o bi~s@sUddlZddlZddlZddlZddlZddlZddlmZmZddl m Z m Z m Z m Z mZmZddlZddlmZmZmZddlmZddlmZddlmZddlmZdd lmZe redd lm Z e!e"Z#da$d e%d <e&Z'ed dGddde(ej)Z*dZ+dZ,dZ-dZ.dZ/dZ0dZ1dZ2dZ3e4ej56ddZ7ej56de*j8Z9eddZ:dZ;dZdZ?e4e@ej56d d!ZAdZBed"dZCed#dZDed$dZEdZFdZGe4e@ej56d%d!ZHed&dZIed'dZJe4e@ej56d(d)ZKe4ed*d ZLed+dZMdZNd,ZOd-ZPd.ZQdZRed/dZSeTej56d0d1ZUdZVd2ZWd3ZXd4ZYd5ZZdZ[ed6d7Z\d8Z]e4e@ej56d9d!Z^ed:d;Z_edd?ZaeTe%d@<edAdBZbeTe%dC<eeGdDdEdEZcdNdHdIZddFe*fdJdKZeeeGdLdMdMZfefZgdS)ON) dataclassfield) TYPE_CHECKINGAnyDictListOptionalUnion)env_bool env_float env_integer) WORKER_MODE) update_dataset_logger_for_worker) DeveloperAPI)log_once)SchedulingStrategyTExecutionOptionszOptional[DataContext]_default_contextalpha) stabilityc@seZdZdZdZdZdZdS)ShuffleStrategyzkShuffle strategy determines shuffling algorithm employed by operations like aggregate, repartition, etcsort_shuffle_pull_basedsort_shuffle_push_based hash_shuffleN)__name__ __module__ __qualname____doc__SORT_SHUFFLE_PULL_BASEDSORT_SHUFFLE_PUSH_BASED HASH_SHUFFLEr"r"D/home/ubuntu/.local/lib/python3.10/site-packages/ray/data/context.pyrs rii@g?iiTFRAY_DATA_PUSH_BASED_SHUFFLE!RAY_DATA_DEFAULT_SHUFFLE_STRATEGY%RAY_DATA_MAX_HASH_SHUFFLE_AGGREGATORS@SPREADDEFAULTi RAY_DATA_EAGER_FREE0 RAY_DATA_DEFAULT_MIN_PARALLELISM(RAY_DATA_ENABLE_TENSOR_EXTENSION_CASTINGRAY_DATA_USE_ARROW_TENSOR_V2RAY_DATA_TRACE_ALLOCATIONS+RAY_DATA_LOG_INTERNAL_STACK_TRACE_TO_STDOUT%RAY_DATA_RAISE_ORIGINAL_MAP_EXCEPTIONRAY_TQDM1RAY_DATA_DISABLE_PROGRESS_BARS,RAY_DATA_ENABLE_PROGRESS_BAR_NAME_TRUNCATION)AWS Error INTERNAL_FAILUREAWS Error NETWORK_CONNECTIONAWS Error SLOW_DOWN#AWS Error UNKNOWN (HTTP status 503))r7r8r9r:zAWS Error SERVICE_UNAVAILABLEl'RAY_DATA_ENABLE_OP_RESOURCE_RESERVATIONRAY_DATA_OP_RESERVATION_RATIOz0.5u⚠️ u✔️ i&RAY_DATA_DEFAULT_WAIT_FOR_MIN_ACTORS_SRAY_DATA_PER_NODE_METRICS3RAY_DATA_MIN_HASH_SHUFFLE_AGGREGATOR_WAIT_TIME_IN_Si,:RAY_DATA_HASH_SHUFFLE_AGGREGATOR_HEALTH_WARNING_INTERVAL_S4RAY_DATA_DEFAULT_ACTOR_POOL_UTIL_UPSCALING_THRESHOLDg@+DEFAULT_ACTOR_POOL_UTIL_UPSCALING_THRESHOLD6RAY_DATA_DEFAULT_ACTOR_POOL_UTIL_DOWNSCALING_THRESHOLDg?-DEFAULT_ACTOR_POOL_UTIL_DOWNSCALING_THRESHOLDc@s&eZdZUeZeed<eZeed<dS)AutoscalingConfig#actor_pool_util_upscaling_threshold%actor_pool_util_downscaling_thresholdN) rrrrFrJfloat__annotations__rHrKr"r"r"r#rIs  rIreturnrcCsddlm}|S)Nrr)'ray.data._internal.execution.interfacesrrr"r"r#_execution_options_factorys rPcCsJtr tdtjSddtD}t|vs#Jdd|dtdtS)NzqRAY_DATA_PUSH_BASED_SHUFFLE is deprecated, please use RAY_DATA_DEFAULT_SHUFFLE_STRATEGY to set shuffling strategycSsg|]}|qSr"r").0sr"r"r# sz5_deduce_default_shuffle_algorithm..z8RAY_DATA_DEFAULT_SHUFFLE_STRATEGY has to be one of the [,z] (got ))DEFAULT_USE_PUSH_BASED_SHUFFLEloggerwarningrr DEFAULT_SHUFFLE_STRATEGYjoin)vsr"r"r#!_deduce_default_shuffle_algorithms r\cseZdZUdZeZeed<eZ eed<e Z eed<e Z eed<eZeed<eZeed<eedZeed <eZeed <eZeed <d Zeed <eZeZe eed<e!Z"eed<e#Z$eed<dZ%e eed<dZ&e'ed<dZ(e'ed<dZ)e'ed<e*Z+e,ed<e-Z.e,ed<e/Z0eed<e1Z2eed<e3Z4eed<e5Z6eed<e7Z8eed<eZ9eed<e:Z;eed<eZ?eed <dZ@e eed!<eAZBeed"<eCZDeed#<eEZFeed$<eeGdZHd%ed&<eIZJeed'<eKZLeed(<d ZMeed)<eNZOeed*<ePZQeed+<eRZSeTeUed,<eVZWeed-<eXZYeZeeTe[fed.<e\Z]eed/<e^Z_e'ed0<e`Zaeed1<ebZceed2<edZeeed3<d Zfeed4<egZheed5<eiZjeed6<ekZle eed7<ed8d9dZmeTeUed:<enZoeed;<dZpe'ed<<d=Zqe e'ed><dZre eUed?<d@ZseedA<dBdCZtdDeUdEeudFdffdGdH Zvewd^dIdJZxewd_dLdMZyezdFefdNdOZ{e{j|dEedFdfdPdOZ{d`dQeUdReudFeufdSdTZ}dQeUdEeudFdfdUdVZ~dQeUdFdfdWdXZd^dYdZZd[eUdFdfd\d]ZZS)a DataContextaGlobal settings for Ray Data. Configure this class to enable advanced features and tune performance. .. warning:: Apply changes before creating a :class:`~ray.data.Dataset`. Changes made after won't take effect. .. note:: This object is automatically propagated to workers. Access it from the driver and remote workers with :meth:`DataContext.get_current()`. Examples: >>> from ray.data import DataContext >>> DataContext.get_current().enable_progress_bars = False Args: target_max_block_size: The max target block size in bytes for reads and transformations. target_shuffle_max_block_size: The max target block size in bytes for shuffle ops like ``random_shuffle``, ``sort``, and ``repartition``. target_min_block_size: Ray Data avoids creating blocks smaller than this size in bytes on read. This takes precedence over ``read_op_min_num_blocks``. streaming_read_buffer_size: Buffer size when doing streaming reads from local or remote storage. enable_pandas_block: Whether pandas block format is enabled. actor_prefetcher_enabled: Whether to use actor based block prefetcher. autoscaling_config: Autoscaling configuration. use_push_based_shuffle: Whether to use push-based shuffle. pipeline_push_based_shuffle_reduce_tasks: scheduling_strategy: The global scheduling strategy. For tasks with large args, ``scheduling_strategy_large_args`` takes precedence. scheduling_strategy_large_args: Scheduling strategy for tasks with large args. large_args_threshold: Size in bytes after which point task arguments are considered large. Choose a value so that the data transfer overhead is significant in comparison to task scheduling (i.e., low tens of ms). use_polars: Whether to use Polars for tabular dataset sorts, groupbys, and aggregations. eager_free: Whether to eagerly free memory. decoding_size_estimation: Whether to estimate in-memory decoding data size for data source. min_parallelism: This setting is deprecated. Use ``read_op_min_num_blocks`` instead. read_op_min_num_blocks: Minimum number of read output blocks for a dataset. enable_tensor_extension_casting: Whether to automatically cast NumPy ndarray columns in Pandas DataFrames to tensor extension columns. use_arrow_tensor_v2: Config enabling V2 version of ArrowTensorArray supporting tensors > 2Gb in size (off by default) enable_fallback_to_arrow_object_ext_type: Enables fallback to serialize column values not suppported by Arrow natively (like user-defined custom Python classes for ex, etc) using `ArrowPythonObjectType` (simply serializing these as bytes) enable_auto_log_stats: Whether to automatically log stats after execution. If disabled, you can still manually print stats with ``Dataset.stats()``. verbose_stats_logs: Whether stats logs should be verbose. This includes fields such as `extra_metrics` in the stats output, which are excluded by default. trace_allocations: Whether to trace allocations / eager free. This adds significant performance overheads and should only be used for debugging. execution_options: The :class:`~ray.data._internal.execution.interfaces.execution_options.ExecutionOptions` to use. use_ray_tqdm: Whether to enable distributed tqdm. enable_progress_bars: Whether to enable progress bars. enable_progress_bar_name_truncation: If True, the name of the progress bar (often the operator name) will be truncated if it exceeds `ProgressBar.MAX_NAME_LENGTH`. Otherwise, the full operator name is shown. enable_get_object_locations_for_metrics: Whether to enable ``get_object_locations`` for metrics. write_file_retry_on_errors: A list of substrings of error messages that should trigger a retry when writing files. This is useful for handling transient errors when writing to remote storage systems. warn_on_driver_memory_usage_bytes: If driver memory exceeds this threshold, Ray Data warns you. For now, this only applies to shuffle ops because most other ops are unlikely to use as much driver memory. actor_task_retry_on_errors: The application-level errors that actor task should retry. This follows same format as :ref:`retry_exceptions ` in Ray Core. Default to `False` to not retry on any errors. Set to `True` to retry all errors, or set to a list of errors to retry. enable_op_resource_reservation: Whether to reserve resources for each operator. op_resource_reservation_ratio: The ratio of the total resources to reserve for each operator. max_errored_blocks: Max number of blocks that are allowed to have errors, unlimited if negative. This option allows application-level exceptions in block processing tasks. These exceptions may be caused by UDFs (e.g., due to corrupted data samples) or IO errors. Data in the failed blocks are dropped. This option can be useful to prevent a long-running job from failing due to a small number of bad blocks. log_internal_stack_trace_to_stdout: Whether to include internal Ray Data/Ray Core code stack frames when logging to stdout. The full stack trace is always written to the Ray Data log file. raise_original_map_exception: Whether to raise the original exception encountered in map UDF instead of wrapping it in a `UserCodeException`. print_on_execution_start: If ``True``, print execution information when execution starts. s3_try_create_dir: If ``True``, try to create directories on S3 when a write call is made with a S3 URI. wait_for_min_actors_s: The default time to wait for minimum requested actors to start before raising a timeout, in seconds. max_tasks_in_flight_per_actor: Max number of tasks that could be submitted for execution to individual actor at the same time. Note that only up to `max_concurrency` number of these tasks will be executing concurrently while remaining ones will be waiting in the Actor's queue. Buffering tasks in the queue allows us to overlap pulling of the blocks (which are tasks arguments) with the execution of the prior tasks maximizing individual Actor's utilization retried_io_errors: A list of substrings of error messages that should trigger a retry when reading or writing files. This is useful for handling transient errors when reading from remote storage systems. enable_per_node_metrics: Enable per node metrics reporting for Ray Data, disabled by default. memory_usage_poll_interval_s: The interval to poll the USS of map tasks. If `None`, map tasks won't record memory stats. target_max_block_sizetarget_shuffle_max_block_sizetarget_min_block_sizestreaming_read_buffer_sizeenable_pandas_blockactor_prefetcher_enabled)default_factoryautoscaling_configuse_push_based_shuffle_shuffle_strategyT(pipeline_push_based_shuffle_reduce_tasksmax_hash_shuffle_aggregators*min_hash_shuffle_aggregator_wait_time_in_s1hash_shuffle_aggregator_health_warning_interval_sN(max_hash_shuffle_finalization_batch_size3join_operator_actor_num_cpus_per_partition_override;hash_shuffle_operator_actor_num_cpus_per_partition_override=hash_aggregate_operator_actor_num_cpus_per_partition_overridescheduling_strategyscheduling_strategy_large_argslarge_args_threshold use_polarsuse_polars_sort eager_freedecoding_size_estimationmin_parallelismread_op_min_num_blocksenable_tensor_extension_castinguse_arrow_tensor_v2(enable_fallback_to_arrow_object_ext_typeenable_auto_log_statsverbose_stats_logstrace_allocationsrexecution_options use_ray_tqdmenable_progress_barsenable_operator_progress_bars#enable_progress_bar_name_truncation'enable_get_object_locations_for_metricswrite_file_retry_on_errors!warn_on_driver_memory_usage_bytesactor_task_retry_on_errorsop_resource_reservation_enabledop_resource_reservation_ratiomax_errored_blocks"log_internal_stack_trace_to_stdoutraise_original_map_exceptionprint_on_execution_starts3_try_create_dirwait_for_min_actors_smax_tasks_in_flight_per_actorcCsttSN)listDEFAULT_RETRIED_IO_ERRORSr"r"r"r#szDataContext.retried_io_errorsenable_per_node_metrics+override_object_store_memory_limit_fractionmemory_usage_poll_interval_sdataset_logger_idF_enable_actor_pool_on_exit_hookcCs`i|_i|_t|_tjddu}|r+tj j t k}|r&t dr&t dd|_dSd|_dS)N RAY_JOB_ID3ray_data_disable_operator_progress_bars_in_ray_jobszDisabling operator-level progress bars by default in Ray Jobs. To enable progress bars for all operators, set `ray.data.DataContext.get_current().enable_operator_progress_bars = True`.FT) _task_pool_data_task_remote_args _kv_configs.DEFAULT_MAX_NUM_BLOCKS_IN_STREAMING_GEN_BUFFER'_max_num_blocks_in_streaming_gen_bufferosenvirongetrayget_runtime_contextworkermoder rrWinfor)self is_ray_job is_driverr"r"r# __post_init__ s zDataContext.__post_init__namevaluerNcs`|dkr|tkrtdtn|dkrtdtn |dkr'tdt||_t||dS)NrzR`write_file_retry_on_errors` is deprecated. Configure `retried_io_errors` instead.rfT`use_push_based_shuffle` is deprecated, please configure `shuffle_strategy` instead.rszH`use_polars` is deprecated, please configure `use_polars_sort` instead.)"DEFAULT_WRITE_FILE_RETRY_ON_ERRORSwarningswarnDeprecationWarningrtsuper __setattr__)rrr __class__r"r#r/s$zDataContext.__setattr__cCs:ttdur tatWdS1swYdS)aGet or create the current DataContext. When a Dataset is created, the current DataContext will be sealed. Changes to `DataContext.get_current()` will not impact existing Datasets. Examples: .. testcode:: import ray context = ray.data.DataContext.get_current() context.target_max_block_size = 100 * 1024 ** 2 ds1 = ray.data.range(1) context.target_max_block_size = 1 * 1024 ** 2 ds2 = ray.data.range(1) # ds1's target_max_block_size will be 100MB ds1.take_all() # ds2's target_max_block_size will be 1MB ds2.take_all() Developer notes: Avoid using `DataContext.get_current()` in data internal components, use the DataContext object captured in the Dataset and pass it around as arguments. N) _context_lockrr]r"r"r"r# get_currentJs $zDataContext.get_currentcontextcCs"trtj|jkr t|j|adS)zSet the current context in a remote worker. This is used internally by Dataset to propagate the driver context to remote workers used for parallelization. N)rrr)rr"r"r# _set_currentos    zDataContext._set_currentcCs|jr tdtjS|jS)Nr)rfrWrXrr rgrr"r"r#shuffle_strategy~s zDataContext.shuffle_strategycCs ||_dSr)rg)rrr"r"r#r keydefaultcCs|j||S)a Get the value for a key-value style config. Args: key: The key of the config. default: The default value to return if the key is not found. Returns: The value for the key, or the default value if the key is not found. )rr)rrrr"r"r# get_configszDataContext.get_configcCs||j|<dS)zSet the value for a key-value style config. Args: key: The key of the config. value: The value of the config. N)r)rrrr"r"r# set_configszDataContext.set_configcCs|j|ddS)z`Remove a key-value style config. Args: key: The key of the config. N)rpop)rrr"r"r# remove_configszDataContext.remove_configcCs t|S)z)Create a copy of the current DataContext.)copydeepcopyrr"r"r#rrzDataContext.copy dataset_idcCs ||_dS)zSet the current dataset logger id. This is used internally to propagate the current dataset logger id to remote workers. N)r)rrr"r"r#set_dataset_logger_ids z!DataContext.set_dataset_logger_id)rNr])rr]rNNr)rrrrDEFAULT_TARGET_MAX_BLOCK_SIZEr^intrM%DEFAULT_SHUFFLE_TARGET_MAX_BLOCK_SIZEr_DEFAULT_TARGET_MIN_BLOCK_SIZEr`"DEFAULT_STREAMING_READ_BUFFER_SIZEraDEFAULT_ENABLE_PANDAS_BLOCKrbbool DEFAULT_ACTOR_PREFETCHER_ENABLEDrcrrIrerVrfr\rgrrhDEFAULT_MIN_PARALLELISM default_hash_shuffle_parallelism$DEFAULT_MAX_HASH_SHUFFLE_AGGREGATORSrir2DEFAULT_MIN_HASH_SHUFFLE_AGGREGATOR_WAIT_TIME_IN_Srj9DEFAULT_HASH_SHUFFLE_AGGREGATOR_HEALTH_WARNING_INTERVAL_SrkrlrmrLrnroDEFAULT_SCHEDULING_STRATEGYrpr&DEFAULT_SCHEDULING_STRATEGY_LARGE_ARGSrqDEFAULT_LARGE_ARGS_THRESHOLDrrDEFAULT_USE_POLARSrsDEFAULT_USE_POLARS_SORTrtDEFAULT_EAGER_FREEru(DEFAULT_DECODING_SIZE_ESTIMATION_ENABLEDrvrwDEFAULT_READ_OP_MIN_NUM_BLOCKSrx'DEFAULT_ENABLE_TENSOR_EXTENSION_CASTINGryDEFAULT_USE_ARROW_TENSOR_V2rzr{DEFAULT_AUTO_LOG_STATSr|DEFAULT_VERBOSE_STATS_LOGr}DEFAULT_TRACE_ALLOCATIONSr~rPrDEFAULT_USE_RAY_TQDMrDEFAULT_ENABLE_PROGRESS_BARSrr+DEFAULT_ENABLE_PROGRESS_BAR_NAME_TRUNCATIONr/DEFAULT_ENABLE_GET_OBJECT_LOCATIONS_FOR_METRICSrrrrstr)DEFAULT_WARN_ON_DRIVER_MEMORY_USAGE_BYTESr"DEFAULT_ACTOR_TASK_RETRY_ON_ERRORSrr BaseException&DEFAULT_ENABLE_OP_RESOURCE_RESERVATIONr%DEFAULT_OP_RESOURCE_RESERVATION_RATIOrDEFAULT_MAX_ERRORED_BLOCKSr*DEFAULT_LOG_INTERNAL_STACK_TRACE_TO_STDOUTr-DEFAULT_RAY_DATA_RAISE_ORIGINAL_MAP_EXCEPTIONrrDEFAULT_S3_TRY_CREATE_DIRrDEFAULT_WAIT_FOR_MIN_ACTORS_Sr%DEFAULT_MAX_TASKS_IN_FLIGHT_PER_ACTORrrDEFAULT_ENABLE_PER_NODE_METRICSrrrrrrrr staticmethodrrpropertyrsetterrrrrr __classcell__r"r"rr#r]s  s                                          % $    r])rNr)hrenumloggingr threadingr dataclassesrrtypingrrrrrr rray._private.ray_constantsr r r ray._private.workerr ray.data._internal.loggingrray.util.annotationsrray.util.debugrray.util.scheduling_strategiesrrOr getLoggerrrWrrMLockrrEnumrrrMAX_SAFE_BLOCK_SIZE_FACTORMAX_SAFE_ROWS_PER_BLOCK_FACTORrrrrrrrrrVrrYrrrrrrrrrrrrrrrrrrrrrrrrrrrLrr WARN_PREFIX OK_PREFIXLEGACY_DEFAULT_BATCH_SIZErrrrrrrrFrHrIrPr\r]DatasetContextr"r"r"r#s