What's Changed
- Optimize tile of DataFrame.setitem by reducing time of generating chunk meta by @qinxuye in #3140
- Increase the default value of alru cache max size by @zhongchun in #3146
- Support scipy special function with tuple output by @RandomY-2 in #3139
- Fix
DAG.to_dot
when reducers have multiple outputs by @chaokunyang in #3150 - Fix deserializing RandomStateField when its value is None by @chaokunyang in #3149
- Patch pandas magic functions to allow reverse operands by @wjsi in #3155
- Run flaky test
test_load_third_party_modules
separately by @chaokunyang in #3162 - Manually install cri-dockerd before installing kubernetes by @wjsi in #3166
- [Shuffle] Add
n_mappers
andn_reducers
toShuffleProxy
by @chaokunyang in #3160 - [Ray] task based shuffle for ray by @chaokunyang in #3040
- Add support for
{DataFrame,Series}.align
by @wjsi in #3147 - Integrate remaining error functions and fresnel integrals except
fresnel_zeros
by @RandomY-2 in #3172 - Improve numexpr fusion by @fyrestone in #3177
- Ensure key is a valid Python identifier by @fyrestone in #3190
- Bump terser from 5.7.1 to 5.14.2 in web component by @dependabot in #3194
- Implement airy functions (except the
ai_zeros
andbi_zeros
functions) by @shantam-8 in #3195 - Disable version updates for dependabot by @wjsi in #3203
- [Ray] Fix ray memory leak by @fyrestone in #3184
- [Ray] Support reducer has inputs which isn't mapper by @chaokunyang in #3206
- Refine UT and logs by @fyrestone in #3204
- release actor lock when set_subtask_result by @chaokunyang in #3210
- Refine apply key generation by @chaokunyang in #3208
- fix remove mapper data by @chaokunyang in #3214
- [Ray] Configurable subtask num_cpus by @fyrestone in #3207
- Fix versionner compatibility with PEP600 by @chaokunyang in #3223
- Support get mappers data without index/mapperids by @chaokunyang in #3222
- [Ray] RayExecutionContext.get_chunk_meta from meta service by @fyrestone in #3212
- [Ray] Share RayTaskState across tasks by @fyrestone in #3219
- [Shuffle] Support shuffle operands mapper whose outputs aren't mapper blocks by @chaokunyang in #3228
- Apply Operand Closure clean up by @vcfgv in #3205
- Fix dataframe sort_values with multiple ascendings bug in pandas < 1.4 by @fyrestone in #3234
- Lifecycle gc task service by @fyrestone in #3230
- Fix dataframe loc with slice returns incorrect results by @fyrestone in #3241
- Fix dataframe setitem bugs when partial indexes exist in target dataframe by @fyrestone in #3240
- [Shuffle] isolate mappers in different subtasks for fetch_by_index mode by @chaokunyang in #3239
- TypeDispatcher support one type multiple serializers by @fyrestone in #3242
- [Shuffle] Skip store shuffle object refs to reduce meta overhead by @chaokunyang in #3209
- [ray] Support scheduling ray tasks in Ray oscar deploy backend by @chaokunyang in #3165
- Dump subtask graph for all backends by @fyrestone in #3245
- [Metrics] Fix metrics and docs by @zhongchun in #3233
- Remove storage service from supervisor by @vcfgv in #3254
- Fix optimization rule memory leak by @fyrestone in #3246
- fsspec integration by @hekaisheng in #3253
- [Ray] Enable CI of mars/dataframe for Ray DAG by @fyrestone in #3250
- Fix minikube installation by @hekaisheng in #3244
- Implements scipy.stats.rankdata by @shantam-8 in #3218
- Add S3 support by @fyrestone in #3258
- Fix tensor frexp by @fyrestone in #3259
- Optimize the display of task process bar by @zhongchun in #3264
- [Ray] Optimize ray executor submit subtask by @fyrestone in #3271
- [Ray] Enable CI of mars/learn for Ray DAG by @fyrestone in #3261
- [Ray] Enable CI of mars/tensor for Ray DAG by @fyrestone in #3275
- Compatible with pandas 1.5.0 by @hekaisheng in #3276
- Remove skip_ray_dag mark for raydataset tests by @vcfgv in #3255
- MapChunk Operand Closure and Callable cleanup by @vcfgv in #3238
- [Ray] Spread scheduling subtasks with empty dependencies by @fyrestone in #3281
- Speedup mars deserialization by new by @chaokunyang in #3283
- A cython-based ordered_set to speedup
discard
operation by @chaokunyang in #3277 - Optimize concat by @fyrestone in #3286
- Fix
md.concat
error when there are same fetch chunk data by @zhongchun in #3285 - [Ray] Improve Ray executor GC by @fyrestone in #3287
- Fix some CI issues by @hekaisheng in #3296
- [Ray] Implement Ray executor subtask GC by @fyrestone in #3294
- [Ray] Add metrics for Ray executor by @fyrestone in #3295
- Bump up required vineyard version to address the CI failure. by @sighingnow in #3298
- [Operand] support loc setitem by @chaokunyang in #3291
- [Ray] Support worker_mem for ray executor by @fyrestone in #3300
- Fix duplicate execution by @fyrestone in #3301
- Fix CI by @hekaisheng in #3306
- [Ray] Basic slow subtask detection by @fyrestone in #3305
- Fix stats tests and pin sphinx version by @hekaisheng in #3313
- Fix s3 client kwargs by @fyrestone in #3316
- Update Mars on Ray doc by @fyrestone in #3311
Full Changelog: v0.10.0a1...v0.10.0