makingsensebali

name: makingsense-data-sync

# Publishes the public air-quality data for
# https://mdg-bali.github.io/makingsensebali/
#
# Why this exists: the data used to be pushed only by a launchd job on the mini
# in Bali. Any power cut / network reset / Tailscale drop / exo outage froze the
# whole site (it stalled for ~a month after the June move). Running the fetch in
# GitHub's cloud removes that single point of failure and self-heals: a failed
# run is irrelevant because the next one runs 15 min later.
#
# Scope: this workflow OWNS sensors.json / history.json / areas.json (public
# networks: Smart Citizen + AirGradient are keyless; AQICN/KLHK + PurpleAir turn
# on when the repo secrets are set). It does NOT touch data/reports/ (those come
# from the private NAS) and does NOT run the exo narrative (MSB_NARRATIVE=0) —
# both need on-network infra and are published separately. Because it only writes
# the sensor files, it never collides with the reports writer, and generate.py's
# own `pull --rebase --strategy-option=ours` handles any interleaving.

on:
  schedule:
    - cron: "*/15 * * * *"   # every 15 min (cron is UTC; interval-based so tz is irrelevant)
  workflow_dispatch: {}       # manual "Run workflow" button in the Actions tab

permissions:
  contents: write             # generate.py commits + pushes the refreshed data

concurrency:
  group: makingsense-data-sync
  cancel-in-progress: false   # let a run finish rather than cancel a mid-push

jobs:
  sync:
    runs-on: ubuntu-latest
    timeout-minutes: 10
    steps:
      - name: Checkout
        uses: actions/checkout@v4
        with:
          ref: main
          fetch-depth: 0        # full history so generate.py's rebase works

      - name: Set up Python
        uses: actions/setup-python@v5
        with:
          python-version: "3.12"

      - name: Configure git identity
        run: |
          git config user.name  "makingsense-sync[bot]"
          git config user.email "makingsense-sync@users.noreply.github.com"

      - name: Generate + publish data
        env:
          MSB_REPO_DIR: ${{ github.workspace }}
          MSB_NARRATIVE: "0"                              # no exo narrative in the cloud
          # api.smartcitizen.me answered in 21-32s from Bali on 2026-08-20, so the
          # 20s default timed out EVERY Smart Citizen call and our own kits silently
          # dropped out of sensors.json while history.json went empty. Presented as
          # dead hardware; the kits were online the whole time.
          MSB_HTTP_TIMEOUT: "60"
          # MSB_NAS_RSYNC_SRC deliberately unset -> reports pull is skipped,
          # existing data/reports/ is left untouched.
          MSB_AQICN_TOKEN: ${{ secrets.MSB_AQICN_TOKEN }} # KLHK / WAQI (optional; secret)
          MSB_AQICN_STATIONS: "-519205"                   # pinned Bali station UID (not secret)
          MSB_PURPLEAIR_KEY: ${{ secrets.MSB_PURPLEAIR_KEY }} # PurpleAir read key (optional; secret)
          MSB_PURPLEAIR_IDS: "36601,46949"                # Bali community sensors (not secret)
        run: python3 tools/makingsense-sync/generate.py