Move basic PR checks like Flake8, Mypy and HelloWorld to Github actio…

…ns (#426)

At present, external contributors don't have any insight into why the PR builds fail because they run on ADO. This PR moves some of the basic checks to Github Actions, where they are fully visible: Flake8, mypy, and training the HelloWorld model.

.github/workflows/linting_and_hello_world.yml

@@ -0,0 +1,69 @@
+name: Flake8, MyPy, and Hello World
+
+# This workflow runs linting via Flake8 and Mypy.
+# It also checks that the HelloWorld model can be trained.
+
+on:
+ pull_request:
+
+jobs:
+ linux:
+ runs-on: ubuntu-latest
+ steps:
+ - uses: actions/checkout@v2
+ with:
+ lfs: true
+
+ - name: flake8
+ run: |
+ pip install flake8
+ python -m flake8
+ shell: bash
+ if: always() 
+
+ # This script also does "conda init" for all shells. For bash, this modifies .bashrc.
+ # However, the default "bash" in a github workflow does not execute bashrc. Hence, all
+ # scripts that use this environment need to either hardcode the path to Python, or use
+ # a customized shell that executes bashrc, like "shell: bash -eo pipefail {0}"
+ - uses: conda-incubator/setup-miniconda@v2
+ with:
+ activate-environment: InnerEye
+ auto-activate-base: false
+ environment-file: environment.yml
+ python-version: 3.7
+ if: always()
+
+ - name: mypy
+ run: |
+ $CONDA/envs/InnerEye/bin/python mypy_runner.py --mypy=$CONDA/envs/InnerEye/bin/mypy
+ if: always()
+
+ - name: Run HelloWorld model
+ run: |
+ $CONDA/envs/InnerEye/bin/python ./InnerEye/ML/runner.py --model=HelloWorld
+ env:
+ PYTHONPATH: ${{ github.workspace }}
+ if: always()
+
+ windows:
+ runs-on: windows-latest
+ steps:
+ - uses: actions/checkout@v2
+ with:
+ lfs: true
+
+ - uses: conda-incubator/setup-miniconda@v2
+ with:
+ activate-environment: InnerEye
+ auto-activate-base: false
+ environment-file: environment.yml
+ python-version: 3.7
+ if: always()
+
+ - name: Run HelloWorld model
+ run: |
+ conda info
+ %CONDA%\envs\InnerEye\bin\python ./InnerEye/ML/runner.py --model=HelloWorld
+ env:
+ PYTHONPATH: ${{ github.workspace }}
+ shell: cmd

CHANGELOG.md

@@ -28,7 +28,12 @@ nodes in AzureML. Example: Add `--num_nodes=2` to the commandline arguments to t
  `dataset.csv` supports multiple labels (indices corresponding to `class_names`) per subject in the label column. 
  Multiple labels should be encoded as a string with labels separated by a `|`, for example "0|2|4".
  Note that this PR does not add support for multiclass models, where the labels are mutually exclusive.
-- ([#425](https://github.com/microsoft/InnerEye-DeepLearning/pull/425)) The number of layers in a Unet is no longer fixed at 4, but can be set via the config field `num_downsampling_paths`. A lower number of layers may be useful for decreasing memory requirements, or for working with smaller images. (The minimum image size in any dimension when using a network of n layers is 2**n.)
+- ([#425](https://github.com/microsoft/InnerEye-DeepLearning/pull/425)) The number of layers in a Unet is no longer 
+ fixed at 4, but can be set via the config field `num_downsampling_paths`. A lower number of layers may be useful 
+ for decreasing memory requirements, or for working with smaller images. 
+ (The minimum image size in any dimension when using a network of n layers is 2**n.)
+- ([#426](https://github.com/microsoft/InnerEye-DeepLearning/pull/426)) Flake8, mypy, and testing the HelloWorld model
+ is now happening in a Github action, no longer in Azure Pipelines.
 
 ### Changed
 - ([#385](https://github.com/microsoft/InnerEye-DeepLearning/pull/385)) Starting an AzureML run now uses the

InnerEye/ML/reports/notebook_report.py

@@ -68,7 +68,9 @@ def generate_notebook(template_notebook: Path, notebook_params: Dict, result_not
  papermill.execute_notebook(input_path=str(template_notebook),
  output_path=str(result_notebook),
  parameters=notebook_params,
- progress_bar=False)
+ progress_bar=False,
+ # Unit tests often fail with cell timeouts when default of 4 is used.
+ iopub_timeout=10)
  return convert_to_html(result_notebook)
 
 

azure-pipelines/build.yaml

@@ -26,31 +26,6 @@ steps:
  condition: succeeded()
  displayName: Create conda environment
 
- - bash: |
- source activate InnerEye
- flake8
- failOnStderr: true
- condition: succeededOrFailed()
- displayName: flake8
-
- - bash: |
- source activate InnerEye
- python mypy_runner.py
- failOnStderr: true
- condition: succeededOrFailed()
- displayName: mypy
-
- # Build the HelloWorld model, without providing the secrets for the ServicePrincipal, to mimic running HelloWorld
- # right after checking out. This must happen before we provide any additional information in settings.yml
- - bash: |
- source activate InnerEye
- python ./InnerEye/ML/runner.py --model=HelloWorld
- env:
- PYTHONPATH: $(Build.SourcesDirectory)/
- failOnStderr: false
- condition: succeededOrFailed()
- displayName: Run HelloWorld model
-
  # Pytest needs subscription information directly in settings.yml. Local package install will cause use of
  # the wrong default project root, hence InnerEyePrivateSettings.yml can't be picked up.
  - bash: |