WatcherEye

RESTful API with Flask that serves images from Raspberry Pi powered by picamera.

Setup

This software is intended to run in the following environment:

1. Raspberry Pi 3 Model B v1.2
2. Picamera v2

The testing environment for this software is a Macbook, the built-in camera is used.

Installation

All dependencies are specified in requirements.txt. Install all python packages:

pip install -r requirements.txt

Running

1. If using virtualenvwrapper: workon WatcherEye
2. Set Flask app environment variable: export FLASK_APP=watchereye
3. Set Flask app development variable: export FLASK_ENV=info
4. Start the Flask server: flask run
5. Point your favorite browser to localhost:5000

Note that flask in debug mode only exposes the web service locally on the host. To have the web service accessible from external IPs (i.e. using the public IP of the server), you may need to run flask with the --host=0.0.0.0 flag.

API Reference

The API can be accessed via HTTP/S requests, see watchereye/api.pyfor up-to-date API endpoint documentation. Briefly:

Method	Endpoint	Description	Return Format
GET	/api/v1/ping	Ping API server	JSON array
GET	/api/v1/stats	Usage statistics	JSON array
GET	/api/v1/image	Capture and return image	image/jpeg
GET	/api/v1/stream	View video stream	multipart/x-mixed-replace; image/jpeg

Logging

Logging at the debug level (recording all incoming requests and IPs) using the default python logging facility. Logs are stored in logs/log.log and streamed to STDOUT

Test

Using PyTest:

1. workon WatcherEye
2. python -m pytest -s -v --disable-pytest-warnings tests/

References

1. Installing Raspbian on Raspberry Pi on Mac OS X
2. Video Streaming with Flask
3. Flask Video Streaming Revisited

Development Notes

Update requirements.txt after installing any new pip modules using: pip freeze > requirements.txt