Sigfox-gcloud-data NPM

sigfox-gcloud-data is a sigfox-gcloud adapter for writing Sigfox messages into SQL databases like MySQL and Postgres.

You may read and update Sigfox messages with other modules (such as sigfox-gcloud-ubidots) before passing to sigfox-gcloud-data for writing to the database. sigfox-gcloud-data works with most SQL databases supported by Knex.js like MySQL, Postgres, MSSQL, MariaDB and Oracle.

sigfox-gcloud-data was built with sigfox-gcloud, an open-source software framework for building a Sigfox server with Google Cloud Functions and Google Cloud PubSub message queues. Check out sigfox-gcloud

sigfox-gcloud-data with MySQL:

sigfox-gcloud-data with Postgres:

Releases

Version 1.0.1 (14 Oct 2017): Supports multiple instances

Getting Started

For development we support Linux, MacOS and Ubuntu on Windows 10. Open a command prompt and enter these commands to download the sigfox-cloud-data source folder to your computer.

git clone https://github.com/UnaBiz/sigfox-gcloud-data.git
cd sigfox-gcloud-data

If you're using Ubuntu on Windows 10, we recommend that you launch "Bash on Ubuntu on Windows" and enter the following commands to download the source files into the folder /mnt/c/sigfox-gcloud-data:

cd /mnt/c
git clone https://github.com/UnaBiz/sigfox-gcloud-data.git
cd sigfox-gcloud-data

That's because /mnt/c/sigfox-gcloud-data under bash is a shortcut to c:\sigfox-gcloud-data under Windows.
So you could use Windows Explorer and other Windows tools to browse and edit files in the folder. Remember to use a text editor like Visual Studio Code that can save files using the Linux line-ending convention (linefeed only: \n), instead of the Windows convention (carriage return + linefeed: \r \n).

Setting up Google Cloud

Install sigfox-gcloud with the base modules (exclude optional modules):
https://github.com/UnaBiz/sigfox-gcloud/blob/master/README.md
Open a bash command prompt. For Windows, open "Bash on Ubuntu on Windows."
Create a file named .env in the sigfox-gcloud-data folder
and populate the GCLOUD_PROJECT variable with your project ID. To do that, you may use this command (change myproject to your project ID):
```
cd sigfox-gcloud-data
echo GCLOUD_PROJECT=myproject >.env
```
Add the following sigfox-route setting to the Google Cloud Project Metadata store. This route says that all received Sigfox messages will be processed by the two steps decodeStructuredMessage and sendToDatabase.
```
gcloud compute project-info add-metadata --metadata=^:^sigfox-route=decodeStructuredMessage,sendToDatabase
```
If you're using sigfox-gcloud-ubidots, the sendToDatabase step should appear last so that the updates from sendToUbidots will be recorded in the database.
```
gcloud compute project-info add-metadata --metadata=^:^sigfox-route=decodeStructuredMessage,sendToUbidots,sendToDatabase
```
Create the Google PubSub message queue that we will use to route the Sigfox messages between the Cloud Functions:
```
gcloud beta pubsub topics create sigfox.types.sendToDatabase
```
sigfox.devices.sendToDatabase is the queue that will receive decoded Sigfox messages to be sent to data via the data API
Go to the Google Cloud Metadata screen to define your database settings:
https://console.cloud.google.com/compute/metadata
- sigfox-dbclient: Database client library to be used e.g mysql, pg. Check this page for the library: http://knexjs.org/#Installation-node
- sigfox-dbhost: Address of database server e.g. 127.127.127.127
- sigfox-dbuser: User ID for accessing the database e.g. user
- sigfox-dbpassword: Password for accessing the database.
- sigfox-dbname: Name of the database that will store the sensor data. Defaults to sigfox
- sigfox-dbtable: Name of the table to store sensor data. Defaults to sensordata
- sigfox-dbversion: Version number of database, used only by Postgres, e.g. 7.2
If the sigfox-dbtable table above does not exist, it will be created automatically.
Install the database library if you are NOT using MySQL or Postgres. Check this page for the library to be used:
http://knexjs.org/#Installation-node
Then run the command npm install LIBRARYNAME --save. For example if you're using MSSQL, you would run this command:
```
npm install mssql --save
```
Deploy the sendToDatabase Cloud Function with the deployall.sh script:
```
chmod +x */*.sh
scripts/deployall.sh
```

How it works

Sigfox messages are pushed by the Sigfox Cloud to the Google Cloud Function sigfoxCallback
Cloud Function sigfoxCallback delivers the message to PubSub message queue sigfox.devices.all, as well as to the device ID and device type queues
Cloud Function routeMessage listens to PubSub message queue sigfox.devices.all and picks up the new message
Cloud Function routeMessage assigns a route to the Sigfox message by reading the sigfox-route from the Google Compute Metadata Store. The route looks like this:
```
decodeStructuredMessage, sendToDatabase
```
This route first sends the message to function decodeStructuredMessage via the queue sigfox.types.decodeStructuredMessage
decodeStructuredMessage contains the logic to decode a compressed message format that we call Structured Message Format. Within a 12-byte Sigfox message, the Structured Message Format can encode efficiently 3 sensor field values and their sensor field names.
For example, the encoded 12-byte message b0513801a421f0019405a500 contains 3 sensor values (temperature, humidity, altitude) and their field names: tmp = 31.2, hmd = 49.6, alt = 16.5
According to sigfox-route above, the resulting decoded message is sent next to function sendToDatabase via the queue sigfox.types.sendToDatabase
sendToDatabase appends the received Sigfox message to the sensordata table that you have defined in the Google Cloud Metadata settings. It calls the Knex.js library to update the database.
sendToDatabase automatically matches the received Sigfox message fields with the sensordata fields. So if your Sigfox message includes a new field (perhaps by decoding a Structured Message) and the sensordata table also contains a field by that name, sendToDatabase will write the new field into the sensordata table.
See this doc for the definition of Structured Messages:
https://unabiz.github.io/unashield/
To understand how Structured Messages may be used with the Ubidots IoT platform, check the UnaShield Tutorial for Ubidots:
https://unabiz.github.io/unashield/ubidots

Viewing `sigfox-gcloud-data` server logs

You may view the logs through the Google Cloud Logging Console
Select "Cloud Function" as the "Resource"

From the screen above you can see the logs generated as each Sigfox message is processed in stages by sigfox-gcloud:

Sigfox Device IDs are shown in square brackets e.g. [ 2C30EB ]
Completed Steps are denoted by _<<_
sigfoxCallback is the Google Cloud Function that listens for incoming HTTPS messages delivered by Sigfox
routeMessage passes the Sigfox message to various Google Cloud Functions to decode and process the message
decodeStructuredMessage decodes a compressed Sigfox message that contains multiple field names and field values
sendToDatabase would appear after decodeStructuredMessage. sendToDatabase writes the decoded sensor data to the database via the Knex.js library.

Tracing `sigfox-gcloud-data` server performance

The Google Cloud Trace Console shows you the time taken by each step of the Sigfox message processing pipeline, tracing the message through every Google Cloud Function.

Each message delivered by Sigfox appears as a separate trace timeline. Messages are shown like 2C30EB seq:1913 where 2C30EB is the Sigfox Device ID and 1913 is the Sigfox Message Sequence Number (seqNumber)

The Google Stackdriver Trace API needs to be enabled manually.

Custom reports may be created in Google Cloud Trace Control to benchmark the performance of each processing step over time.

Understanding and troubleshooting the `sigfox-gcloud-data` server

To understand each processing step in the sigfox-gcloud-data server, you may use the Google Cloud Debug Console to set breakpoints and capture in-memory variable values for each Google Cloud Function, without stopping or reconfiguring the server.

In the example below, we have set a breakpoint in the sigfoxCallback Google Cloud Function. The captured in-memory values are displayed at right - you can see the Sigfox message that was received by the callback. The Callback Stack appears at the lower right.

Google Cloud Debug is also useful for troubleshooting your custom message processing code without having to insert the debugging code yourself.

Testing the `sigfox-gcloud-data` server

Send some Sigfox messages from the Sigfox devices. Monitor the progress of the processing through the Google Cloud Logging Console.
Select "Cloud Function" as the "Resource"
Processing errors will be reported to the Google Cloud Error Reporting Console.
We may configure Google Cloud Stackdriver Monitoring to create incident reports upon detecting any errors. Stackdriver may also be used to generate dashboards for monitoring the PubSub message processing queues.

Demo

To send messages from a Sigfox device into your database, you may use this Arduino sketch:
https://github.com/UnaBiz/unabiz-arduino/blob/master/examples/send-light-level/send-light-level.ino
The sketch sends 3 field names and field values, packed into a Structured Message:
```
ctr - message counter
lig - light level, based on the Grove analog light sensor
tmp - module temperature, based on the Sigfox module's embedded temperature sensor        
```

Alternatively, you may test by sending a Sigfox message from your Sigfox device with the data field set to:

920e82002731b01db0512201

We may also use a URL testing tool like Postman to send a POST request to the sigfoxCallback URL e.g. (change myproject to your Google Cloud Project ID)

https://us-central1-myproject.cloudfunctions.net/sigfoxCallback

Set the Content-Type header to application/json. If you're using Postman, click Body -> Raw -> JSON (application/json) Set the body to:

{
  "device":"1A2345",
  "data":"920e82002731b01db0512201",
  "time":"1476980426",
  "duplicate":"false",
  "snr":"18.86",
  "station":"0000",
  "avgSnr":"15.54",
  "lat":"1",
  "lng":"104",
  "rssi":"-123.00",
  "seqNumber":"1492",
  "ack":"false",
  "longPolling":"false"
}

where device is your Sigfox device ID.

Here's the request in Postman:

We may use the curl command as well. Remember to change myproject and 1A2345 to your project ID and device ID.

curl --request POST \
  --url https://us-central1-myproject.cloudfunctions.net/sigfoxCallback \
  --header 'cache-control: no-cache' \
  --header 'content-type: application/json' \
  --data '{"device":"1A2345", "data":"920e82002731b01db0512201", "time":"1476980426", "duplicate":"false", "snr":"18.86", "station":"0000", "avgSnr":"15.54", "lat":"1", "lng":"104", "rssi":"-123.00", "seqNumber":"1492", "ack":"false", "longPolling":"false"}'

The response from the callback function should look like this:
```
{
  "1A2345": {
    "noData": true
  }
}
```
```
   
```
The test message sent above will be decoded and written to your sensordata table as
```
ctr (counter): 13
lig (light level): 760
tmp (temperature): 29        
```
The other fields of the Sigfox message will be written as well.

Adding one or more instances of `sendToDatabase`

It's possible to run 2 or more Cloud Functions that will update different databases. The Cloud Functions should be named:

sendToDatabase, sendToDatabase2, sendToDatabase3, ...

and the configuration for each function shall be set in the Google Cloud Metadata screen as

sigfox-dbclient, sigfox-dbclient2, sigfox-dbclient3, ...

For example, this metadata screen defines 2 databases settings for MySQL and Postgres:

To deploy the second instance of sendToDatabase, edit the script scripts/deploy.sh and uncomment the second functiondeploy so it looks like:

./scripts/functiondeploy.sh ${name}2  ${localpath} ${trigger} ${topic}

Run scripts/deploy.sh. This will deploy a new function sendToDatabase2 that uses the second database setting in the Google Cloud Metadata screen.

To deploy sendToDatabase3, sendToDatabase4, ... you may edit scripts/deploy.sh accordingly:

./scripts/functiondeploy.sh ${name}3  ${localpath} ${trigger} ${topic}
./scripts/functiondeploy.sh ${name}4  ${localpath} ${trigger} ${topic}

Note that all instances of sendToDatabase will read Sigfox messages from the sigfox.types.sendToDatabase queue simultaneously. The database updates will run in parallel.