Integration: WatchDuty app alerts

[chrisdavis2110]

I found a public API for the WatchDuty app. I have it set up so that it polls the API at a configurable time (default 5 min) within a bounding box and sends messages when it finds new events. You can set a single channel where all the messages are sent, but with WatchDuty sending a lot of reports for a single fire, and if there are multiple fires going at once, it gets overwhelming, so I also included a "feed" channel and a "report" channel. The feed channel just gets stats (Fire name | acreage | containment % | location | link to fire) as the acreage or containment gets updated. The reports channel gets the notifications of the detailed reports WatchDuty sends out (example: Carbon Fire: The Red Cross is staffing a care and reception center at the Chino Hills Community C... | https://app.watchduty.org/i/95436).

I've also added two commands: fire and evac. Fire with no args returns a list of active fires. Fire with either the list number, fire name, or fire id will return the stats of the fire. Evac is similar to fire, but lists the evacuation info for the fire. Still need to test that more, just waiting on more big fires to happen.

Feed channel:

Report channel:

[hickey]

I have build a new alerts architecture (https://github.com/hickey/meshcore-bot/tree/feat/new-alert-architecture) to allow multiple alerting backends to use the same alert command. I have yet to create a PR yet as I am still thinking and experimenting a bit with how the alert command queries the backend alert services. I should be to the point where I could create a draft PR so others can start looking at it.

I am just thinking that this may be useful to you also. There is documentation on creating the backend alert service and interfacing to the alert command.

In my case, Volusia County, FL does not use the PulsePoint API for alerting, but there is a page that the Sheriff's department publishes that shows the active calls out of their CAD system. So in the re-architect I have made the PulsePoint code a backend service and the alert command an interface to the backend services. Then I created a vcso_alert_service.py to answer queries. An added benefit is that the backend service can poll and add messages to a channel when there are incidents (i.e. broadcast that an incident has occurred).

[hickey]

I have created draft PR #284.