Sending pictures with Yowsup

One feature that could be quite useful in home automation is the ability to send pictures took with a webcam to our smartphone.

This could happen on request, by just asking to take a photo with our parser, or automatically when the script is programmed to do that (at specified intervals or maybe when detecting movements).

Sending pictures using Yowsup is not too difficult, but to make it easier for all of you I created the file with class declaration. You can download it here:

http://www.mascal.it/public/wamedia.py

With this script you will be able to send pictures. But there's more: you can use it to send also audio messages and videos!

If you wish you can examine that script, but you can use it as is without analyzing what's inside. If you have an updated Raspbian you should have all the dependecies already installed. Otherwise you should install Pillow

To use it you first need to import the script into your project, so at the beginning section of the parser add the following line:

from wamedia import *

To be able to take pictures from the camera module you should also add the following:

import picamera

and remember to enable the camera module using raspi-config.

I will not show you how to parse a command to take a photo as now you should be able to add it by yourself. You can always refer to http://rpihome.blogspot.com/2015/01/using-whatsapp-for-home-automation.html if you don't remember how to proceed.

Let's see a simple function to take a photo and send it to our smartphone:

def TakePhoto(received):

    Answer("Your photo will be sent soon")

    camera=picamera.PiCamera()

    camera.rotation=180

    camera.resolution=(640, 480)

    time.sleep(2)

    camera.capture("lastimage.jpg", format='jpeg', quality=75, thumbnail=None)

    camera.close()

    if os.path.exists("lastimage.jpg"):

        try:

            stack = SendMediaStack(credential(), [(["390000000000", "lastimage.jpg"])])

            stack.start()

        except ImageSent as e:

            if ("ERROR" in e.value): Answer("Could not send the photo, sorry!")

    else:

        Answer("I found no photo to send!")

    return

The first thing I'm doing is sending a message to inform that the photo is arriving, as the process could take some time. You can delete this if you want.

Then I create a camera object and set a couple of options. I rotate the image as my camera is upside down and set the resolution to values lower than the defaults, to keep the file small and speed up the things a bit.

There are several more options you could set. Just refer to the official documentation.

Now it's time to get the picture and save it to a file. You can save the image wherever you wish of course. You can also change the quality and the format of the resulting file. Just remember that Whatsapp accepts just jpeg and png images.

After closing the camera object I check if the picture file exists, just in case there were some problem with its creation. If there is no file I send an error message to the smartphone.

If the file exists we can try to send it. The statement is quite simple and you should just change the target phone number and the image filename to adapt to your system.

Afer the execution of the class an ImageSent exception is raised with the result of the operation. If there was any problem we can find the word "ERROR" inside the error value and we can send an error message.

Of course you can add some features by just checking for key words in the received message. For example you could add an option to take high resolution pictures or maybe one to take photos in low light conditions...

As i wrote, you can also send video and audio files. Just change the filename and the class will recognize what to send by checking the file extension.

These are the valid extension for the class:

• image: jpg, png
• audio: mp3, wav, aac, wma, ogg, oga
• video: mp4

So you could ask your RPi to make a video and send it to you, instead of a photo. Think about this setup:

A Raspberry Pi with a camera module mounted on a servo. You ask for a video and the script records it while rotating the camera. Then RPi will send a panoramic shot to your smartphone.

Another thing you could do (a bit less useful, bit nice!) is sending you vocal messages instead of text ones. You just need a tts engine able to create an audio file and change the Answer function to send the voice message.
In this case it would be useful to have the file in a compressed format to save space and reduce the time needed to send it (so avoid uncompressed wav if you can).

On one of the next posts I will describe how to install the Pico TTS engine from Android. This is a quite good engine, but unfortunately you can't install it in an easy way, so probably a post about its installation could be useful to some of you.