Reply
Highlighted
<a href="http://community.silabs.com/t5/Welcome-and-Announcements/Community-Ranking-System-and-Recognition-Program/m-p/140490#U140490"><font color="#000000"><font size="2">Star Employee</font></font> </a> DDB
Posts: 3
Registered: ‎01-29-2017

Thunderboard Sense with Raspberry Pi and Python

The goal of this project was to make a very simple python script that runs on a Raspberry Pi and collects data from one or more Thunderboard Sense devices, using the same Google Firebase backend and web application that the official app uses.

 

To get started, go to https://github.com/siliconlabs/thundercloud, and clone the repository. Follow the instructions and set up a firebase account and project, replacing the database name in src/main.js as instructed. For this example, I have not yet gotten authentication working, so change the database.rules.json for now to allow anyone to write (basically replacing "auth.uid === 'YsGcsiI8hkwjImSrr25uZdqlNw3Qkgi8vUWx9MU6'": with "true")

 

{
  "rules": {
    ".read": false,
    ".write": false,
    "thunderboard":{
      ".read": true,
      ".write": true
    },
    "topCharts":{
      ".read": true,
      ".write": true
    },
    "sessions":{
      ".write": true,
      ".indexOn": ["startTime", "contactInfo/deviceName"],
      "$session":{//2592000000 is 30 days
        ".read": "data.child('startTime').val() > (now - 2992000000)",
        ".write":  true
      }
    }
  }
}

 

Now, deploy the app to Firebase using the firebase tools

 

 

sudo npm install -g firebase
firebase login
firebase deploy --project <your firebase project id>

 

In the Firebase console you should now see your deployed rules, as well as the application url. Following the URL should bring you to the default page:

Screenshot 2017-03-09 01.35.43.png

Now the next step is to log onto the Raspberry Pi, and install the necessary tools. The script uses bluepy to communicate with the sensor, and python-firebase to push data to the cloud. I did have some trouble installing python-firebase because of the specific version of the requests library, but eventually got it installed.

 

Remember to put in your own database URL in the thundercloud.py script:

from firebase import firebase
import uuid
import time

class Thundercloud:

   def __init__(self):

      self.addr     = 'https://-- Your database URL --.firebaseio.com/'
      self.firebase = firebase.FirebaseApplication(self.addr, None)

 

The tbsense_scan.py script continuously looks for advertising thunderboards, and automatically connects to the board if a new one is discovered.

 

$ sudo python3 tbsense_scan.py 
No Thunderboard Sense devices found!
No Thunderboard Sense devices found!
Starting thread <Thread(Thread-1, initial)> for 37020

Thunder Sense #37020
Ambient Light:	0.25 Lux
Sound Level:	37.83
tVOC:		0
Humidity:	28.25 %RH
Temperature:	29.86 C
Pressure:	951.586
UV Index:	0
eCO2:		0

When a Thunderboard Sense has been discovered and connected to, the script will print out the read data periodically. Take note of the number in "Thunder Sense #37020". We will need to give the number to the web application.

 

The script generates a session ID for each board connected, and then continuously generates json strings for the data read from the Thunderboard Sense. The json string is then inserted into the appropriate location in the database. Firebase has a useful live database view that shows us that our data is indeed being pushed into the cloud:

Screenshot 2017-03-09 01.20.16.png

Finally if you go to the app url and append your Thunderboard Sense id, you should see the data being displayed https://<project id>.firebaseapp.com/#/37020

Screenshot 2017-03-09 01.19.34.png

 Screenshot 2017-03-09 01.19.47.png

 tbsense.py simply discovers and sets up handles to the different characteristics. It also contains functions to read these characteristics:

 

 

from bluepy.btle import *
import struct
from time import sleep

class Thunderboard:

   def __init__(self, dev):
      self.dev  = dev
      self.char = dict()
      self.name = ''
      self.session = ''
      self.coinCell = False

      # Get device name and characteristics

      scanData = dev.getScanData()

      for (adtype, desc, value) in scanData:
         if (desc == 'Complete Local Name'):
            self.name = value

      ble_service = Peripheral()
      ble_service.connect(dev.addr, dev.addrType)
      characteristics = ble_service.getCharacteristics()

      for k in characteristics:
         if k.uuid == '2a6e':
            self.char['temperature'] = k
            
         elif k.uuid == '2a6f':
            self.char['humidity'] = k
            
         elif k.uuid == '2a76':
            self.char['uvIndex'] = k
            
         elif k.uuid == '2a6d':
            self.char['pressure'] = k
            
         elif k.uuid == 'c8546913-bfd9-45eb-8dde-9f8754f4a32e':
            self.char['ambientLight'] = k

         elif k.uuid == 'c8546913-bf02-45eb-8dde-9f8754f4a32e':
            self.char['sound'] = k

         elif k.uuid == 'efd658ae-c401-ef33-76e7-91b00019103b':
            self.char['co2'] = k

         elif k.uuid == 'efd658ae-c402-ef33-76e7-91b00019103b':
            self.char['voc'] = k

         elif k.uuid == 'ec61a454-ed01-a5e8-b8f9-de9ec026ec51':
            self.char['power_source_type'] = k

   
   def readTemperature(self):
      value = self.char['temperature'].read()
      value = struct.unpack('<H', value)
      value = value[0] / 100
      return value

   def readHumidity(self):
      value = self.char['humidity'].read()
      value = struct.unpack('<H', value)
      value = value[0] / 100
      return value

   def readAmbientLight(self):
      value = self.char['ambientLight'].read()
      value = struct.unpack('<L', value)
      value = value[0] / 100
      return value

   def readUvIndex(self):
      value = self.char['uvIndex'].read()
      value = ord(value)
      return value

   def readCo2(self):
      value = self.char['co2'].read()
      value = struct.unpack('<h', value)
      value = value[0]
      return value

   def readVoc(self):
      value = self.char['voc'].read()
      value = struct.unpack('<h', value)
      value = value[0]
      return value

   def readSound(self):
      value = self.char['sound'].read()
      value = struct.unpack('<h', value)
      value = value[0] / 100
      return value

   def readPressure(self):
      value = self.char['pressure'].read()
      value = struct.unpack('<L', value)
      value = value[0] / 1000
      return value

 

 

thundercloud.py handles the connection to the Firebase database. getSession() generates a new session ID and is called once for every new Thunderboard Sense connection. putEnvironmentData() inserts the data and updates the timestamps:

 

 

from firebase import firebase
import uuid
import time

class Thundercloud:

   def __init__(self):

      self.addr     = 'https://'-- Firebase Database Name --'.firebaseio.com/'
      self.firebase = firebase.FirebaseApplication(self.addr, None)

   def getSession(self, deviceId):

      timestamp = int(time.time() * 1000)
      guid = str(uuid.uuid1())

      url = 'thunderboard/{}/sessions'.format(deviceId)
      self.firebase.put(url, timestamp, guid)

      
      d = {
            "startTime" : timestamp,
            "endTime" : timestamp,
            "shortURL": '',
            "contactInfo" : {
                 "fullName":"First and last name",
                 "phoneNumber":"12345678",
                 "emailAddress":"name@example.com",
                 "title":"",
                 "deviceName": 'Thunderboard #{}'.format(deviceId)
             },
             "temperatureUnits" : 0,
             "measurementUnits" : 0,
         }

      url = 'sessions'
      self.firebase.put(url, guid, d)

      return guid

   def putEnvironmentData(self, guid, data):

      timestamp = int(time.time() * 1000)
      url = 'sessions/{}/environment/data'.format(guid)
      self.firebase.put(url, timestamp, data)

      url = 'sessions/{}'.format(guid)
      self.firebase.put(url, 'endTime', timestamp)

 

Finally, tbsense_scan.py continuously searches for new Thunderboard Sense devices, and spawns a new thread for each one successfully connected to:

 

from bluepy.btle import *
import struct
from time import sleep
from tbsense import Thunderboard
from thundercloud import Thundercloud
import threading

def getThunderboards():
    scanner = Scanner(0)
    devices = scanner.scan(3)
    tbsense = dict()
    for dev in devices:
        scanData = dev.getScanData()
        for (adtype, desc, value) in scanData:
            if desc == 'Complete Local Name':
                if 'Thunder Sense #' in value:
                    deviceId = int(value.split('#')[-1])
                    tbsense[deviceId] = Thunderboard(dev)

    return tbsense

def sensorLoop(fb, tb, devId):

    session = fb.getSession(devId)
    tb.session = session
    
    value = tb.char['power_source_type'].read()
    if ord(value) == 0x04:
        tb.coinCell = True

    while True:

        text = ''

        text += '\n' + tb.name + '\n'
        data = dict()

        try:

            for key in tb.char.keys():
                if key == 'temperature':
                        data['temperature'] = tb.readTemperature()
                        text += 'Temperature:\t{} C\n'.format(data['temperature'])

                elif key == 'humidity':
                    data['humidity'] = tb.readHumidity()
                    text += 'Humidity:\t{} %RH\n'.format(data['humidity'])

                elif key == 'ambientLight':
                    data['ambientLight'] = tb.readAmbientLight()
                    text += 'Ambient Light:\t{} Lux\n'.format(data['ambientLight'])

                elif key == 'uvIndex':
                    data['uvIndex'] = tb.readUvIndex()
                    text += 'UV Index:\t{}\n'.format(data['uvIndex'])

                elif key == 'co2' and tb.coinCell == False:
                    data['co2'] = tb.readCo2()
                    text += 'eCO2:\t\t{}\n'.format(data['co2'])

                elif key == 'voc' and tb.coinCell == False:
                    data['voc'] = tb.readVoc()
                    text += 'tVOC:\t\t{}\n'.format(data['voc'])

                elif key == 'sound':
                    data['sound'] = tb.readSound()
                    text += 'Sound Level:\t{}\n'.format(data['sound'])

                elif key == 'pressure':
                    data['pressure'] = tb.readPressure()
                    text += 'Pressure:\t{}\n'.format(data['pressure'])

        except:
            return

        print(text)
        fb.putEnvironmentData(session, data)
        sleep(1)


def dataLoop(fb, thunderboards):
    threads = []
    for devId, tb in thunderboards.items():
        t = threading.Thread(target=sensorLoop, args=(fb, tb, devId))
        threads.append(t)
        print('Starting thread {} for {}'.format(t, devId))
        t.start()


if __name__ == '__main__':
    
    fb = Thundercloud()

    while True:
        thunderboards = getThunderboards()
        if len(thunderboards) == 0:
            print("No Thunderboard Sense devices found!")
        else:
            dataLoop(fb, thunderboards)

 

Posts: 2
Registered: ‎01-09-2017

Re: Thunderboard Sense with Raspberry Pi and Python

Hi Mr. DDB,

you did a gat job in publishing this code.

 

I'm trying to interface the Thunderboard Sense board using Node.js with the final target to deliver a nodered node for it, based on NOBLE for Bluetooth 4.0 interface.

I've been partially sucessfull at the moment in reading some environmnet data.

 

I'd like to be able to drive GPIO signals, so to be able to drive relays (or somthing else) directrly soldered  to the board.

I cannot find any hint about which UUIS Services and Characteristics that should be used in order to achieve this.

I've done no changes to the Sense factory firmware.

Sure you have far btter experience then me ... so I'd appreciate a suggestion.

 

Thanks very much.

Marco

Posts: 17
Registered: ‎11-11-2016

Re: Thunderboard Sense with Raspberry Pi and Python

Hi DDB,

Nice job.Is there any update on this project?

Thanks

Posts: 17
Registered: ‎11-11-2016

Re: Thunderboard Sense with Raspberry Pi and Python

Hi @DDB,

Anybody from Silicon Labs please upgrade thundercloud github.

https://github.com/SiliconLabs/thundercloud

 

The files from github are different from thundercloud.silabs.com site.

 

Best regards,

Marcio

 

 

 

Posts: 6
Registered: ‎06-07-2017

Re: Thunderboard Sense with Raspberry Pi and Python

I was able to get through the tutorial and make a little help doc where I ran into troubles. It might help you. I am still very new with all of this stuff so I am by no means an expert at this stuff

 

https://gist.github.com/hornej/bf56d8103a1c2d685579f3628949a442


msam wrote:

Hi @DDB,

Anybody from Silicon Labs please upgrade thundercloud github.

https://github.com/SiliconLabs/thundercloud

 

The files from github are different from thundercloud.silabs.com site.

 

Best regards,

Marcio