Skip to content

DB Refresh (Objects API) #24861

DB Refresh (Objects API)

DB Refresh (Objects API) #24861

Workflow file for this run

name: DB Refresh (Objects API)
on:
schedule:
# - cron: '0 5 * * *' # every day 5am GMT (12/1am EST/EDT)
- cron: '*/20 * * * *' # every 20 minutes
# - cron: '0 * * * *' # every hour on the hour
workflow_dispatch:
inputs:
since:
# A timestamp by which new data should be fetched. If not present, will look up the last time the DB was
# refreshed and fetch new data from that time. Valid formats:
# - Simple date timestamp, YYYY-MM-DD, e.g. 2022-02-22.
# - ISO 8601 datetime with timezone offset, YYYY-MM-DDTHH:MM:SS.SSSSSS+HH:MM, e.g. 2022-02-22T22:22:22.222222+00:00.
description: 'Optional: YYYY-MM-DD or YYYY-MM-DDTHH:MM:SS.SSSSSS+HH:MM'
required: false
type: string
cset-ids:
# Space-delimited list of integers representing codeset IDs
description: 'Optional: Space-delimited list of integers representing codeset IDs'
required: false
type: string
repository_dispatch:
types: [refresh-db]
jobs:
refresh-db:
# runs-on: ubuntu-18.04
runs-on: ubuntu-latest
steps:
# Set up
- name: Checkout repository and submodules
uses: actions/checkout@v2
- name: Set up Python version
uses: actions/setup-python@v2
with:
python-version: '3.10'
- name: Print commit hash & branch for rollbacks & troubleshooting
run: |
echo "Commit hash: ${{ github.sha }}"
echo "Branch: ${{ github.ref }}"
- name: 'Create env file'
run: |
mkdir env
echo "${{ secrets.ENV_FILE }}" > env/.env
- name: Create and start virtual environment
run: |
python3 -m venv venv
source venv/bin/activate
- name: Install dependencies
run: |
# Reactivate install upgrades of pip, wheel, &/or setuptools if needed
# python -m pip install --upgrade pip
# pip install --upgrade wheel
# pip install --upgrade setuptools
# VS Hub SDK w/ OAuth disabled: https://github.com/jhu-bids/TermHub/issues/863
# TOKEN=$(grep '^PALANTIR_ENCLAVE_AUTHENTICATION_BEARER_TOKEN=' env/.env | cut -d'=' -f2)
# python3 -m pip install vshub_sdk --upgrade --extra-index-url "https://:${TOKEN}@unite.nih.gov/artifacts/api/repositories/ri.artifacts.main.repository.9bc9cc56-4b8c-4560-9bfb-fba8ade55246/contents/release/pypi/simple" --extra-index-url "https://:${TOKEN}@unite.nih.gov/artifacts/api/repositories/ri.foundry-sdk-asset-bundle.main.artifacts.repository/contents/release/pypi/simple"
pip install -r requirements.txt
# Refresh
- name: Refresh db
run: |
if [ -n "${{ github.event.inputs.cset-ids }}" ] && [ -n "${{ github.event.inputs.since }}" ]; then
echo "Error: Both 'cset-ids' and 'since' were provided. These inputs are mutually exclusive."
exit 1
elif [ -n "${{ github.event.inputs.cset-ids }}" ]; then
python backend/db/refresh.py --cset-ids ${{ github.event.inputs.cset-ids }}
elif [ -n "${{ github.event.inputs.since }}" ]; then
python backend/db/refresh.py --since ${{ github.event.inputs.since }}
else
python backend/db/refresh.py
fi
- name: Commit changes to documentation
uses: stefanzweifel/git-auto-commit-action@v4
with:
commit_message: Updating documentation in docs/backend/db/analysis.md
# fetch-missing-csets: Deactivated because various tweaks and default re-refresh buffer solve this problem, and
# also test-missing-csets is there as a redundancy. Original context:
# https://github.com/jhu-bids/TermHub/issues/573 https://github.com/jhu-bids/TermHub/issues/574
# - name: Fetch missing csets
# run: make fetch-missing-csets
# Test
# todo: add more tests, ideally
- name: Test missing csets
run: make test-missing-csets