Tiledb-ml

What's Changed
* Updates for Tensorflow 2.9+ & Pytorch 1.13 by gsakkis in https://github.com/TileDB-Inc/TileDB-ML/pull/195
* Update mypy to 0.991 by gsakkis in https://github.com/TileDB-Inc/TileDB-ML/pull/197
* Gsk/ch21380/upgrade ml model schema by georgeSkoumas in https://github.com/TileDB-Inc/TileDB-ML/pull/196
* Moderate model refactoring by gsakkis in https://github.com/TileDB-Inc/TileDB-ML/pull/199


**Full Changelog**: https://github.com/TileDB-Inc/TileDB-ML/compare/v0.8.0...v0.9.0

What's Changed
* Introduce PyTorch DataPipes by gsakkis in https://github.com/TileDB-Inc/TileDB-ML/pull/173
* Refactor models design and add TB as extra array by ktsitsi in https://github.com/TileDB-Inc/TileDB-ML/pull/175
* [Hotfix] Fixed example notebooks by ktsitsi in https://github.com/TileDB-Inc/TileDB-ML/pull/176
* Fix PyTorchTileDBDataLoader to be iterable more than once by gsakkis in https://github.com/TileDB-Inc/TileDB-ML/pull/177
* Bump pytorch version to 1.12 by gsakkis in https://github.com/TileDB-Inc/TileDB-ML/pull/178
* Test _csr_to_coo_collate by gsakkis in https://github.com/TileDB-Inc/TileDB-ML/pull/180
* Default to TensorKind.SPARSE_COO for sparse tensors by gsakkis in https://github.com/TileDB-Inc/TileDB-ML/pull/181
* Add support for slicing on non-key indices [final] by gsakkis in https://github.com/TileDB-Inc/TileDB-ML/pull/184
* Unbatched data loading support for PyTorchTileDBDataLoader by gsakkis in https://github.com/TileDB-Inc/TileDB-ML/pull/183
* Reduce parametric tests by gsakkis in https://github.com/TileDB-Inc/TileDB-ML/pull/185
* Remove SparseData by gsakkis in https://github.com/TileDB-Inc/TileDB-ML/pull/186
* Support pytorch 1.11 by gsakkis in https://github.com/TileDB-Inc/TileDB-ML/pull/188
* Support reading dense tensors from sparse arrays by gsakkis in https://github.com/TileDB-Inc/TileDB-ML/pull/187
* Package _tensor_schema by gsakkis in https://github.com/TileDB-Inc/TileDB-ML/pull/189
* [Enhancement] Update Pull Request Guide by georgeSkoumas in https://github.com/TileDB-Inc/TileDB-ML/pull/182
* Add support for slicing on non key indices [sparse arrays] by gsakkis in https://github.com/TileDB-Inc/TileDB-ML/pull/190
* Add support for slicing on key dimension [sparse arrays] by gsakkis in https://github.com/TileDB-Inc/TileDB-ML/pull/191
* Add ConstrainedPartitionsIntRange by gsakkis in https://github.com/TileDB-Inc/TileDB-ML/pull/192
* Improve DenseTensorSchema.key_range by gsakkis in https://github.com/TileDB-Inc/TileDB-ML/pull/193
* Update precommit-hooks by gsakkis in https://github.com/TileDB-Inc/TileDB-ML/pull/194


**Full Changelog**: https://github.com/TileDB-Inc/TileDB-ML/compare/v0.7.0...v0.8.0

What's Changed
* Fix ArrayParams imports by georgeSkoumas in https://github.com/TileDB-Inc/TileDB-ML/pull/162
* Array params import to readers by ktsitsi in https://github.com/TileDB-Inc/TileDB-ML/pull/163
* Take into account var-sized fields in SparseTensorSchema.max_partition_weight by gsakkis in https://github.com/TileDB-Inc/TileDB-ML/pull/164
* Bump minimum tensorflow to 2.6 and pytorch to 1.10 by gsakkis in https://github.com/TileDB-Inc/TileDB-ML/pull/168
* pass dataloader params in Dataloader, pytorch tests update by georgeSkoumas in https://github.com/TileDB-Inc/TileDB-ML/pull/165
* Add ArrayParams.tensor_kind by gsakkis in https://github.com/TileDB-Inc/TileDB-ML/pull/167
* Refactor parametric test utils by gsakkis in https://github.com/TileDB-Inc/TileDB-ML/pull/169
* Pass Through Arguments Tensorflow by georgeSkoumas in https://github.com/TileDB-Inc/TileDB-ML/pull/170
* Avoid intermediate `sparse.COO` instances if possible by gsakkis in https://github.com/TileDB-Inc/TileDB-ML/pull/172
* Read from sparse TileDB arrays to ragged tensors by gsakkis in https://github.com/TileDB-Inc/TileDB-ML/pull/171
* CI for notebooks by gsakkis in https://github.com/TileDB-Inc/TileDB-ML/pull/174


**Full Changelog**: https://github.com/TileDB-Inc/TileDB-ML/compare/v0.6.0...v0.7.0

What's Changed
* [Hotfix] Re-run notebook examples for warnings and errors by ktsitsi in https://github.com/TileDB-Inc/TileDB-ML/pull/143
* Refactor reader tests by gsakkis in https://github.com/TileDB-Inc/TileDB-ML/pull/146
* Custom key dimension by gsakkis in https://github.com/TileDB-Inc/TileDB-ML/pull/145
* Adding early stopping callback to model notebooks by ktsitsi in https://github.com/TileDB-Inc/TileDB-ML/pull/144
* Support integer dimensions with non-zero based domain by gsakkis in https://github.com/TileDB-Inc/TileDB-ML/pull/147
* Determine default buffer size for sparse arrays by gsakkis in https://github.com/TileDB-Inc/TileDB-ML/pull/148
* Non-empty domain instead of full domain by gsakkis in https://github.com/TileDB-Inc/TileDB-ML/pull/149
* Refactor TensorSchema by gsakkis in https://github.com/TileDB-Inc/TileDB-ML/pull/150
* Privatization of models/base and PytorchTileDBDataset by ktsitsi in https://github.com/TileDB-Inc/TileDB-ML/pull/151
* [CI] Pin protobuf by gsakkis in https://github.com/TileDB-Inc/TileDB-ML/pull/152
* Generate tensors for dimensions by gsakkis in https://github.com/TileDB-Inc/TileDB-ML/pull/153
* InclusiveRange abstraction by gsakkis in https://github.com/TileDB-Inc/TileDB-ML/pull/154
* Refactor data loaders to use InclusiveRange by gsakkis in https://github.com/TileDB-Inc/TileDB-ML/pull/155
* Remove `buffer_bytes` parameter by gsakkis in https://github.com/TileDB-Inc/TileDB-ML/pull/156
* Support non-integer key dimension by gsakkis in https://github.com/TileDB-Inc/TileDB-ML/pull/157
* Introducing ArrayParams as TF-PyTorch readers argument by ktsitsi in https://github.com/TileDB-Inc/TileDB-ML/pull/158
* Update precommit hooks by gsakkis in https://github.com/TileDB-Inc/TileDB-ML/pull/160
* Make TensorSchema a dataclass by gsakkis in https://github.com/TileDB-Inc/TileDB-ML/pull/159
* Read any number of arrays by gsakkis in https://github.com/TileDB-Inc/TileDB-ML/pull/161


**Full Changelog**: https://github.com/TileDB-Inc/TileDB-ML/compare/v0.5.0...v0.6.0

What's Changed
* Implement parallelized TensorflowTileDBDataset by gsakkis in https://github.com/TileDB-Inc/TileDB-ML/pull/132
* Expose TileDB-ML version at runtime by georgeSkoumas in https://github.com/TileDB-Inc/TileDB-ML/pull/134
* Bump tiledb to 0.14+ by gsakkis in https://github.com/TileDB-Inc/TileDB-ML/pull/135
* Separate tensor generation for x and y arrays by gsakkis in https://github.com/TileDB-Inc/TileDB-ML/pull/133
* scipy.sparse for 2D arrays and optional CSR sparse tensors by gsakkis in https://github.com/TileDB-Inc/TileDB-ML/pull/136
* Always generate (x,y) tuples by gsakkis in https://github.com/TileDB-Inc/TileDB-ML/pull/137
* Improve dependency versioning, add CI for multiple dependency versions, fix support for older ML versions by gsakkis in https://github.com/TileDB-Inc/TileDB-ML/pull/139
* Τensorboard Callback Support by ktsitsi in https://github.com/TileDB-Inc/TileDB-ML/pull/138
* Pytorch tensorboard support save and load by ktsitsi in https://github.com/TileDB-Inc/TileDB-ML/pull/141
* Extract tensorboard callback logic in a separate module by gsakkis in https://github.com/TileDB-Inc/TileDB-ML/pull/142
* Support tensorflow>2.5 by gsakkis in https://github.com/TileDB-Inc/TileDB-ML/pull/140


**Full Changelog**: https://github.com/TileDB-Inc/TileDB-ML/compare/v0.4.0...v0.5.0

What's Changed
* Return correctly sized sparse tensors by gsakkis in https://github.com/TileDB-Inc/TileDB-ML/pull/116
* Adding unit tests for read sparse order when batch_shuffle=False by ktsitsi in https://github.com/TileDB-Inc/TileDB-ML/pull/115
* Model API and Model Notebooks Update by georgeSkoumas in https://github.com/TileDB-Inc/TileDB-ML/pull/114
* Separate buffer_size for x/y arrays by gsakkis in https://github.com/TileDB-Inc/TileDB-ML/pull/117
* Specify buffer size in bytes instead of rows by gsakkis in https://github.com/TileDB-Inc/TileDB-ML/pull/118
* Reenable batch shuffling by gsakkis in https://github.com/TileDB-Inc/TileDB-ML/pull/119
* Single shuffle option by gsakkis in https://github.com/TileDB-Inc/TileDB-ML/pull/120
* Better default buffer_size for dense arrays by gsakkis in https://github.com/TileDB-Inc/TileDB-ML/pull/121
* Use get_max_buffer_size() for dense arrays even when buffer_bytes is specified by gsakkis in https://github.com/TileDB-Inc/TileDB-ML/pull/122
* Add PyTorchTileDBDataLoader extending torch.utils.data.DataLoader by gsakkis in https://github.com/TileDB-Inc/TileDB-ML/pull/123
* Convert numpy arrays to dense tensors implicitly by gsakkis in https://github.com/TileDB-Inc/TileDB-ML/pull/125
* Remove batch size from tensor generator by gsakkis in https://github.com/TileDB-Inc/TileDB-ML/pull/124
* Replace shuffle parameter with shuffle buffer size by gsakkis in https://github.com/TileDB-Inc/TileDB-ML/pull/126
* Faster batching for PyTorchTileDBDataLoader by gsakkis in https://github.com/TileDB-Inc/TileDB-ML/pull/127
* Drop ThreadPoolExecutor for reading from x and y by gsakkis in https://github.com/TileDB-Inc/TileDB-ML/pull/128
* Split _batch_utils module by gsakkis in https://github.com/TileDB-Inc/TileDB-ML/pull/129
* Support nD-dim sparse arrays by gsakkis in https://github.com/TileDB-Inc/TileDB-ML/pull/130
* Add prefetch parameter to PyTorchTileDBDataLoader and TensorflowTileDBDataset by gsakkis in https://github.com/TileDB-Inc/TileDB-ML/pull/131


**Full Changelog**: https://github.com/TileDB-Inc/TileDB-ML/compare/v0.3.0...v0.4.0