From d785ee9d055836da511f53ed1e2d55b796ee9df4 Mon Sep 17 00:00:00 2001 From: Xuchen Pan <32844285+pan-x-c@users.noreply.github.com> Date: Wed, 8 Jan 2025 12:31:25 +0800 Subject: [PATCH] [Hot Fix] Update Ray version (#532) --- data_juicer/core/ray_data.py | 6 +++--- environments/dist_requires.txt | 2 +- 2 files changed, 4 insertions(+), 4 deletions(-) diff --git a/data_juicer/core/ray_data.py b/data_juicer/core/ray_data.py index b77ed4e69..c5ec8c727 100644 --- a/data_juicer/core/ray_data.py +++ b/data_juicer/core/ray_data.py @@ -15,7 +15,7 @@ from data_juicer.utils.process_utils import calculate_np rd = LazyLoader('rd', 'ray.data') -ds = LazyLoader('ds', 'ray.data.datasource') +ds = LazyLoader('ds', 'ray.data.read_api') def get_abs_path(path, dataset_dir): @@ -235,7 +235,7 @@ def read_json_stream( arrow_open_stream_args: Optional[Dict[str, Any]] = None, meta_provider=None, partition_filter=None, - partitioning=ds.partitioning.Partitioning('hive'), + partitioning=ds.Partitioning('hive'), include_paths: bool = False, ignore_missing_paths: bool = False, shuffle: Union[Literal['files'], None] = None, @@ -245,7 +245,7 @@ def read_json_stream( **arrow_json_args, ) -> rd.Dataset: if meta_provider is None: - meta_provider = ds.file_meta_provider.DefaultFileMetadataProvider() + meta_provider = ds.DefaultFileMetadataProvider() datasource = JSONStreamDatasource( paths, diff --git a/environments/dist_requires.txt b/environments/dist_requires.txt index 90f28ef7a..2ec0908c7 100644 --- a/environments/dist_requires.txt +++ b/environments/dist_requires.txt @@ -1,2 +1,2 @@ -ray>=2.31.0 +ray==2.40.0 redis>=5.0.0