o $io+@sddlZddlZddlZddlZddlmZddlmZddlm Z m Z m Z m Z m Z mZddlZddlZddlZddlmZddlmZmZmZddlmZeeZeedd Gd d d ZdS) N) dataclass)Path)AnyDictListOptionalTupleUnionunflattened_lookup)EXPR_ERROR_PICKLE_FILEEXPR_PROGRESS_FILEEXPR_RESULT_FILE) PublicAPIstable stabilityc@s^eZdZUdZeeeefed<eded<ee ed<eed<dZ eded <dZ ee e deeeffed <dZeejjed <gd Zed eeeeffddZed ejjfddZd$ded efddZd efddZedejjded efddZe d%deeejfdeejjd dfddZe ddd ed!ed edfd"d#Z!dS)&Resulta{The final result of a ML training run or a Tune trial. This is the output produced by ``Trainer.fit``. ``Tuner.fit`` outputs a :class:`~ray.tune.ResultGrid` that is a collection of ``Result`` objects. This API is the recommended way to access the outputs such as: - checkpoints (``Result.checkpoint``) - the history of reported metrics (``Result.metrics_dataframe``, ``Result.metrics``) - errors encountered during a training run (``Result.error``) The constructor is a private API -- use ``Result.from_path`` to create a result object from a directory. Attributes: metrics: The latest set of reported metrics. checkpoint: The latest checkpoint. error: The execution error of the Trainable run, if the trial finishes in error. path: Path pointing to the result directory on persistent storage. This can point to a remote storage location (e.g. S3) or to a local location (path on the head node). The path is accessible via the result's associated `filesystem`. For instance, for a result stored in S3 at ``s3://bucket/location``, ``path`` will have the value ``bucket/location``. metrics_dataframe: The full result dataframe of the Trainable. The dataframe is indexed by iterations and contains reported metrics. Note that the dataframe columns are indexed with the *flattened* keys of reported metrics, so the format of this dataframe may be slightly different than ``Result.metrics``, which is an unflattened dict of the latest set of reported metrics. best_checkpoints: A list of tuples of the best checkpoints and their associated metrics. The number of saved checkpoints is determined by :class:`~ray.train.CheckpointConfig` (by default, all checkpoints will be saved). metricszray.tune.Checkpoint checkpointerrorpathNz pd.DataFramemetrics_dataframebest_checkpoints_storage_filesystem)rrr filesystemrreturncCs|jsdS|jddS)z&The config associated with the result.Nconfig)rgetselfr!K/home/ubuntu/veenaModal/venv/lib/python3.10/site-packages/ray/air/result.pyrIsz Result.configcCs|jptjS)zReturn the filesystem that can be used to access the result path. Returns: pyarrow.fs.FileSystem implementation. )rpyarrowfsLocalFileSystemrr!r!r"rPszResult.filesystemrindentcsddlm}ddlm}fddjD}jr"tjj|d<n|d|dj |d<j rHt | |fddj D|d <d |}d |d fd d | D}d|}d|tj|S)zCConstruct the representation with specified number of space indent.r)BLACKLISTED_KEYS)AUTO_RESULT_KEYScsi|]}|t|qSr!)getattr).0krr!r" ^sz Result._repr..rrcsi|] \}}|vr||qSr!r!)r*r+v)excluder!r"r,isr cs"g|] \}}|d|qS)=r!)r*keyvalue) kws_indentr!r" psz Result._repr..z, z{0}{1}( {2} {0}))ray.tune.experimental.outputr'ray.tune.resultr(_items_to_reprrtype__name__pop type_namersetupdateitemsjoinformat)r r&r'r(shown_attributes cls_indentkwskws_reprr!)r.r4r r"_reprYs(         z Result._reprcCs |jddS)Nr)r&)rFrr!r!r"__repr__vs zResult.__repr__storage_filesystem storage_pathcCs:||}|WdS1swYdS)aOpens a file as an input stream reading all byte content sequentially and decoding read bytes as utf-8 string. Args: storage_filesystem: The filesystem to use. storage_path: The source to open for reading. N)open_input_streamreadalldecode)rHrIfr!r!r"_read_file_as_strys $zResult._read_file_as_strc sddlmddlm}m}m}ddlm}|||\|s*tddt t  }t t  }||r]| |d} dd | D} tj| d d } | rZ| d ni} n)||r{tt| |} | jsx| jd ni} n td t dt dt|ddd} | rԇfdd | D}g}| D]%}| | ||k}|jrtd|d||jrin|jd q|d }tt||}nd}}d}t t }||r|}tj !|}Wdn1swYt"| || ||dS)a3Restore a Result object from local or remote trial directory. Args: path: A path of a trial directory on local or remote storage (ex: s3://bucket/path or /tmp/ray_results). storage_filesystem: A custom filesystem to use. If not provided, this will be auto-resolved by pyarrow. If provided, the path is assumed to be prefix-stripped already, and must be a valid path on the filesystem. Returns: A :py:class:`Result` object of that trial. r) Checkpoint)_exists_at_fs_path_list_at_fs_pathget_fs_and_path)CHECKPOINT_DIR_NAMEz Trial folder z doesn't exist! cSsg|] }|rt|qSr!)jsonloads)r*liner!r!r"r5sz$Result.from_path../)sepz-Failed to restore the Result object: Neither z nor z exists in the trial folder!cSs|jtjjjko |jdS)N checkpoint_)r9r#r$FileType Directory base_name startswith) file_infor!r!r"s z"Result.from_path..) file_filtercs"g|] }t|dqS))rr)ras_posix)r*checkpoint_dir_namerOr$fs_pathr!r"r5s z(Could not find metrics corresponding to z&. These will default to an empty dict.N)rrrrrrr)# ray.trainrOray.train._internal.storagerPrQrRray.train.constantsrS RuntimeErrorrrrcr rNsplitpdjson_normalizeread_csvioStringIOemptyilocto_dictsortedloggerwarningappendlistzipr rJray cloudpickleloadr)clsrrHrPrQrRrSresult_json_fileprogress_csv_filelines json_list metrics_dflatest_metricscheckpoint_dir_names checkpointsrrd#metrics_corresponding_to_checkpointlatest_checkpointrrerror_file_pathrMr!rer" from_paths          zResult.from_pathalpharmetricmodecs|jstd|dvrtd|d|dkrtnt}fdd|jD}|s5tdd |jd ||fd d d dS)aGet the best checkpoint from this trial based on a specific metric. Any checkpoints without an associated metric value will be filtered out. Args: metric: The key for checkpoints to order on. mode: One of ["min", "max"]. Returns: :class:`Checkpoint ` object, or None if there is no valid checkpoint associated with the metric. z,No checkpoint exists in the trial directory!)maxminzUnsupported mode: z$. Please choose from ["min", "max"]!rcs&g|]}t|ddddur|qS)N)defaultr )r* ckpt_inforr!r"r5s z.Result.get_best_checkpoint..zInvalid metric name z-! You may choose from the following metrics: .cst|dS)Nrr )xrr!r"rasz,Result.get_best_checkpoint..)r2r)rrj ValueErrorrrrkeys)r rropvalid_checkpointsr!rr"get_best_checkpoints"  zResult.get_best_checkpoint)r)N)"r: __module__ __qualname____doc__rrstrr__annotations__ Exceptionrrrrrr#r$ FileSystemr8propertyrrintrFrG staticmethodrN classmethodr osPathLikerrrr!r!r!r"rsX #      pr)rorUloggingr dataclassesrpathlibrtypingrrrrrr pandasrlr#rzray._private.dictr ray.air.constantsr r rray.util.annotationsr getLoggerr:rurr!r!r!r"s"