Incident Free Days

A simple Flask app that visualizes incident-free days across a year. It renders a year-long calendar that colors each day based on whether an incident was reported and provides a form to add new incidents.

Features

• Year view calendar with green (incident-free), red (incident), and grey (future/unknown) days.
• Tabbed interface to switch between the calendar and the incident submission form.
• Incident list showing everything recorded in incidents.json, sorted newest first.
• Accepts incident dates as either YYYY-MM-DD (native date input) or MM/DD/YYYY.

Getting started

1. Install dependencies (Flask, Pillow, and related libraries)

   sudo apt-get update && sudo apt-get install -y fonts-dejavu-core
   python -m venv .venv
   source .venv/bin/activate
   pip install -r requirements.txt

2. Run the app

   python app.py

   The server listens on http://0.0.0.0:8080/.

Run as a system service (systemd)

1. Create a dedicated user (optional but recommended)

   sudo useradd --system --shell /usr/sbin/nologin --home /opt/incident-free-days incident
   sudo mkdir -p /opt/incident-free-days
   sudo chown -R incident:incident /opt/incident-free-days

2. Copy the app to the service directory

   sudo cp -r /workspace/incident-free-days /opt/incident-free-days/app
   sudo chown -R incident:incident /opt/incident-free-days/app

3. Create the virtual environment and install dependencies

   cd /opt/incident-free-days/app
   python -m venv .venv
   source .venv/bin/activate
   pip install -r requirements.txt
   deactivate

4. Create the systemd unit file

   sudo tee /etc/systemd/system/incident-free-days.service > /dev/null <<'EOF'
   [Unit]
   Description=Incident Free Days Flask app
   After=network.target
   
   [Service]
   Type=simple
   User=incident
   WorkingDirectory=/opt/incident-free-days/app
   ExecStart=/opt/incident-free-days/app/.venv/bin/python app.py
   Restart=on-failure
   Environment=FLASK_ENV=production
   
   [Install]
   WantedBy=multi-user.target
   EOF

5. Start and enable the service

   sudo systemctl daemon-reload
   sudo systemctl start incident-free-days.service
   sudo systemctl status incident-free-days.service
   sudo systemctl enable incident-free-days.service

   The app will now start on boot and can be managed via systemctl.

6. Report an incident

• Open the app in a browser and switch to the Configuration tab.

Data storage

• Incidents are stored in incidents.json in the project root. The file is created on first save.
• Each incident is a JSON object with keys inc_number, date (YYYY-MM-DD), severity, and pillar.
• Days are colored red when one or more incidents exist for that date; past days without incidents are green.

Configuration

• The default year rendered on the calendar is controlled by DEFAULT_YEAR in app.py (currently 2025).
• The calendar uses Sunday as the first day of the week (calendar.Calendar(firstweekday=6)).
• To sync incidents from incident.io, set INCIDENT_IO_API_TOKEN (and optionally INCIDENT_IO_BASE_URL if you use a non-default host).
• To configure an ESP32-driven OSHA display, visit the hidden /dpost route and provide the display IP/port plus automation preferences.

Syncing from incident.io

The app can pull incidents directly from incident.io so you do not need to manually upload CSV files.

Via the Configuration tab

1. Open the app and go to Configuration.
2. Enter your INCIDENT_IO_API_TOKEN and optional base URL override. Tokens are stored locally in sync_config.json and are never displayed back in the UI.
3. Pick a sync cadence (once an hour, once a day [default], or once a week) and optionally set a start/end date window.
4. Click Dry run mapping to preview the first 10 normalized payloads and verify pillars/severities.
5. Click Import incidents to pull all incidents in the selected window into incidents.json/others.json.

Automatic sync cadence

• The Flask server now runs a lightweight background worker that honors your selected cadence as long as the app process is running. It will trigger /sync/incidents with the stored token/base URL/date window and update the Last sync status on completion.
• To keep the worker alive between reboots, run the app under a supervisor such as the systemd unit described above; you don't need an extra cron entry.
• If you prefer to offload scheduling, you can alternatively point an external scheduler (cron, systemd timer, Kubernetes CronJob) at POST http://<host>:8080/sync/incidents with the token/base URL payload. In that mode, set cadence to "off" or ignore the in-app cadence.

Via the CLI or API

1. Export your incident.io API token to the environment:

   export INCIDENT_IO_API_TOKEN="<token>"
   # Optional: override the API host
   # export INCIDENT_IO_BASE_URL="https://eu.api.incident.io"

2. Run a dry run from the feature branch to verify parsing and mapping without writing to disk:

   python app.py --sync-incidents --dry-run

3. Sync data to incidents.json/others.json once the output looks correct:

   python app.py --sync-incidents

You can also trigger the sync over HTTP using GET /sync/incidents?dry_run=1 (dry run) or POST /sync/incidents with a JSON payload (token, optional date range) to persist results.

E-Paper display pushes

• Visit /dpost to enter the display IP (e.g., 192.168.1.25 or 192.168.1.25:80) and enable pushing. Settings are stored locally alongside the incident.io token.
• When enabled, the OSHA tab shows a Send to display button that rerenders the OSHA sign and pushes it over the ESP32 start/chunk/end protocol.
• Automation options:
  • Daily push at a configured local time (defaults to 05:00 America/Phoenix). A background worker in the Flask process handles the schedule; keep the service running (e.g., via systemd) to preserve the cadence.
  • Push on new procedural incident: after each incident.io sync, the app detects a new procedural RCA incident and sends the updated sign once per incident number.
• The OSHA status panel records the last push time, incident number, and any error text from the ESP32.

Troubleshooting display pushes

• Verify the IP/port in /dpost and that the ESP32 is reachable over HTTP.
• Timeouts or display request failed errors usually indicate network issues or a busy display; retry after a few seconds.
• If pushes do not run daily, confirm the server process is running (or managed by systemd) so the background scheduler can fire.

Repository structure

• app.py — Flask application entrypoint and routing.
• templates/index.html — Calendar layout, tabs, and incident table.
• static/styles.css — Styling for the calendar, tabs, and forms.
• static/tabs.js — Simple tab switcher for the UI and HTML-to-image export logic.

Importing data and pillar key mapping

• Upload a product → pillar key JSON file using the Upload product → pillar key form; the key is saved alongside incidents.json.
• Duplicate incident numbers are skipped when adding incidents via sync.

Exporting the calendar to an image

• The calendar tab includes an Export as Image button that captures the visible calendar into a PNG.
• The export runs completely in the browser (no external CDN dependency), so it works even in offline environments.
• The server also provides /calendar/eink.png, which returns a 1600×1200 PNG calendar for the current month by default. You can override the period with ?year=YYYY&month=MM. This endpoint is designed for consumption by external clients such as a Raspberry Pi driving a 13.3" e-ink display.