NHL_DB / .github /workflows /update-data.yaml
RentoSaijo's picture
Refactor NHL_DB update automation
d196151
Raw
History Blame Contribute Delete
3.37 kB
name: Update HuggingFace datasets
on:
schedule:
- cron: '0 6 * * *'
timezone: 'America/New_York'
workflow_dispatch:
inputs:
target_repo:
description: 'HuggingFace dataset repo to upload generated parquet files to.'
required: true
default: 'RentoSaijo/NHL_DB'
seasons:
description: 'Optional comma-separated season IDs, such as 20252026.'
required: false
default: ''
datasets:
description: 'Optional comma-separated datasets: gc_raw,gc,wsc_raw,wsc,scs,scss.'
required: false
default: ''
dry_run:
description: 'Generate and validate files without uploading.'
required: true
default: 'false'
type: choice
options:
- 'false'
- 'true'
permissions: read-all
concurrency:
group: nhl-db-update
cancel-in-progress: false
jobs:
update-data:
runs-on: ubuntu-latest
timeout-minutes: 360
env:
HF_TOKEN: ${{ secrets.HF_TOKEN }}
TARGET_REPO: ${{ github.event.inputs.target_repo || 'RentoSaijo/NHL_DB' }}
NHL_DB_OUTPUT_DIR: output
NHL_DB_HF_SOURCE_REPO: ${{ github.event.inputs.target_repo || 'RentoSaijo/NHL_DB' }}
NHL_DB_SEASONS: ${{ github.event.inputs.seasons || '' }}
NHL_DB_DATASETS: ${{ github.event.inputs.datasets || '' }}
NHL_DB_TIMEZONE: America/New_York
NHL_DB_DOWNLOAD_TIMEOUT: '3600'
DRY_RUN: ${{ github.event.inputs.dry_run || 'false' }}
steps:
- uses: actions/checkout@v4
- uses: r-lib/actions/setup-r@v2
with:
use-public-rspm: true
- uses: r-lib/actions/setup-pandoc@v2
- name: Install R dependencies
shell: Rscript {0}
run: |
install.packages(c('arrow', 'httr2', 'remotes'))
remotes::install_git(
'https://github.com/RentoSaijo/nhlscraper.git',
dependencies = TRUE,
upgrade = 'never'
)
- name: Install HuggingFace CLI
run: python -m pip install --upgrade 'huggingface_hub[cli]'
- name: Generate game parquet files
run: Rscript R/game.R
- name: Generate event parquet files
run: Rscript R/event.R
- name: Validate generated files
shell: Rscript {0}
run: |
files <- list.files(Sys.getenv('NHL_DB_OUTPUT_DIR'), recursive = TRUE, full.names = TRUE)
bad <- files[grepl('\\.(csv|csv\\.gz)$', files)]
if (length(bad)) {
stop('CSV output detected: ', paste(bad, collapse = ', '))
}
parquet_files <- files[grepl('\\.parquet$', files)]
if (!length(parquet_files)) {
message('No parquet files were generated; nothing to upload.')
quit(save = 'no')
}
invisible(lapply(parquet_files, arrow::read_parquet))
- name: Upload HuggingFace dataset
if: env.DRY_RUN != 'true'
run: |
test -n "$HF_TOKEN"
if ! find "$NHL_DB_OUTPUT_DIR" -name '*.parquet' -type f | grep -q .; then
echo "No parquet files were generated; skipping upload."
exit 0
fi
hf upload-large-folder "$TARGET_REPO" "$NHL_DB_OUTPUT_DIR" \
--repo-type dataset \
--include '*.parquet' \
--token "$HF_TOKEN" \
--num-workers 4