Scrapes/parses for new Covid-19 Inquiry data #2047

	name: update-site
	run-name: Scrapes/parses for new Covid-19 Inquiry data

	on:
	workflow_dispatch:
	schedule:
	- cron: '13 9,12,15,18,21 * * 1-5'

	permissions:
	contents: write

	jobs:
	build:
	runs-on: ubuntu-latest
	steps:
	- uses: actions/checkout@v3

	- name: Install pdftotext
	run: \|
	sudo apt-get update
	sudo apt-get install poppler-utils

	- uses: actions/setup-python@v4
	with:
	python-version: '3.11'
	cache: 'pip'
	cache-dependency-path: 'requirements-parse.txt'

	- name: Install dependencies
	run: pip install -r requirements-parse.txt

	- name: Make data directory
	run: mkdir data

	- uses: actions/cache@v3
	with:
	path: data
	key: data

	- name: Scrape transcript data
	run: python scrape.py

	- name: Parse transcript data
	run: python parse.py

	- name: Check if there are any changes
	id: verify_diff
	run: \|
	git add -N .
	git diff --quiet . \|\| echo "changed=true" >> $GITHUB_OUTPUT

	- name: Commit
	if: steps.verify_diff.outputs.changed == 'true'
	run: \|
	git config --global user.name 'Matthew Somerville'
	git config --global user.email '[email protected]'
	git commit -am "Data update."
	git push

Provide feedback