Tests workflow (bethgelab/foolbox)
The Tests workflow from bethgelab/foolbox, explained and optimized by Latchkey.
CI health: F - at risk
Point runs-on at Latchkey and get caching, run de-duplication, job timeouts, SHA-pinned actions, self-healing for flaky steps, and up to 58% lower cost, applied automatically.
What it does
This is the Tests workflow from the bethgelab/foolbox repository, a real project running GitHub Actions. It is shown here with attribution under its MIT license.
Below, Latchkey shows a faster, safer version produced by its optimization engine.
The workflow
name: Tests
on:
push:
branches:
- master
pull_request:
jobs:
build:
runs-on: ubuntu-latest
strategy:
max-parallel: 20
matrix:
python-version: ['3.9', '3.10']
backend: ["none", "pytorch", "tensorflow", "jax", "numpy"]
steps:
- uses: actions/checkout@v2
- uses: actions/cache@v1
with:
path: ~/.cache/pip
key: ${{ runner.os }}-pip-${{ hashFiles('**/requirements.txt') }}
restore-keys: |
${{ runner.os }}-pip-
- uses: actions/cache@v1
with:
path: ~/.cache/torch
key: ${{ runner.os }}-cache-torch
- uses: actions/cache@v1
with:
path: ~/.torch
key: ${{ runner.os }}-torch
- uses: actions/cache@v1
with:
path: ~/.keras
key: ${{ runner.os }}-keras
# - name: debugging
# run: |
# python3 -c 'from urllib.request import urlopen; r = urlopen("https://download.pytorch.org/models/resnet18-5c106cde.pth"); print(r.status, r.reason, r.msg)' || true
# - name: workaround for https://github.com/pytorch/vision/issues/1876
# uses: wei/wget@v1
# with:
# args: https://download.pytorch.org/models/resnet18-5c106cde.pth
# - name: workaround (step 2)
# run: |
# mkdir -p ~/.cache/torch/checkpoints || true
# mv resnet18-5c106cde.pth ~/.cache/torch/checkpoints || true
- name: Set up Python ${{ matrix.python-version }}
uses: actions/setup-python@v1
with:
python-version: ${{ matrix.python-version }}
- name: Install requirements.txt
run: |
function retry-with-backoff() {
for BACKOFF in 0 1 2 4 8 16 32 64; do
sleep $BACKOFF
if "$@"; then
return 0
fi
done
return 1
}
python -m pip install --upgrade pip setuptools wheel
retry-with-backoff pip install -r requirements.txt
- name: flake8
run: |
flake8 . --count --show-source --statistics
- name: black
run: |
black --check --verbose .
- name: Install package
run: |
pip install -e .
- name: Install tests/requirements.txt
run: |
function retry-with-backoff() {
for BACKOFF in 0 1 2 4 8 16 32 64; do
sleep $BACKOFF
if "$@"; then
return 0
fi
done
return 1
}
retry-with-backoff pip install -r tests/requirements.txt
- name: mypy (package)
run: |
mypy --install-types --non-interactive foolbox/
mypy -p foolbox
- name: mypy (tests)
run: |
mypy tests/
- name: Test with pytest (backend ${{ matrix.backend }})
run: |
pytest --durations=0 --cov-report term-missing --cov=foolbox --verbose --backend ${{ matrix.backend }}
- name: Codecov
continue-on-error: true
env:
CODECOV_TOKEN: "60c0d7ac-8ec1-47c6-b3b1-ac2ad2dea76f"
run: |
codecov
- name: Coveralls
continue-on-error: true
env:
COVERALLS_REPO_TOKEN: "2r76Cn01kW1sSkEirJ3SRpp478NJtPNdA"
run: |
coveralls
The same workflow, on Latchkey
Estimated ~20% faster on cache hits, plus fewer wasted runs and a safer supply chain. Added and changed lines are highlighted.
name: Tests on: push: branches: - master pull_request: concurrency: group: ${{ github.workflow }}-${{ github.ref }} cancel-in-progress: true jobs: build: timeout-minutes: 30 runs-on: latchkey-small strategy: max-parallel: 20 matrix: python-version: ['3.9', '3.10'] backend: ["none", "pytorch", "tensorflow", "jax", "numpy"] steps: - uses: actions/checkout@v2 - uses: actions/cache@v1 with: path: ~/.cache/pip key: ${{ runner.os }}-pip-${{ hashFiles('**/requirements.txt') }} restore-keys: | ${{ runner.os }}-pip- - uses: actions/cache@v1 with: path: ~/.cache/torch key: ${{ runner.os }}-cache-torch - uses: actions/cache@v1 with: path: ~/.torch key: ${{ runner.os }}-torch - uses: actions/cache@v1 with: path: ~/.keras key: ${{ runner.os }}-keras # - name: debugging # run: | # python3 -c 'from urllib.request import urlopen; r = urlopen("https://download.pytorch.org/models/resnet18-5c106cde.pth"); print(r.status, r.reason, r.msg)' || true # - name: workaround for https://github.com/pytorch/vision/issues/1876 # uses: wei/wget@v1 # with: # args: https://download.pytorch.org/models/resnet18-5c106cde.pth # - name: workaround (step 2) # run: | # mkdir -p ~/.cache/torch/checkpoints || true # mv resnet18-5c106cde.pth ~/.cache/torch/checkpoints || true - name: Set up Python ${{ matrix.python-version }} uses: actions/setup-python@v1 with: cache: 'pip' python-version: ${{ matrix.python-version }} - name: Install requirements.txt run: | function retry-with-backoff() { for BACKOFF in 0 1 2 4 8 16 32 64; do sleep $BACKOFF if "$@"; then return 0 fi done return 1 } python -m pip install --upgrade pip setuptools wheel retry-with-backoff pip install -r requirements.txt - name: flake8 run: | flake8 . --count --show-source --statistics - name: black run: | black --check --verbose . - name: Install package run: | pip install -e . - name: Install tests/requirements.txt run: | function retry-with-backoff() { for BACKOFF in 0 1 2 4 8 16 32 64; do sleep $BACKOFF if "$@"; then return 0 fi done return 1 } retry-with-backoff pip install -r tests/requirements.txt - name: mypy (package) run: | mypy --install-types --non-interactive foolbox/ mypy -p foolbox - name: mypy (tests) run: | mypy tests/ - name: Test with pytest (backend ${{ matrix.backend }}) run: | pytest --durations=0 --cov-report term-missing --cov=foolbox --verbose --backend ${{ matrix.backend }} - name: Codecov continue-on-error: true env: CODECOV_TOKEN: "60c0d7ac-8ec1-47c6-b3b1-ac2ad2dea76f" run: | codecov - name: Coveralls continue-on-error: true env: COVERALLS_REPO_TOKEN: "2r76Cn01kW1sSkEirJ3SRpp478NJtPNdA" run: | coveralls
What changed
- Run on Latchkey managed runners with one line (
runs-on), which apply the fixes below automatically and self-heal transient failures. This example useslatchkey-small; pick the runner size that fits the job. - Cancel superseded runs when a branch or PR gets a newer push.
- Cache dependency installs on the setup step so they are served from cache.
- Add a job timeout so a hung step cannot burn hours of runner time.
1 third-party action is referenced by a movable tag. Pin it to the commit SHA (Latchkey resolves and applies this automatically) so a repointed tag cannot change what runs.
What Latchkey heals here
This workflow has steps that commonly fail on transient issues (network, registries, flaky browsers). On Latchkey managed runners they are detected, retried, and self-healed instead of failing your build:
- Dependency installs
This workflow runs 1 job (10 with the matrix expanded) per trigger. On Latchkey the same minutes cost up to 58% less than GitHub-hosted, with zero queue time.