JTK – Notes and Ramblings Things we have found …

The original garage pi was a model 1 – 26 pin GPIO, 2 USB A, composite video, full size SD card. It served admirably for many years but I was looking to add to its responsibilities. So a Raspberry Pi 3 is replacing it.

The Pi 3 had been in the drawer for a year or so with raspian from late 2018. A full sudo apt-get update/upgrade/dist-upgrade later and we are ready to start.

sudo apt-get install mosquitto mosquitto-clients libmosquitto-dev gets MQTT ready to go.

The outdoor i2c temp sensor will need the library from https://bitbucket.org/loblik/libam2315/src/master/ . So wget https://bitbucket.org/loblik/libam2315/get/701be2837287.zip downloads the archive, a quick unzip and we are ready to install.

sudo make install is the magic for installing the AM2315 library we will need for the sensor.

Using the gateway C code we long ago modified from the original RFM69 Gateway Pi port, we compile with first gcc -lam2315 am2315-get.c -o am2315-get and then gcc piGateway.c -o piGateway -lwiringPi -lmosquitto -lam2315 -DRASPBERRY -DDEBUG.

A simple shell script to start things and we are done with the temp/humidity sensor. sudo ~/piGateway2/piGateway /dev/i2c-1

For clarity – the connection to the sensor uses the Pi I2C connection:

• Pin 4 – Vcc (5v) [orange & orange/white]
• Pin 6 – GND [brown & brown/white]
• Pin 3 – SDA1 (GPIO2) [green]
• Pin 5 – SCL1 (GPIO3) [green/white]

Update: So all of that is brilliant but the real reason for the upgrade was to use the RTL-SDR – well the reception is very poor from the garage so we are back to the original Pi 1 Model B V 2.0.

The code at https://github.com/K2DLS/noaacap looks like a good start on grabbing the relevant Weather Alerts from https://alerts.weather.gov.

For my current location I am using https://alerts.weather.gov/cap/wwaatmget.php?x=OHC055&y=1 or https://alerts.weather.gov/cap/wwaatmget.php?x=OHZ013&y=1. Changing the parameter y=1 to y=0 will yield the atom/xml feed instead of the html.

After getting the Python sorted out to read the Seneye Home it was time resolve a few outstanding issues with the unit. Seneye has only released the partial API discussed previously. This API lacks the ability to 1) register a slide 2) clear the “offline” readings stored in the unit.

Following ideas found online I setup VirtualHere. There is a Raspberry Pi server and a Windows 10 client. The free version will allow for a for sharing a single device. The free version also doesn’t support running the client as a service on Windows 10. Neither of these limitation was a problem for our setup.

I setup a cron job to run every 30 minutes. It calls a shell script which stops the VirtualHere server, runs the Seneye Python code to send reading to Openhab via MQTT and restarts the VirtualHere server.

crontab -e
*/30 * * * * ~/fish.sh

fish.sh
date >>~/Seneye-MQTT/suddriver.log
sudo killall vhusbdarm
sleep 10
python ~/Seneye-MQTT/suddriver.py >>~/Seneye-MQTT/suddriver.log
sleep 10
sudo ~/VirtualHere/vhusbdarm -b

It would be nice not to need sudo in this script but at the moment I haven’t noodled out the permissions required to get the VitrualHere server to run without sudo.

The Seneye Home normally wants to be connected to a Windows PC to log its reading to the cloud. Naturally, I see no reason for the cloud to be involved. {Insert rant about unnecessary cloud connections for home devices here} Fortunately, Seneye has released at least a partial API for reading on a Raspberry Pi.

https://github.com/seneye/SUDDriver

There is a handy page describing compiling and testing the samples from the above: https://raspberrypidiyblog.wordpress.com/2018/01/17/seneye-suddriver-raspberrypi/

Compiling the code required I change line 464 of linux/hid.c from:

tmp = (hid_device_info*)malloc(sizeof(struct hid_device_info));

to

tmp = (struct hid_device_info *)malloc(sizeof(struct hid_device_info));

Over at https://doughall.me/2017/06/30/seneyemqttpt1/ is a discussion on using Python on the Pi to get the device readings. His code can be found at https://github.com/dhallgb/Seneye-MQTT.

Following these instructions I installed python modules pprint, pyusb, paho-mqtt and bitstring(not mentioned but required) using pip install. It was necessary to update the 10-local.rules file to indicate the correct idVendor and idProduct: idVendor=24f7 and idProduct=2204. These need to be in hex in the udev files. They are in decimal in the python source.

I have also updated the python script to not use bitstring and to use struct instead.

def mungReadings(p):
   s = {}
   scrap, scrap, TS, flag1, flag2, scrap, pH, NH3, thermo, scrap  = unpack('<BBLBBHHHL46s', p)
   s['InWater']=bool(flag1 & 0b00000100)
   s['SlideNotFitted']=bool(flag1 & 0b00001000)
   s['SlideExpired']=bool(flag1 & 0b00010000)
   s['pH']=float(pH)/100   # divided by 100
   s['NH3']=float(NH3)/1000  # divided by 1000
   s['Temp']=float(thermo)/1000 # divided by 1000
   j = json.dumps(s, ensure_ascii=False)
   return(j)

Still working on how to register a Seneye slide without the Windows PC. From the Seneye SUDDriver wiki:

Slide codes are retrieved from an API hosted on the seneye cloud. In order to use it, your account will need activating to allow permission. You can request access to this by E-mail by asking for “SUD Developer Access” and we will send you the terms and conditions to be sent to you for agreement.

https://github.com/seneye/SUDDriver/wiki/Activating-a-seneye-slide-for-your-SUD#retrieving-a-slide-code