README

This Docker Compose configuration sets up a development environment for Pyronear's API along with supporting services like a PostgreSQL database, LocalStack for S3 emulation, Pyro Engine, and Promtail for log shipping.

Services

1. pyro-api: Runs the Pyronear API using uvicorn.

2. db: PostgreSQL database for the API.

3. localstack: Emulates AWS S3 using LocalStack.

4. pyro-engine: Pyro Engine service.

5. reolinkdev: a service which imitate a reolink camera by sending back pictures of fire.

6. frontend: our webapp available on the 8085 port.

   Additional services (helpers):

7. notebooks : Python notebook server to run scripts on api without having to install python

8. db-ui (pgadmin): UI to visualize and manipulate the data in PostgreSQL database

9. s3manager: UI to visualize and manipulate the data in S3 LocalStack

Installation

Prerequisites

• Docker and Docker Compose installed on your system.
• Precommit hook installed on this repo

Running everything (engine + predefined alerts)

Start the Docker services using the following commands:

make build
make run-all

This will launch the full stack including the engine and the predefined alert generation.

---

you can check that everyhing is working thanks to the following commands :

docker logs init
docker logs engine

Running services partially

If you want to launch only the engine and two dev-cameras you can use :

make run-engine

If you want to launch only the additional tools, you can use :

make run-tools

Running customized alerts using personal notebooks (not in docker)

Install the notebook dependencies:

pip install -r notebooks/requirements.txt

Access

Accessing the API

Once the services are up and running, you can access the Pyronear API at http://localhost:5050/docs.

Accessing the web-app

First you need to tell your computer where your S3 is. For that you will have to add this line to you /etc/hosts :

127.0.0.1 www.localstack.com localstack

Since Dash can be a bit capricious, you should launch a private window from you browser and access the web app at http://localhost:8050

29/01/2024 : For the moment, the ADMIN access doesn't show the alerts sent by the camera. For that you will have to use a user account which are defined in data/csv/users.csv

Launch the web app manually from the pyro-platform directory

You can launch the API :

make run-api

And, in your pyro-platform/.env use this API_URL env var :

API_URL=http://localhost:5050

Access the service notebooks

Access at the address : http://localhost:8889/

For the first connection, open the logs of the container. Copy the token and save it for later (to login in the service)

Access the service db-ui

You can access the db-ui service (pgadmin) at http://localhost:8888/browser/

Log in with the mail/pwd specified in the env file (DB_UI_MAIL/DB_UI_PWD)

At the first connection, the db server must be configured:

• Register a server with those data :
• Name: "pyro-db"
• Host name/address: "db"
• Maintenance database : see POSTGRES_DB in .env
• User : see POSTGRES_USER in .env
• Password : see POSTGRES_PASSWORD in .env

Access the service S3manager

You can access the s3-ui service (s3manager) at http://localhost:8080 And then upload/download and delete files

How to use data

How to update the last image for a camera

• In s3manager, open the directory finishing by "...-alert-api-{organisation_id}" and upload the image
• In db-ui, open the table "cameras" and update
  • the column last_image with the filename from above
  • the column last_active_at

More images in the Reoling Dev Camera

you need to create a directory data/images before launching the env, with the images inside !

How to create alerts

Use one of the provided notebooks to send custom alerts manually.

For example, to send real alerts based on selected examples, run:

notebooks/send_real_alerts.ipynb

Then, you will be able to connect to the API thanks to the credentials in the .env file Note : if the notebook is run in a container, please change the following variable in the files .ipynb API_URL = "http://api:5050"

Cleanup

To stop and remove the Docker services, run:

make stop