Merge pull request #1632 from AdeelH/prerelease-0.20.1

Update changelog for 0.20.1 release

docs/changelog.rst

@@ -1,9 +1,30 @@
 CHANGELOG
 =========
 
-Raster Vision 0.20.0
+Raster Vision 0.20.1
 --------------------
 
+Fixes
+~~~~~
+
+* Do not install ``rastervision_gdal_vsi`` by default (`#1622 <https://github.com/azavea/raster-vision/pull/1622>`__)
+* Do not set ``cfg.model.pretrained=False`` in ``Learner.from_model_bundle()`` (`#1626 <https://github.com/azavea/raster-vision/pull/1626>`__)
+* Fix docker build errors (`#1629 <https://github.com/azavea/raster-vision/pull/1629>`__)
+* Documentation:
+
+  * Improve docstrings for most commonly used classes and configs (`#1630 <https://github.com/azavea/raster-vision/pull/1630>`__)
+  * Minor textual fixes for the pre-chipped datasets tutorial (`#1623 <https://github.com/azavea/raster-vision/pull/1623>`__)
+  * Add comment about password for the ISPRS Potsdam dataset (`#1627 <https://github.com/azavea/raster-vision/pull/1627>`__)
+
+* README:
+
+  * fix broken links (`#1608 <https://github.com/azavea/raster-vision/pull/1608>`__)
+  * make CV-tasks image slightly smaller (`#1624 <https://github.com/azavea/raster-vision/pull/1624>`__)
+
+
+Raster Vision 0.20
+------------------
+
 This release brings major improvements to Raster Vision's **usability** as well as its **usefulness**.
 
 Whereas previously Raster Vision was a **framework** where users could configure a *pipeline* and then let it run, it is now *also* a **library** from which users can pick individual components and use them to build new things.
@@ -12,7 +33,7 @@ We have also significantly improved the documentation. Most notably, it now cont
 
 In terms of features, some highlights are:
 
-- Support for multiband imagery, introduced in v0.13 for semantic segmentation, is now also available in chip classification and object detection. (`#1345 <https://github.com/azavea/raster-vision/pull/1345>`__)
+- Support for multiband imagery, introduced in v0.13 for semantic segmentation, is now also available for chip classification and object detection. (`#1345 <https://github.com/azavea/raster-vision/pull/1345>`__)
 - Improved data fusion: the :class:`~rastervision.core.data.raster_source.multi_raster_source.MultiRasterSource` can now combine :class:`RasterSources <rastervision.core.data.raster_source.raster_source.RasterSource>` with varying extents and resolutions. (`#1308 <https://github.com/azavea/raster-vision/pull/1308>`__)
 - You can now discard edges of predicted chips in semantic segmentation in order to reduce boundary artifacts (`#1486 <https://github.com/azavea/raster-vision/pull/1486>`__). This can be used *in addition* to the `previously introduced <https://github.com/azavea/raster-vision/pull/1057>`__ ability to average overlapping regions in adjacent chips.
 - Progress-bars will now be shown for all downloads and uploads as well as other time-consuming operations that take longer than 5 seconds.
@@ -28,7 +49,7 @@ Features
 - Allow ``MultiRasterSource`` to read from sub raster sources with non-identical extents and resolutions (`#1308 <https://github.com/azavea/raster-vision/pull/1308>`__)
 - Allow discarding edges of predicted chips in semantic segmentation (`#1486 <https://github.com/azavea/raster-vision/pull/1486>`__)
 - Add numpy-like array indexing and slicing to ``RasterSource`` and ``LabelSource`` (`#1470 <https://github.com/azavea/raster-vision/pull/1470>`__)
-- Make ``RandomWindowGeoDataset`` more efficient when sampling chips from scenes with AOIs (`#1225 <https://github.com/azavea/raster-vision/pull/1225>`__)
+- Make ``RandomWindowGeoDataset`` more efficient when sampling chips from scenes with sparse AOIs (`#1225 <https://github.com/azavea/raster-vision/pull/1225>`__)
 - Add support for Albumentations' lambda transforms (`#1368 <https://github.com/azavea/raster-vision/pull/1368>`__)
 - Provide grouping mechanism for scenes and use it in the ``analyze`` and ``eval`` stages (`#1375 <https://github.com/azavea/raster-vision/pull/1375>`__)
 - Update STAC-reading functionality to make it compatible with STAC v1.0.* (`#1243 <https://github.com/azavea/raster-vision/pull/1243>`__)
@@ -62,19 +83,19 @@ Fixes
 
 - Speed up ``RGBClassTransformer`` by an order of magnitude (`#1485 <https://github.com/azavea/raster-vision/pull/1485>`__)
 - Fix ``rastervision_pipeline`` entry point to ensure commands from other plugins are available (`#1250 <https://github.com/azavea/raster-vision/pull/1250>`__)
-- Make the semantic segmentation integration test more deterministic (`#1261 <https://github.com/azavea/raster-vision/pull/1261>`__)
-- Misc. fixes (`#1260 <https://github.com/azavea/raster-vision/pull/1260>`__, `#1281 <https://github.com/azavea/raster-vision/pull/1281>`__, `#1453 <https://github.com/azavea/raster-vision/pull/1453>`__)
 - Fix incorrect F1 scores when aggregating evals for scenes in the eval stage (`#1386 <https://github.com/azavea/raster-vision/pull/1386>`__)
 - Fix bug in semantic segmentation prediction output paths (`#1354 <https://github.com/azavea/raster-vision/pull/1354>`__)
 - Do not zero out null class pixels when creating semantic segmentation training chips (`#1556 <https://github.com/azavea/raster-vision/pull/1556>`__)
 - Fix a bug in ``DataConfig`` validation and refactor ``ClassConfig`` (`#1436 <https://github.com/azavea/raster-vision/pull/1436>`__)
 - Fix `#1052 <https://github.com/azavea/raster-vision/pull/1052>`__ (`#1451 <https://github.com/azavea/raster-vision/pull/1451>`__)
 - Fix `#991 <https://github.com/azavea/raster-vision/pull/991>`__ and `#1452 <https://github.com/azavea/raster-vision/pull/1452>`__ (`#1484 <https://github.com/azavea/raster-vision/pull/1484>`__)
 - Fix `#1430 <https://github.com/azavea/raster-vision/pull/1430>`__ (`#1495 <https://github.com/azavea/raster-vision/pull/1495>`__)
+- Misc. fixes (`#1260 <https://github.com/azavea/raster-vision/pull/1260>`__, `#1281 <https://github.com/azavea/raster-vision/pull/1281>`__, `#1453 <https://github.com/azavea/raster-vision/pull/1453>`__)
 
 Development/maintenance
 ~~~~~~~~~~~~~~~~~~~~~~~
 
+- Make the semantic segmentation integration test more deterministic (`#1261 <https://github.com/azavea/raster-vision/pull/1261>`__)
 - Migrate from Travis to GitHub Actions (`#1218 <https://github.com/azavea/raster-vision/pull/1218>`__)
 - Add Github issue templates (`#1242 <https://github.com/azavea/raster-vision/pull/1242>`__, `#1288 <https://github.com/azavea/raster-vision/pull/1288>`__, `#1420 <https://github.com/azavea/raster-vision/pull/1420>`__)
 - Switch from Gitter to Github Discussions (`#1464 <https://github.com/azavea/raster-vision/pull/1464>`__, `#1465 <https://github.com/azavea/raster-vision/pull/1465>`__)

docs/release.rst

@@ -11,7 +11,7 @@ Minor or Major Version Release
 #.  It's a good idea to update any major dependencies before the release.
 #.  Update the docs if needed. See the `docs README <{{ repo }}/docs/README.md>`__ for instructions.
 #.  Checkout the ``master`` branch, re-build the docker image (``docker/build``), and push it to ECR (``docker/ecr_publish``).
-#.  Execute all notebooks in the `tutorial notebooks <{{ repo }}/docs/usage/tutorials/>` and make sure they work correctly. Do not commit output changes unless code behavior has changed.
+#.  Execute all `tutorial notebooks <{{ repo }}/docs/usage/tutorials/>`__ and make sure they work correctly. Do not commit output changes unless code behavior has changed.
 #.  Run all :ref:`rv examples` and check that evaluation metrics are close to the scores from the last release. (For each example, there should be a link to a JSON file with the evaluation metrics from the last release.) This stage often uncovers bugs, and is the most time consuming part of the release process. There is a `script <{{ repo_examples }}/test.py>`__ to help run the examples and collect their outputs. See the associated `README <{{ repo_examples }}/README.md>`__ for details.
 #.  Collect all model bundles, and check that they work with the ``predict`` command and sanity check output in QGIS.
 #.  Update the :ref:`model zoo` by uploading model bundles and sample images to the right place on S3. If you use the ``collect`` command (`described here <{{ repo_examples }}/README.md>`__), you should be able to sync the ``collect_dir`` to ``s3://azavea-research-public-data/raster-vision/examples/model-zoo-<version>``.
@@ -116,6 +116,7 @@ Minor or Major Version Release
         twine upload dist/*
 
 #.  Announce new release in our `forum <https://github.com/azavea/raster-vision/discussions>`_, and with blog post if it's a big release.
+#.  Make a PR to the master branch that updates the version number to the next development version. For example, if the last release was ``0.20.1``, update the version
 
 Bug Fix Release
 -----------------

rastervision_pytorch_backend/rastervision/pytorch_backend/examples/README.md

@@ -25,8 +25,8 @@ Some examples require preprocessing data. Remember to upload that first. If unch
 
 ```sh
 aws s3 cp \
-s3://raster-vision/examples/0.13/processed-data/ \
 s3://raster-vision/examples/0.20/processed-data/ \
+s3://raster-vision/examples/0.20.1/processed-data/ \
 --recursive
 ```
 
@@ -52,16 +52,16 @@ This currently only compares the respective `eval.json`'s, but can be extended i
 ```sh
 python "rastervision_pytorch_backend/rastervision/pytorch_backend/examples/test.py" \
 compare \
---root_uri_old "s3://raster-vision/examples/0.13/output/spacenet-rio-cc/" \
---root_uri_new "s3://raster-vision/examples/0.20/output/spacenet-rio-cc/"
+--root_uri_old "s3://raster-vision/examples/0.20/output/spacenet-rio-cc/" \
+--root_uri_new "s3://raster-vision/examples/0.20.1/output/spacenet-rio-cc/"
 ```
 
 **All examples**:
 ```sh
 python "rastervision_pytorch_backend/rastervision/pytorch_backend/examples/test.py" \
 compare \
---examples_root_old "s3://raster-vision/examples/0.13/output/" \
---examples_root_new "s3://raster-vision/examples/0.20/output/"
+--examples_root_old "s3://raster-vision/examples/0.20/output/" \
+--examples_root_new "s3://raster-vision/examples/0.20.1/output/"
 ```
 
 ### Collect

rastervision_pytorch_backend/rastervision/pytorch_backend/examples/test.py

@@ -11,8 +11,10 @@
     file_to_json, sync_from_dir, upload_or_copy, download_or_copy, file_exists,
     sync_to_dir, NotReadableError, download_if_needed)
 
-OLD_VERSION = '0.13'
-NEW_VERSION = '0.20'
+OLD_VERSION = '0.20'
+NEW_VERSION = '0.20.1'
+NEW_VERSION_MAJOR_MINOR = '0.20'
+
 EXAMPLES_MODULE_ROOT = 'rastervision.pytorch_backend.examples'
 EXAMPLES_PATH_ROOT = '/opt/src/rastervision_pytorch_backend/rastervision/pytorch_backend/examples'  # noqa
 REMOTE_PROCESSED_ROOT = f's3://raster-vision/examples/{NEW_VERSION}/processed-data'
@@ -21,7 +23,7 @@
 LOCAL_PROCESSED_ROOT = '/opt/data/examples/processed-data'
 LOCAL_OUTPUT_ROOT = '/opt/data/examples/output'
 LOCAL_COLLECT_ROOT = '/opt/data/examples/collect'
-ZOO_UPLOAD_ROOT = f's3://azavea-research-public-data/raster-vision/examples/model-zoo-{NEW_VERSION}'  # noqa
+ZOO_UPLOAD_ROOT = f's3://azavea-research-public-data/raster-vision/examples/model-zoo-{NEW_VERSION_MAJOR_MINOR}'  # noqa
 SAMPLE_IMG_DIR = f's3://azavea-research-public-data/raster-vision/examples/sample_images'  # noqa
 
 ######################
@@ -463,11 +465,11 @@ def _compare_runs(root_uri_old: str,
             _compare_evals(cmd_root_uri_old_local, cmd_root_uri_new_local)
 
 
-def _compare_evals(
-        root_uri_old: str,
-        root_uri_new: str,
-        float_tol: float = 1e-3,
-        exclude_keys: list = ['conf_mat', 'count_error', 'per_scene']) -> None:
+def _compare_evals(root_uri_old: str,
+                   root_uri_new: str,
+                   float_tol: float = 1e-3,
+                   exclude_keys: list = ['conf_mat', 'count',
+                                         'per_scene']) -> None:
     """Compare outputs of the eval command for two runs of an example."""
     console_heading('Comparing keys and values in eval.json files...')
     try:
@@ -590,8 +592,6 @@ def _compare_dicts(dict_old: dict,
     """
     dict_old = flatten_dict(dict_old)
     dict_new: Dict[str, Any] = flatten_dict(dict_new)
-    # TODO this hack is specific to 0.13 vs 0.20 comparison
-    dict_new = {k.replace('.metrics', ''): v for k, v in dict_new.items()}
     keys_old, keys_new = set(dict_old.keys()), set(dict_new.keys())
     diff1, diff2 = keys_new - keys_old, keys_old - keys_new
     if len(diff1) > 0: