Archive New Conference URLs #6

	name: Archive New Conference URLs

	on:
	schedule:
	# Runs at 00:00 UTC on Friday
	# https://crontab.guru/#0_0___6
	- cron: '0 0 * * 6'
	workflow_dispatch:
	# Allows manual triggering

	jobs:
	check-new-urls:
	runs-on: ubuntu-latest

	steps:
	- uses: actions/checkout@v4
	with:
	fetch-depth: 0 # Full history for finding last run

	- name: Find last successful run commit
	id: last-run
	run: \|
	last_sha=$(gh run list --workflow="Check New Conference URLs" --json conclusion,headSha --jq '.[] \| select(.conclusion=="success") \| .headSha' \| head -n1)
	if [ -n "$last_sha" ]; then
	echo "sha=$last_sha" >> $GITHUB_OUTPUT
	else
	# If no previous run, use the first commit
	echo "sha=$(git rev-list --max-parents=0 HEAD)" >> $GITHUB_OUTPUT
	fi
	env:
	GITHUB_TOKEN: ${{ secrets.GITHUB_TOKEN }}

	- name: Extract and archive URLs
	id: url-check
	run: \|
	# Extract URLs
	URLS=$(git diff ${{ steps.last-run.outputs.sha }} -- _data/conferences.yml \| \
	grep -E '^\+.*https?://[^ ]+' \| \
	sed -E 's/^\+[[:space:]][^:]:[[:space:]](https?:\/\/[^ ]+)./\1/')

	if [ -n "$URLS" ]; then
	echo "Found new URLs, archiving to Wayback Machine..."

	# Process each URL
	echo "$URLS" \| while read -r url; do
	echo "Archiving: $url"
	# Call Wayback Machine Save API
	response=$(curl -s -L -X POST \
	-A "Python-Deadlines-Bot/1.0 (https://pythondeadlin.es; [email protected])" \
	"https://web.archive.org/save/$url")
	status=$?

	if [ $status -eq 0 ]; then
	echo "✓ Successfully archived"
	else
	echo "⚠ Failed to archive (error $status)"
	fi

	# Wait between requests
	sleep 5
	done
	else
	echo "No new URLs found"
	fi

Provide feedback