https://github.com/hursey013/casa-del-cardinal
Friendly bot logging visitors to our window mounted bird feeder in Charlottesville, VA
https://github.com/hursey013/casa-del-cardinal
Last synced: 3 months ago
JSON representation
Friendly bot logging visitors to our window mounted bird feeder in Charlottesville, VA
- Host: GitHub
- URL: https://github.com/hursey013/casa-del-cardinal
- Owner: hursey013
- Created: 2021-03-21T15:30:08.000Z (about 4 years ago)
- Default Branch: main
- Last Pushed: 2021-12-06T17:00:46.000Z (over 3 years ago)
- Last Synced: 2025-01-22T10:44:08.714Z (4 months ago)
- Language: JavaScript
- Homepage: https://twitter.com/casadelcardinal
- Size: 5.36 MB
- Stars: 4
- Watchers: 3
- Forks: 0
- Open Issues: 0
-
Metadata Files:
- Readme: README.md
Awesome Lists containing this project
README
# casadelcardinal
> Friendly bot logging visitors to our window mounted bird feeder in Charlottesville, VA
## Introduction
`casadelcardinal` is a Firebase Cloud Function and Tensorflow.js bird classification model which provides an API endpoint to identify the species in a given image. The results are stored in a Firebase Realtime Database and passed to the Twitter API to update the connected account's status.
## Requirements
- [Home Assistant](https://github.com/home-assistant/core)
- [Frigate](https://github.com/blakeblackshear/frigate)
- [eclipse-mosquitto](https://hub.docker.com/_/eclipse-mosquitto) or another MQTT broker
## Initial setup
### Clone this repo
- Clone or download this repo and open the `casadelcardinal` directory.
### Create a Firebase project
- Create a Firebase Project using the [Firebase Developer Console](https://console.firebase.google.com)
- Enable billing on your project by switching to the Blaze or Flame plan. See [pricing](https://firebase.google.com/pricing/) for more details. This is required to allow requests to non-Google services within the Function.
- Install [Firebase CLI Tools](https://github.com/firebase/firebase-tools) if you have not already, and log in with `firebase login`.
- Configure the function to use your project using `firebase use --add` and select your project.
### Install dependencies and add environment variables
- Install dependencies locally by running: `cd functions; npm i; cd -`
- [Add your Twitter API credentials](https://developer.twitter.com/) to the Firebase config:
```bash
firebase functions:config:set \
twitter.consumer_key= \
twitter.consumer_secret= \
twitter.access_token= \
twitter.access_token_secret= \
```
### Customize function
There are a number of variables which can be customized within the function. Begin by copying `config.sample.js` to `config.js`.
- `apiUrl` (string) - set to your public Home Assistant URL
- `cooldown` (number) - minutes to wait before reposting the same species
- `known` - (array) - IDs of species which will always be considered valid
- `threshold` (number) - if not included in `known`, the score must be above this threshold to be valid
`casadelcardinal` uses [RosaeNLG](https://rosaenlg.org/), a Natural Language Generation library, to generate the text for the Twitter status updates. For more information about the syntax used, visit [RosaeNLG](https://rosaenlg.org/).
Within the template, the following values can be used:
- `known` (boolean) - is the species included in the `known` array
- `now` (number) - the current hour (from 0 to 23)
- `results` (object) - includes the following keys: `id`, `name`, `common_name`, `score`
- `threshold` (number) - value set in config.js
### Deploy the app to production
- Deploy your function using `firebase deploy --only functions`
After deploying the function POST requests can be sent to:
```
https://.firebaseapp.com/app
```
### Configure Frigate
The steps to configure Frigate will vary based on your setup. Review the [Frigate docs](https://blakeblackshear.github.io/frigate/) for more information. The following settings should be included to support `casadelcardinal`:
```
...
mqtt:
host: xxx.xxx.xxx.xxx
clips:
max_seconds: 0
cameras:
feeder:
ffmpeg:
inputs:
- path: rtsp://user:[email protected]/live
roles:
- clips
- detect
best_image_timeout: 300
snapshots:
crop: True
enabled: True
clips:
enabled: True
pre_capture: 0
post_capture: 0
detect:
max_disappeared: 50
objects:
track:
- bird
filters:
bird:
min_area: 90000
...
```
### Create Home Assistant automation
The steps to configure your Home Assistant automation will vary based on your preferences. The following example uses MQTT and a [RESTful Command](https://www.home-assistant.io/integrations/rest_command/) to trigger the Firebase Function at the end of a Frigate event:
#### automations.yaml
```
trigger:
- platform: mqtt
topic: frigate/events
condition:
- condition: template
value_template: '{{ trigger.payload_json["type"] == "end" }}'
action:
- service: rest_command.casa_del_cardinal
data_template:
payload: '{{trigger.payload}}'
```
#### configuration.yaml
```
rest_command:
casa_del_cardinal:
url: "https://.firebaseapp.com/app"
method: POST
payload: '{{ payload }}'
content_type: "application/json; charset=utf-8"
timeout: 300
```