indices Get Data Stream Settings and Put Data Stream Settings APIs
• Add allow_no_indices, expand_wildcards and ignore_available to Indices Recovery API
• Add Indices Remove Block API
• Add Amazon Sagemaker to Inference API
• Add input_type to Inference API
• Add timeout to all Inference Put APIs
• Add Inference Put Custom API
• Add Inference Put DeepSeek API
• Add task_settings to Put HuggingFace API
• Add refresh to Security Grant API Key API
• Add wait_for_completion to the Snapshot Delete API
• Add state to Snapshot Get API
• Add refresh to Synonyms Put Synonym, Put Synonym Rule and Delete Synonym Rule APIs

• Handle lists in copy_to option in DSL field declarations correctly (Fixes #2992) (#2993)
• Add index_options to SparseVector type
• Add SparseVectorIndexOptions type
• Add key to FiltersBucket type

Other changes

• Drop support for Python 3.8 (#3001)

• ES|QL query builder (technical preview) (#2997)
• Add option to disable accurate reporting of file and line location in warnings (Fixes #3003) (#3006)

• Remove if_primary_term, if_seq_no and op_type from Create API
• Remove stored_fields from Get Source API
• Remove master_timeout from Ingest Get Ip Location Database API
• Remove application, priviledge and username from the Security Get User API
• Rename type_query_string to type in License Post Start Trial API
• Add require_data_stream to Index API
• Add settings_filter to Cluster Get Component Template API
• Add cause to Cluster Put Component Template API
• Add ccs_minimize_roundtrips to EQL Search API
• Add keep_alive and keep_on_completion to ES|QL Async Query API
• Add format to ES|QL Async Query Get API
• Add allow_no_indices, expand_wildcards and ignore_available to Indices Recovery API
• Add timeout to all Inference Put APIs
• Add refresh to Security Get User Profile API
• Add wait_for_completion to the Snapshot Delete API

• Handle lists in copy_to field option correctly (Fixes #2992) (#2993)
• Add key to FiltersBucket type

• Add rescore_vector to DenseVectorIndexOptions

• Surface caused_by in ApiError (#2932)
• Clarify Elasticsearch 9.x compatibility (#2928)
• Reorganize Sphinx docs to only include reference pages (#2776)

• Remove deprecated Elasticsearch() options (#2840)
• Remove deprecated url_prefix and use_ssl options (#2797)

See the breaking changes page for more details.

• Merge elasticsearch-dsl-py _ package (#2736)
• Add Python DSL documentation (#2761)
• Autogenerate DSL field classes from schema (#2780)
• Improve DSL documentation examples with class-based queries and type hints (#2857)
• Document the use of param() in Python DSL methods (#2861)
• Migrate documentation from AsciiDoc to Markdown format (#2806)
• Document use of sub-clients (#2798)
• Document how to make API calls (#2843)
• Fix simulate sub-client documentation (#2749)

• Remove deprecated /_knn_search API
• Remove Unfreeze an index API
• Remove min_compatible_shard_node from Search and Async Search Submit APIs
• Remove local parameter from cat alias, Alias exists, and Get alias APIs
• Remove verbose from Index segments API
• Remove include_model_definition from Get trained model configuration info API
• Remove wait_for_active_shards from experimental Get field usage stats API
• Support soft-deletes in connectors:
  • Add hard to Delete connector API
  • Add include_deleted to Get and List Connector APIs
• Add master_timeout to Migrate to data tiers routing APIs
• Add master_timeout to the Alias exists and Get alias APIs.
• Add expand_wildcards to Create snapshot API
• Rename incorrect access_token to token in Logout of OpenID Connect API
• Add inference APIs: Alibaba Cloud AI Search, Amazon Bedrock, Anthropic, Azure AI Studio, Azure OpenAI, Cohere, Elastic Inference Service (EIS), Elasticsearch, ELSER, Google AI Studio, Google Vertex AI, Hugging Face, Jina AI, Mistral, OpenAI, and Voyage AI
• Add Elastic Inference Service (EIS) chat completion API
• Add Reindex legacy backing indices APIs
• Add Create an index from a source index API
• Add include_source_on_error to Create, Index, Update and Bulk APIs
• Add Stop async ES|QL query API
• Add timeout to Resolve Cluster API
• Add adaptive_allocations body field to Start and Update a trained model deployment API
• Rename index_template_subtitutions to index_template_substitutions in Simulate data ingestion API
• Add if_primary_term, if_seq_no, op_type, require_alias and require_data_stream to Create API
• Add max_concurrent_shard_requests to Open point in time API
• Add local and flat_settings to Check index templates API
• Add reopen to Update index settings API
• Add resource to Reload search analyzer API
• Add lazy to Roll over to a new index API
• Add cause and create to Simulate index template APIs
• Add inference APIs: Alibaba Cloud AI Search, Amazon Bedrock, Anthropic, Azure AI Studio, Azure OpenAI, Cohere, Elasticsearch, ELSER, Google AI Studio, Google Vertex AI, Hugging Face, Jina AI, Mistral, OpenAI, and Voyage AI

• Add ignore_malformed, script, on_script_error and time_series_dimension to Boolean field
• Add index to GeoShape field
• Add search_inference_id to SemanticText field