Updating package version number and release docs

Added release notes and migration guide

docs/autodoc/index.rst

@@ -24,6 +24,7 @@ training frameworks.
    benchmarks_tutorial_include
    troubleshoot_include
    release_notes_include
+   migration_guide_include
 
 
 Indices and tables

docs/autodoc/migration_guide_include.rst

@@ -0,0 +1,3 @@
+Migration notes
+===============
+.. include:: ../migrating-0.5.0.rst

docs/migrating-0.5.0.rst

@@ -0,0 +1,45 @@
+.. inclusion-marker-start-do-not-remove
+
+==================
+To petastorm 0.5.0
+==================
+
+Petastorm 0.5.0 has some breaking changes from previous versions. These include:
+
+- Users should use :func:`~petastorm.reader.make_reader`, instead of instantiating :class:`~petastorm.reader.Reader`
+  directly to create a new instances
+- It is still possible (although discouraged in most cases) to instantitate :class:`~petastorm.reader.Reader`. Some of
+  its argument has changed.
+
+Use :func:`~petastorm.reader.make_reader` to instantiate a reader instance
+--------------------------------------------------------------------------
+
+Use :func:`~petastorm.reader.make_reader` to create a new instance of a reader. :func:`~petastorm.reader.make_reader`
+takes arguments that are almost similar to constructor arguments of :class:`~petastorm.reader.Reader`. The following
+list enumerates the differences:
+
+- ``reader_pool``: takes one of the strings: ``'thread'``, ``'process'``, ``'dummy'``
+  (instead of ``ThreadPool()``, ``ProcessPool()`` and ``DummyPool()`` object instances). Pass number of workers using
+  ``workers_count`` argument.
+- ``training_partition`` and ``num_training_partitions`` were renamed into ``cur_shard`` and ``shard_count``.
+- ``shuffle`` and ``shuffle_options`` were replaced by ``shuffle_row_groups=True, shuffle_row_drop_partitions=1``
+
+.. code-block:: python
+
+    from petastorm.reader import Reader
+    reader = Reader(dataset_url,
+                    reader_pool=ThreadPool(5),
+                    training_partition=1, num_training_partitions=5,
+                    shuffle_options=ShuffleOptions(shuffle_row_groups=False))
+
+To:
+
+.. code-block:: python
+
+    from petastorm import make_reader
+    reader = make_reader(dataset_url,
+                         reader_pool='thread',
+                         workers_count=5,
+                         cur_shard=1, shard_count=5,
+                         shuffle_row_groups=False)
+

docs/release-notes.rst

@@ -4,6 +4,30 @@
 Release notes
 =============
 
+Release 0.5.0
+=============
+
+Breaking changes
+----------------
+- :func:`~petastorm.reader.make_reader` should be used to create new instance of a reader.
+- It is still possible, but not recommended to use :class:`~petastorm.reader.Reader` in most cases. Its constructor arguments
+  has changed:
+  -- ``training_partition`` and ``num_training_partitions`` were renamed into ``cur_shard`` and ``shard_count``.
+  -- ``shuffle`` and ``shuffle_options`` were replaced by ``shuffle_row_groups=True, shuffle_row_drop_partitions=1``
+  -- ``sequence`` argument was removed
+
+
+New features and bug fixes
+--------------------------
+- It is possible to read non-Petastorm Parquet datasets (created externally to Petastorm). Currently most of the
+  scalar types are supported.
+- Support s3 as the protocol in a dataset url strings (e.g. 's3://...')
+- PyTorch: support collating decimal scalars
+- PyTorch: promote integer types that are not supported by PyTorch to the next larger integer types that is supported
+  (e.g. int8 -> int16). Booleans are promoted to uint8.
+- Support running ``petastorm-generate-metadata.py`` on datasets created by Hive.
+- Fix incorrect dataset sharding when using Python 3.
+
 Release 0.4.3
 =============
 

petastorm/__init__.py

@@ -14,4 +14,4 @@
 
 from petastorm.reader import make_reader, make_batch_reader  # noqa: F401
 
-__version__ = '0.4.3'
+__version__ = '0.5.0rc0'