v0.33.2

v0.33.2 patch release with bug fixes.

• Check for empty data for add/delete/+async (#5242)
• Generalize Python InputTable class (#5236)
• Update web version 0.68.0 (#5239)
• Update web version 0.67.0 (#5218)
• Update web version 0.66.1 (#5205)
• Update web version 0.66.0 (#5202)
• Fix Python 3.8 compatibility issue (#5235)
• eval string annotation with fn's globals (#5227)
• Address Blink table usage of i, ii, k, and column arrays; address improper memoization sharing between blink and non-blink table copies (#5216)
• Reduce memory usage by using a temporary context. (#5215)
• Fix bugs in moveColumns and renameColumns (#5193)
• Improve WindowCheck memory usage. (#5197)

Full Changelog: v0.33.1...v0.33.2

v0.33.1

84264a5 Bump to 0.33.1
6e48391 Support use of DType instance in UDF annotation (#5200)
4698b9f Fix learn dependency issue (#5201)
589eaae Added py.typed marker for deephaven (#5196)
c4053f0 Added fix for managing context objects (#5195)
63e383e More helpful error when UDF arg type check failed (#5175)
aac191f Added support to read mixed encoded parquet files (#5176)
cd27d4d Update publish-ci to continue on certain failures (#5191)
21e39dc Make 'numba' a soft dependency for the Python server API (#5169)
3fb6fd8 Add optional S3 dependency to py embedded-server (#5180)

Full Changelog: v0.33.0...v0.33.1

v0.33.0

What's Changed

• Run JS API unit tests and integration tests in the Gradle build by @niloc132 in #4988
• Post release bump to 0.33.0 by @stanbrub in #5049
• JS API should expose if a table is refreshing by @niloc132 in #4999
• Eliminate Double Error Notifications in jsapi and Error Log by @nbauernfeind in #5048
• Extract qst-type from qst by @devinrsmith in #5054
• Refactor TypedFunction, ObjectProcessor, and ObjectProcessorFunctions by @devinrsmith in #4992
• Increase GWT test timeouts to avoid CI timeouts by @niloc132 in #5057
• Chunk pools should be easier to emulate in GWT by @niloc132 in #5053
• Extract variable api from ScriptSession, let ScriptSession guard reads by @niloc132 in #4970
• Go Client: Rewrite TestAsOfJoinQuery in query_test.go by @kosak in #5059
• Re-implement ParallelWhere using JobScheduler semantics. by @lbooker42 in #5032
• Add a local copy of build-dependencies.sh for cpp-client, and add a PR check that is kept up to date. by @jcferretti in #5072
• Fix Where performance regression by @lbooker42 in #5074
• Use a Deque to avoid StackOverflow in ImmediateJobScheduler by @nbauernfeind in #5077
• Update web version 0.60.0 by @github-actions in #5078
• Added support to read single parquet file hosted in AWS S3 by @malhotrashivam in #4972
• Fixed a parquet python module failure when s3 extension is not included by @malhotrashivam in #5092
• Fix ReadableByteChannel.read usages by @devinrsmith in #5089
• Submit gradle-aware dependency information to Github by @devinrsmith in #5086
• Update slf4j and logback versions by @devinrsmith in #5102
• Bump commons-compress by @devinrsmith in #5108
• Fixed a bug in reading metadata files by @malhotrashivam in #5103
• Increase one more failing test timeout by @niloc132 in #5098
• Update web version 0.61.0 by @github-actions in #5099
• Fix S3ChannelContext.FragmentState#cancelAndRelease by @devinrsmith in #5095
• Update avro version by @devinrsmith in #5104
• Update xerial snappy-java version by @devinrsmith in #5106
• Bump postgresql version by @devinrsmith in #5107
• Update web version 0.61.1 by @github-actions in #5111
• Add Py version check for using eval_str flag of the inspect.signature() by @jmao-denver in #5083
• Update web version 0.62.0 by @github-actions in #5114
• DEEPHAVEN_VERSION must be set when installing py-client wheel by @niloc132 in #5110
• Add support of 1D arrays in UDF vectorization by @jmao-denver in #5100
• Ensure wildcard imports work against JRE classes by @niloc132 in #5117
• Add endpointOverride S3 support by @devinrsmith in #5087
• Update web version 0.63.0 by @github-actions in #5126
• Add AggWAvg support for rollup() by @lbooker42 in #5121
• Propagate Blink Attribute Downstream of View/UpdateView/Select/UpdateView by @nbauernfeind in #5125
• Bump base image, update numba cutoff to Python 3.12 by @devinrsmith in #5130
• Increase another failing test timeout by @niloc132 in #5134
• Replace usage of trove in Chunk APIs by @niloc132 in #5058
• Deploy @deephaven/jsapi-types during release, provide tools for nightly builds by @niloc132 in #4907
• Reduce public visibility into io.deephaven.parquet.base by @devinrsmith in #5093
• Add RollingFormula feature to UpdateBy by @lbooker42 in #5085
• PythonScopeJpyImpl: Disable Conversion Cache by @nbauernfeind in #5123
• Fix unrendered HTML in py client doc by @jjbrosnan in #5149
• Ensure docs invokes typedoc with correct properties by @devinrsmith in #5146
• Improve S3, parquet, and SeekableChannelsProvider by @devinrsmith in #5137
• Reduce Verbose Logging in Generators by @nbauernfeind in #5155
• Clarify responsibilities when calling this method by @niloc132 in #5152
• Add TrustCustom by @devinrsmith in #5141
• Update web version 0.64.0 by @github-actions in #5156
• Add exception for exponential operators providing non-descending timestamps. by @lbooker42 in #5144
• GRPC support for MultiJoin, improved support for RangeJoin by @lbooker42 in #5153
• Drop Usage of ensureGil from PythonDeephavenSession and Copy Current Scope by @nbauernfeind in #5145
• Introduce TableInitializationException to Propagate JobScheduler Failures by @nbauernfeind in #5165
• When Sorting with a Static Symbol Table use a RowKey or RowPosition ColumnSource by @nbauernfeind in #5154
• Fix test failures from wrapping job scheduler errors in TableInitializationExceptions by @nbauernfeind in #5166
• Use reader's schema for kafka avro deserialization by @devinrsmith in #5167
• Update web version 0.65.0 by @github-actions in #5172
• Provide more context for failed aggregation operators. by @cpwright in #5119

Full Changelog: v0.32.0...v0.33.0

v0.32.1

• 6ab3cd3 - Propagate Blink Attribute Downstream of View/UpdateView/Select/UpdateView (#5125)
• f3ebf39 - Add AggWAvg support for rollup() (#5121)
• c7dc934 - Ensure wildcard imports work against JRE classes (#5117)
• 6f73f76 - Add Py version check for using eval_str flag of the inspect.signature() (#5083)
• d7e3482 - Fix Where performance regression (#5074)
• f12fac4 - Eliminate Double Error Notifications in jsapi and Error Log (#5048)
• 45e14f2 - Re-implement ParallelWhere using JobScheduler semantics. (#5032)

Full Changelog: v0.32.0...v0.32.1

v0.32.0

What's Changed

• feat: Add iframe/widget endpoint for displaying widgets by @mofojed in #4938
• Bump to 0.32.0-SNAPSHOT by @devinrsmith in #4951
• Update image refs to match base image bump to protobuf 3.21.2 by @kosak in #4956
• Correctly allow base64 prefixes for "s/" by @niloc132 in #4958
• Distinguish between np type hints from Py ones by @jmao-denver in #4953
• Bump actions/upload-artifact from 3 to 4 by @dependabot in #4954
• fix: JsPartitionedTable keyColumn and keyColumnTypes ordering by @georgecwan in #4931
• Remove unused/broken tests by @niloc132 in #4961
• Disable Parallel Where for ConditionFilter by @nbauernfeind in #4963
• Move test generation code into README by @devinrsmith in #4965
• Rework server tests to avoid using test sources from other projects by @niloc132 in #4962
• Address dependency and liveness issues in KafkaTools.StreamPartitionedTable by @rcaudy in #4968
• Go Client: Rewrite the AsOfJoin test to use static tables by @kosak in #4971
• Add dictionary (mapping) as a valid argument for new_table() by @jmao-denver in #4966
• Include column name in errors when creating InputColumn by @jmao-denver in #4979
• UpdateGraph: Log Dependency Information On Failure to Proceed by @nbauernfeind in #4980
• Update web version 0.58.0 by @github-actions in #4982
• OperationInitializer usage bug fix and associated cleanups by @rcaudy in #4976
• fix: Add BigInteger and BigDecimal types to MatchFilter.java by @georgecwan in #4912
• Go Client: Update code to match changes in TimeTableRequest, and reenable CI by @kosak in #4990
• HashedRunFinder run-finding improvements by @lbooker42 in #4974
• Go Client: Change TimeTable method to allow a variety of time/period types by @kosak in #4998
• Adjust parameters for floating point tests by @lbooker42 in #4996
• Bump to jetty 11.0.19 by @devinrsmith in #4967
• Add additional JsPlugin documentation by @devinrsmith in #4950
• Bump actions/github-script from 6 to 7 by @dependabot in #4861
• Revise auto-generated doc issues by @mattrunyon in #5008
• PagingChunkSource context protocol documentation, many silly cleanups, hints for channel context integration by @rcaudy in #5000
• Fix pydeephaven-ticking package name and link by @rbasralian in #5013
• Python Client: avoid thread self-join on stop() by @kosak in #5005
• Move OperationInitializer Construction to Dagger Object Graph by @nbauernfeind in #4986
• Fix Server-Side ABORT That Should be CANCELLED by @nbauernfeind in #5018
• Allow ExportObject to be RELEASED When Already Managed by @nbauernfeind in #5014
• Calendar API cleanup by @chipkent in #4201
• docs: Add details about how to debug server by @mofojed in #4927
• Flight: Disable Batching Static Tables by @nbauernfeind in #4985
• Log User-Friendly Error When AutoComplete is not Available by @nbauernfeind in #4316
• Fix broken calendar unit tests over midnight. by @chipkent in #5025
• ReplayTable: Only Fail Table (not the server) on Error by @nbauernfeind in #5020
• Add dependency on jpy>=0.15.0 by @devinrsmith in #5022
• Share PyArrow table data with the server through ByteBuffer by @jmao-denver in #4936
• Calendar API changes so that methods accepting strings return strings by @chipkent in #5029
• fix: Don't memoize retrieval of tables from JsPartitionedTable#getTable by @mofojed in #5009
• Better error logging when calendars do not load properly by @chipkent in #5034
• Support the use of 'from future import annotations' in scripts where Python UDFs are defined. by @jmao-denver in #5033
• Parse legacy calendar XML files. by @chipkent in #5036
• BarrageTable: Use WritableRowRedirectionLockFree in Most Cases by @nbauernfeind in #5038
• Move calendar initialization to server startup by @devinrsmith in #5037
• Add column source reinterpretation for UpdateBy Delta operator. by @lbooker42 in #5043
• Update web version 0.59.0 by @github-actions in #5044
• Fix a few additional failing calendar tests by @devinrsmith in #5046
• Add Groovy container w/ examples and Redpanda by @jjbrosnan in #4673
• BarrageTable: Fix testCoalescingLargeUpdates out-of-memory by @nbauernfeind in #5047

Full Changelog: v0.31.0...v0.32.0

v0.31.0

Note: the docker images for 0.31.0 have changed from java 17 to java 21.

What's Changed

• Post Release Bump to 0.31.0 by @stanbrub in #4793
• C++ Client: improve proto build script test for missing envars by @kosak in #4796
• C++ Client: some cleanups found while working on Windows port by @kosak in #4797
• Add support for MultiXYErrorBarSeries, MultiOHLCSeries figures by @mofojed in #4763
• Properly pass along exceptional state in BatchHandler by @devinrsmith in #4802
• Update protoc and cpp client base images. by @jcferretti in #4805
• Apply typing.Literal whereever appropriate by @jmao-denver in #4803
• Bug fixes and improved support for Parquet TIMESTAMP by @malhotrashivam in #4801
• Update web version 0.54.0 by @github-actions in #4812
• Bump images post 0.30.0 release by @devinrsmith in #4807
• Avoid extra copying of r docs by @niloc132 in #4811
• TableServiceAsyncTest: Eliminate Compiler Overhead Variance; Close QueryCompiler's JavaFileManager by @nbauernfeind in #4808
• JS TreeTable must always track key column values by @niloc132 in #4791
• Ensure LongWrapper is formatted as a Java long by @niloc132 in #4820
• Always export scope objects before performing other interactions by @niloc132 in #4816
• Server logs should be written when integration tests fail by @niloc132 in #4804
• Moved Parquet specific methods out of DateTimeUtils by @malhotrashivam in #4819
• Clarify standard deviation / variance documentation with Bessel's correction by @alexpeters1208 in #4786
• Support calling numpy ufuncs assigned to top-level names in formulas by @jmao-denver in #4759
• change default dtype_backend for to_pandas by @jmao-denver in #4815
• Stop requesting column statistics for preview columns by @niloc132 in #4828
• Add TableDefinition column name helpers by @devinrsmith in #4813
• Use chunks in DynamicWhereFilter, fix a chunk leak, and cleanup some code generally by @rcaudy in #4826
• Fix a docstring error introduced in to_pandas() default dtype_backend change by @jmao-denver in #4833
• ExportObject PUBLISHING State Change Bug by @nbauernfeind in #4835
• README rework by @jjbrosnan in #4209
• Added PR verification step after cherry pick by @stanbrub in #4846
• JS API must support SKIP as an aggregation type by @niloc132 in #4780
• JS API member/documentation cleanup by @niloc132 in #4787
• Add parquet read TableDefinition support by @devinrsmith in #4831
• Fix testing tools NaN comparison and correct new failing tests. by @lbooker42 in #4794
• R HTML docs source code by @alexpeters1208 in #4782
• Apply query scope ctx on formula aggregation by @jmao-denver in #4839
• Add topic, partition, and timestamp column kafka publishing support by @devinrsmith in #4771
• Website integration for R HTML documentation by @alexpeters1208 in #4851
• Invoke typedoc to produce browsable JS API documentation by @niloc132 in #4757
• Correct the path to read generated typedoc when deploying by @niloc132 in #4854
• QueryPerformanceRecorder: Group Batched Operations as a Single Query by @nbauernfeind in #4760
• C++ Client: fix one incorrect mutex unlock, remove one needless unlock by @kosak in #4855
• C++ Client: avoid deprecated Arrow methods, plus conform to coding style by @kosak in #4857
• C++ Client: more avoid deprecated Arrow, more conform to style guide by @kosak in #4858
• Fix URL in pkgdown file by @alexpeters1208 in #4863
• Update web version 0.55.0 by @github-actions in #4864
• Correct R Docsite by @alexpeters1208 in #4867
• QueryPerformanceRecorder: Changes for CorePlus Integration by @nbauernfeind in #4862
• Docs CI Upload Versioned Artifacts by @stanbrub in #4853
• ConstructSnapshot + BarrageMessageProducer: Use Static ConstructSnapshot Fast Path by @nbauernfeind in #4876
• Added offset index based parquet reading support by @malhotrashivam in #4844
• Add in-development version support by @devinrsmith in #4865
• Introduce Blink-To-Append-Only Memoization and Apply to TableLoggers by @nbauernfeind in #4880
• UPDATE: Change timeTable pydoc to match signature by @elijahpetty in #4870
• Add information about not requiring WSL for pip-installed Deephaven by @jjbrosnan in #4873
• Collect JsPlugin instances before starting Jetty by @niloc132 in #4885
• Fix github action release versioning by @devinrsmith in #4886
• Address PartitionedTable transform and partitionedTransform dependency issues by @rcaudy in #4889
• C++ Client: vendor the "fmt" and "date" projects and start to use them by @kosak in #4883
• C++ Client: Arrow deprecation; more style guide by @kosak in #4859
• Provide JsPlugins as Singleton by @devinrsmith in #4892
• Add the optional dependencies parameters in the transform/partitioned_transform methods by @jmao-denver in #4891
• C++ Client: update examples, demos, and docs with style guide, Arrow deprecation by @kosak in #4860
• Add notes on releasing jsapi-types by @niloc132 in #4887
• Added note to pydeephaven example by @elijahpetty in #4895
• C++ client: remove all Stringf/Streamf and related by @kosak in #4900
• Clean up some legacy code and improve build by @rcaudy in #4674
• Add key column and column properties to JsPartitionedTable by @georgecwan in #4789
• auto convert Java values(arrays/scalar) to Numpy ones and convert DH nulls based on the annotations of the params of a Py UDF by @jmao-denver in #4502
• Don't run run_script/_refresh_token under session lock by @jmao-denver in #4906
• fix: Add BigDecimal and BigInteger as numeric aggregation options by @mofojed in #4878
• Bump actions/setup-java from 3 to 4 by @dependabot in #4910
• EventDrivenUpdateGraph. by @cpwright in #4613
• Add a note in C++ client's README.md about build-dependencies.sh version. by @jcferretti in #4926
• Simplify Input Table Interface. by @cpwright in #4923
• Update web version 0.56.0 by @github-actions in #4930
• Docs CI Rework for Versioned Deploy by @stanbrub in #4904
• Fix JS client leaking exported references, clarify dh.Widget API by @niloc132 in #4939
• fix: LongWrapper comparison through value instead of object by @georgecwan in #4940
• Docs CI V2 Rework For Old Rsync by @stanbrub in #4941
• Bump server-base images by @devinrsmith in #4943
• Revert change ...

v0.30.4

• Bump to 0.30.4
• fix: Add BigDecimal and BigInteger as numeric aggregation options (#4878)
• Don't run run_scrip/_refresh_token under lock (#4906)

Full Changelog: v0.30.3...v0.30.4

v0.30.3

Please note, 0.30.2 had issues syncing to Maven Central, so 0.30.3 was cut and released instead.

• Add the optional dependencies parameters in the transform/partitioned_transform methods (#4891)
• Add optional extra dependencies to PartitionedTable.transform and PartitionedTable.p (#4889)
• Apply query scope ctx on formula aggregation (#4839)
• Ensure LongWrapper is formatted as a Java long (#4820)

Full Changelog: v0.30.1...v0.30.3

v0.30.1

• ExportObject PUBLISHING State Change Bug (#4835)
• Bump to 0.30.1
• Stop requesting column statistics for preview columns (#4828)
• Support calling numpy ufuncs assigned to top-level names in formulas (#4759)
• Moved Parquet specific methods out of DateTimeUtils (#4819)
• TableServiceAsyncTest: Eliminate Compiler Overhead Variance; Close QueryCompiler's JavaFileManager (#4808)
• Bug fixes for Parquet DATE and TIME, and improved support for TIMESTAMP (#4801)
• Update protoc and cpp client base images. (#4805)
• Add better asynchronous impl for TableHandleFuture (#4802)

Full Changelog: v0.30.0...v0.30.1

v0.30.0

What's Changed

• Upgrade jetty to 11.0.17 by @devinrsmith in #4617
• Fix incorrect super method invocation in OperationSnapshotControlEx.snapshotCompletedConsistently by @rcaudy in #4668
• Remove the table size comparison based on wrong assumption by @jmao-denver in #4672
• Bump to gradle 7.6.3 by @devinrsmith in #4659
• Post release bump to 0.30.0 by @devinrsmith in #4666
• Fortify numpy array typehint parsing by @jmao-denver in #4671
• Python client: add a simple demo that is similar to the C++ demo "chapter3" by @kosak in #4679
• Fix a bug in the use of liveness_scope in tests by @jmao-denver in #4684
• Documentation for the R Client by @alexpeters1208 in #4610
• Make TableUpdateValidator composable in a sequence of operations. by @cpwright in #4682
• Create proper JS plugins URI by @devinrsmith in #4697
• BarrageSession Subscription/Snapshot Methods now Return Future<Table> by @nbauernfeind in #4676
• Added buffered stream for all parquet writes by @malhotrashivam in #4669
• Update web version 0.50.0 by @github-actions in #4637
• Fix intermittent build failure, ensuring docker test output is still usable by @niloc132 in #4692
• Add ObjectProcessor noop impl and ObjectProcessor tests by @devinrsmith in #4685
• Gradle build target and CI checks for python ticking client by @jcferretti in #4688
• Improve plotting error messages when the input type is bad. by @chipkent in #4698
• Missing TableUpdateValidator unit test. by @cpwright in #4699
• Update nightly testing to include Java 21 by @devinrsmith in #4702
• Support time zone literals by @chipkent in #4704
• AsOfJoin cookie migration bug fix by @lbooker42 in #4706
• Update web version 0.51.0 by @github-actions in #4712
• Barrage Protocol: Fix Writer of byte[] to Use Var Binary Encoding by @nbauernfeind in #4715
• Create a dockerized gradle task to update the generated R client documentation. by @jcferretti in #4713
• Support Optional annotation by @jmao-denver in #4707
• Don't leak the viewport subscription when the viewport data is read by @niloc132 in #4420
• Re-read from start if a read from proc doesn't fill the buffer by @niloc132 in #4716
• Vector Expansion Kernel O(n^2) -> O(n) by @nbauernfeind in #4722
• C++ Client: Keep FlightStreamWriter alive for the duration of the DoExchange by @kosak in #4725
• Fix NullPointerException if connection is closed after server goes away by @mattrunyon in #4727
• Make Barrage URI resolver return a Table by @devinrsmith in #4732
• Update web version 0.52.0 by @github-actions in #4736
• Bump actions/setup-node from 3 to 4 by @dependabot in #4740
• Pydoc formatting fixes by @elijahpetty in #4678
• C++ Client: changes for compatibility with Arrow version 13 by @kosak in #4726
• Eliminate initialization race condition in LiveAttributeMap.immutableAttributes, clean up a few things, and add a unit test by @rcaudy in #4747
• Modify layout of grouping files by @malhotrashivam in #4525
• Add FilterIn support via java-client by @devinrsmith in #4748
• docs: Add better docs to JsPartitionedTable by @mofojed in #4744
• Ensure defaults for float/double generation result in random values by @niloc132 in #4735
• Server-generated column statistics by @niloc132 in #4680
• Update web version 0.53.0 by @github-actions in #4770
• Unit-test thread cleanup by @devinrsmith in #4764
• More explicit exception message for invalid key columns by @rbasralian in #4769
• Added support for Parquet DATE by @malhotrashivam in #4755
• Select: Reuse Results When Possible by @nbauernfeind in #4752
• Python Client import_table Should not also Close the Writer by @nbauernfeind in #4777
• ExportObject: Add onSuccess Callback for use in TableService#batch by @nbauernfeind in #4772
• Added support for Parquet TIME by @malhotrashivam in #4775
• Add java client TableServiceAsync by @devinrsmith in #4756
• Use last lexigraphical key for parquet schema inference by @devinrsmith in #4783
• Correct EmStd empty table initialiation error. by @lbooker42 in #4790

Full Changelog: v0.29.1...v0.30.0