Patriot Geek: NodeMCU

Showing posts with label NodeMCU. Show all posts

Sunday, November 26, 2017

Node-RED: Displaying MQTT Data in a Dashboard

Last time on Patriot Geek, our intrepid blogger built a MQTT client in Node-RED. Can we now present the incoming data in a useful and attractive manner? Stay tuned!

To visualize the data being sent by the BME280, we will use the "node-red-dashboard" module, which adds various nodes that generate interactive dashboards. Here's the dashboard we'll be building:

This post is divided into the following sections:

Install the dashboard extension
Understand dashboard nomenclature
Design Overview
The MQTT Subscriber Flow
Create Dashboard Tabs and Groups
The Dashboard Flow
Test the Dashboard
Conclusions and Source Code

Installation
The Node-RED IDE can be extended in various ways, from adding a single new node to adding a family of new nodes as well as adding significantly to the user interface. That's what the Node-RED Dashboard module does: it adds 16 new nodes, a new sidecar tab, and makes other changes to the IDE.

To install, click the Menu Button and choose "Manage palette". In the property sheet that opens, click the "Install" tab. Do a search for "Dashboard", and scroll until you see the "node-red-dashboard" package, and click the "install" button, then click the red "Done" button. Note: there is no need to install the "node-red-dashboard-es" package.

Dashboard Nomenclature
The use of the word "Dashboard" in Node-RED is kind of confusing, because it is used in several ways:

The module we just installed
The sidebar tab
The section of the palette, and the nodes inside it
The UIs that are built with those nodes

In addition, there are two specific terms used to described the parts of the dashboards we build: "tab" and "group".

A tab is a top-level part of the dashboard, and every dashboard must have one. When a dashboard has more than one, a menu will be shown on the left side of the title bar. Each tab has one or more "groups", which are... groups of UI controls.

Here's a tab with three groups inside it:

Here's the menu showing the two available tabs:

Design Overview
We will implement the MQTT data dashboard using two workspace tabs: one called "MQTT Subscriber" which will be a flow similar to that in the previous blog post; the other called "MQTT Dashboard" which will be for the dashboard nodes.

This design approach follows the Model-View-Controller (MVC) design pattern. The dashboard user interface is a view, and the MQTT client is essentially a controller - it processes incoming events and passes data to the view.

How do we send messages from the controller tab over to the view tab? There are a pair of nodes for doing this: the Link Out node and the Link In node. These are found under the "output" and "input" sections of the palette, respectively.

The MQTT Subscriber Flow
Create two new tabs by pressing the New Tab button. Double click on the first, and set the name to be "MQTT Subscriber". Double click the second and name it "MQTT Dashboard". Click back on the "MQTT Subscriber" tab.

Again, the MQTT Subscriber flow is a modification of the one in the previous post - all we'll be doing is adding a Link Out node and removing the two Debug nodes. Here's what the final result will look like:

After deleting the two Debug nodes, add a new Link Out node, which is found in the "Input" section of the palette. Wire it as shown. Double click it and set the name to be "To Dashboard" and click "Done". For quick reference, here are the properties of all four nodes in this flow:

Node Type

Property

Value

MQTT In

Server:

192.168.1.11:1883

Topic:

home/living-room

QoS:

CSV

Columns:

temperature,humidity,barometric-pressure

Name:

Parse CSV into JSON

Function

Name:

Add Time Info

Function:

var now = new Date();
msg.payload["timestamp"] = now.getTime();
msg.payload["fomratted-time"] = now.toUTCString();
return msg;

Link Out

Name:

To Dashboard

Create the Dashboard's Tabs and Groups
Before drawing the view flow, set up the tabs and groups for the dashboard. We will have only one tab called "Living Room Data". Inside that tab will be three groups, called "Temperature", "Humidity", and "Barometric Pressure". Here's how this will look in the Layout subtab when we're done:

Click the dashboard sidebar tab, then choose "Layout" sub tab, and use the "+ tab" button to create the tab. Set the name to be "Living Room Data". This determines the text shown in the title bar of the dashboard.

Then, hover-over the tab and click the "+ group" button three times. Hover-over each and click the "edit" button and set the names to be "Temperature", "Humidity", and "Barometric Pressure". Make sure they're all under the "Living Room Data" tab.

Click the "Site" subtab, and set the web page's title to "MQTT Data". This will be shown in the browser title bar.

While we're here, click the Theme subtab, and set the style to be "Dark". Why? Because Dark UIs Matter:

The Dashboard Flow
Click the "MQTT Dashboard" tab so we can build the dashboard flow. On it, drop one Link In node, one Debug node, three Change nodes, two Chart nodes, one Gauge node, and three Text nodes. The Chart, Gauge, and Text nodes and found in the "dashboard" section of the palette, and the Link In node is found under the "Input" section. Wire them up as follows:

We want the Link In node to receive data from the Link Out node on the other tab. Double click the Link In node, and the name we set to the Link Out node ("To Dashboard") is listed there as a checkbox. Click that checkbox. Also, set the node's name to be "From Subscriber":

Click the Link In node. When it has been successfully connected to the Link Out in the other tab, it is displayed like this, but only when selected:

Double click the Debug node, and set the out to be "complete msg object"

Now, the Chart and Gauge nodes expect their input to be in a particular format: the input must be numeric and the incoming msg.payload must be set equal that numeric value. Trouble is, the data coming from the MQTT subscriber tab (via the Link In node) is a JSON object like this:

{
    "temperature": 71.55,
    "humidity": 41.46,
    "barometric-pressure": 29.73
}

That's the purpose of the three Change nodes, to extract the numeric values from msg.payload, and replace msg.payload with those values.

Double click the first Change node, and set the properties as follows:

Update the second Change node so that it uses msg.payload.humidity. Give it the name "Extract Humidity", and set msg.payload to be msg.payload.humidity.

Name the third Change node to be "Extract Pressure", and set msg.payload to be msg.payload["barometric-pressure"]. Note the different syntax here - "barometric-pressure" is a perfectly fine JSON field name, but it is an invalid JavaScript field name!

Notice the dichotomy between visual programming languages and most traditional languages: altering msg.payload in the first Change node only has consequences for the "downstream" nodes!

The Gauge node and the Chart nodes are now getting the right data!

Now we configure the charts and the gauge. Double click the Chart node that's attached to the "Extract Temperature" Change node, and set the properties as follows:

The Chart attached to the "Extract Pressure" Change node will be set in a similar fashion, except that it will be set to appear in the "Barometric Pressure [Living Room Data]" group, and the name will be "Pressure Chart".

For the Gauge node, set the properties as shown:

Note that range for the Humidity Gauge is set to be from 0 to 100. This is because humidity runs from 0% to 100%. It is also possible to preconfigure the y-axis range on the Chart nodes but the chart will draw values outside that range as a flat line.

Our last task is to configure the Text nodes. Set their properties as follows:

Node	Property	Value
Text node for Temperature
	Group:	Temperature [Living Room Data]
	Label:	Temperature:
	Value format:	{{msg.payload}} °F
	Layout:	"row-left"
	Name:	Temperature Text
Text node for Humidity
	Group:	Humidity [Living Room Data]
	Label:	Humidity:
	Value format:	{{msg.payload}} %
	Layout:	"row-left"
	Name:	Humidity Text
Text node for Pressure
	Group:	Barometric Pressure [Living Room Data]
	Label:	Barometric Pressure:
	Value format:	{{msg.payload}} InHg
	Layout:	"row-left"
	Name:	Pressure Text

Here's what the flow will now look like:

Test the Dashboard
To see the result of all this, be sure that the MQTT broker is running, and open a new browser window and go to:

http://127.0.0.1:1880/ui

Data will start appearing in the gauge and charts. If there's not a lot of variability in the data, try (get this) breathing on the BME280! Here's what the dashboard should look like:

It works on an iPhone 5c, too:

Conclusions and Source Code
This post demonstrated how to make a dashboard that displays MQTT data coming from the BME280 and ESP8266. Something else demonstrated was how to use two separate tabs for the application, one for the controller, the other for the user interface. This use of the Model-View-Controller (MVC) design pattern adds some clarity to our flows.

The source code for the modified MQTT Subscriber flow is as follows:

[
    {
        "id": "fdb325cd.8b5748",
        "type": "mqtt in",
        "z": "553bcc1e.c97b54",
        "name": "",
        "topic": "home/living-room",
        "qos": "0",
        "broker": "a481358c.031c18",
        "x": 128,
        "y": 218,
        "wires": [
            [
                "3c175b0f.97a134"
            ]
        ]
    },
    {
        "id": "3c175b0f.97a134",
        "type": "csv",
        "z": "553bcc1e.c97b54",
        "name": "Parse CSV into JSON",
        "sep": ",",
        "hdrin": "",
        "hdrout": "",
        "multi": "one",
        "ret": "\\n",
        "temp": "temperature,humidity,barometric-pressure",
        "x": 340,
        "y": 218,
        "wires": [
            [
                "eae5505.8945fb"
            ]
        ]
    },
    {
        "id": "eae5505.8945fb",
        "type": "function",
        "z": "553bcc1e.c97b54",
        "name": "Add Time Info",
        "func": "var now = new Date();\nmsg.payload[\"timestamp\"] = now.getTime();\nmsg.payload[\"fomratted-time\"] = now.toUTCString();\nreturn msg;",
        "outputs": 1,
        "noerr": 0,
        "x": 554,
        "y": 218,
        "wires": [
            [
                "3a476f0b.d6bea"
            ]
        ]
    },
    {
        "id": "3a476f0b.d6bea",
        "type": "link out",
        "z": "553bcc1e.c97b54",
        "name": "To Dashboard",
        "links": [
            "492fdf69.9cb38",
            "4e7d2e3c.c3427"
        ],
        "x": 684,
        "y": 218,
        "wires": []
    },
    {
        "id": "a481358c.031c18",
        "type": "mqtt-broker",
        "z": "",
        "broker": "192.168.1.11",
        "port": "1883",
        "clientid": "",
        "usetls": false,
        "compatmode": true,
        "keepalive": "60",
        "cleansession": true,
        "willTopic": "",
        "willQos": "0",
        "willPayload": "",
        "birthTopic": "",
        "birthQos": "0",
        "birthPayload": ""
    }
]

Notice how there is an extra node in the JSON that is not shown in the workspace - this is where info about the MQTT broker is stored. Nodes like this are called "configuration nodes".

And here's the source code for the Dashboard flow:

[
    {
        "id": "4e7d2e3c.c3427",
        "type": "link in",
        "z": "6fc2e4b5.acb35c",
        "name": "",
        "links": [
            "3a476f0b.d6bea"
        ],
        "x": 55,
        "y": 140,
        "wires": [
            [
                "43a49c20.b2b534",
                "2d87fae4.e4c7b6",
                "ec5f8a6c.7cef28",
                "432ea3c2.6ca94c"
            ]
        ]
    },
    {
        "id": "43a49c20.b2b534",
        "type": "debug",
        "z": "6fc2e4b5.acb35c",
        "name": "",
        "active": true,
        "console": "false",
        "complete": "true",
        "x": 210,
        "y": 40,
        "wires": []
    },
    {
        "id": "2d87fae4.e4c7b6",
        "type": "change",
        "z": "6fc2e4b5.acb35c",
        "name": "Extract Temperature",
        "rules": [
            {
                "t": "set",
                "p": "payload",
                "pt": "msg",
                "to": "payload.temperature",
                "tot": "msg"
            }
        ],
        "action": "",
        "property": "",
        "from": "",
        "to": "",
        "reg": false,
        "x": 260,
        "y": 100,
        "wires": [
            [
                "5294196b.20d158",
                "d925dd10.68321"
            ]
        ]
    },
    {
        "id": "ec5f8a6c.7cef28",
        "type": "change",
        "z": "6fc2e4b5.acb35c",
        "name": "Extract Humidity",
        "rules": [
            {
                "t": "set",
                "p": "payload",
                "pt": "msg",
                "to": "payload.humidity",
                "tot": "msg"
            }
        ],
        "action": "",
        "property": "",
        "from": "",
        "to": "",
        "reg": false,
        "x": 250,
        "y": 200,
        "wires": [
            [
                "f65a9085.ae81",
                "78782f2d.880f1"
            ]
        ]
    },
    {
        "id": "432ea3c2.6ca94c",
        "type": "change",
        "z": "6fc2e4b5.acb35c",
        "name": "Extract Pressure",
        "rules": [
            {
                "t": "set",
                "p": "payload",
                "pt": "msg",
                "to": "payload[\"barometric-pressure\"]",
                "tot": "msg"
            }
        ],
        "action": "",
        "property": "",
        "from": "",
        "to": "",
        "reg": false,
        "x": 240,
        "y": 300,
        "wires": [
            [
                "ee70b8b0.828fc8",
                "c65aff2.f086b"
            ]
        ]
    },
    {
        "id": "5294196b.20d158",
        "type": "ui_chart",
        "z": "6fc2e4b5.acb35c",
        "name": "Temperature Chart",
        "group": "637855dc.2229bc",
        "order": 0,
        "width": 0,
        "height": 0,
        "label": "",
        "chartType": "line",
        "legend": "false",
        "xformat": "HH:mm:ss",
        "interpolate": "linear",
        "nodata": "",
        "dot": false,
        "ymin": "",
        "ymax": "",
        "removeOlder": "5",
        "removeOlderPoints": "",
        "removeOlderUnit": "60",
        "cutout": 0,
        "useOneColor": false,
        "colors": [
            "#1f77b4",
            "#aec7e8",
            "#ff7f0e",
            "#2ca02c",
            "#98df8a",
            "#d62728",
            "#ff9896",
            "#9467bd",
            "#c5b0d5"
        ],
        "useOldStyle": false,
        "x": 530,
        "y": 80,
        "wires": [
            [],
            []
        ]
    },
    {
        "id": "ee70b8b0.828fc8",
        "type": "ui_chart",
        "z": "6fc2e4b5.acb35c",
        "name": "Pressure Chart",
        "group": "9728196c.705ff8",
        "order": 0,
        "width": 0,
        "height": 0,
        "label": "",
        "chartType": "line",
        "legend": "false",
        "xformat": "HH:mm:ss",
        "interpolate": "linear",
        "nodata": "",
        "dot": false,
        "ymin": "",
        "ymax": "",
        "removeOlder": "5",
        "removeOlderPoints": "",
        "removeOlderUnit": "60",
        "cutout": 0,
        "useOneColor": false,
        "colors": [
            "#1f77b4",
            "#aec7e8",
            "#ff7f0e",
            "#2ca02c",
            "#98df8a",
            "#d62728",
            "#ff9896",
            "#9467bd",
            "#c5b0d5"
        ],
        "useOldStyle": false,
        "x": 520,
        "y": 280,
        "wires": [
            [],
            []
        ]
    },
    {
        "id": "f65a9085.ae81",
        "type": "ui_gauge",
        "z": "6fc2e4b5.acb35c",
        "name": "Humidity Gauge",
        "group": "f468b186.a264b",
        "order": 0,
        "width": 0,
        "height": 0,
        "gtype": "gage",
        "title": "Gauge",
        "label": "%",
        "format": "{{value}}",
        "min": 0,
        "max": "100",
        "colors": [
            "#00b500",
            "#e6e600",
            "#ca3838"
        ],
        "seg1": "",
        "seg2": "",
        "x": 520,
        "y": 180,
        "wires": []
    },
    {
        "id": "d925dd10.68321",
        "type": "ui_text",
        "z": "6fc2e4b5.acb35c",
        "group": "637855dc.2229bc",
        "order": 0,
        "width": 0,
        "height": 0,
        "name": "Temperature Text",
        "label": "Temperature:",
        "format": "{{msg.payload}} &deg;F",
        "layout": "row-left",
        "x": 530,
        "y": 120,
        "wires": []
    },
    {
        "id": "78782f2d.880f1",
        "type": "ui_text",
        "z": "6fc2e4b5.acb35c",
        "group": "f468b186.a264b",
        "order": 0,
        "width": 0,
        "height": 0,
        "name": "Humidity Text",
        "label": "",
        "format": "{{msg.payload}} %",
        "layout": "row-left",
        "x": 520,
        "y": 220,
        "wires": []
    },
    {
        "id": "c65aff2.f086b",
        "type": "ui_text",
        "z": "6fc2e4b5.acb35c",
        "group": "9728196c.705ff8",
        "order": 0,
        "width": 0,
        "height": 0,
        "name": "Pressure Text",
        "label": "Barometric Pressure:",
        "format": "{{msg.payload}} InHg",
        "layout": "row-left",
        "x": 520,
        "y": 320,
        "wires": []
    },
    {
        "id": "637855dc.2229bc",
        "type": "ui_group",
        "z": "",
        "name": "Temperature",
        "tab": "d376d602.4e3538",
        "order": 1,
        "disp": true,
        "width": "6"
    },
    {
        "id": "9728196c.705ff8",
        "type": "ui_group",
        "z": "",
        "name": "Barometric Pressure",
        "tab": "d376d602.4e3538",
        "order": 3,
        "disp": true,
        "width": "6"
    },
    {
        "id": "f468b186.a264b",
        "type": "ui_group",
        "z": "",
        "name": "Humidity",
        "tab": "d376d602.4e3538",
        "order": 2,
        "disp": true,
        "width": "6"
    },
    {
        "id": "d376d602.4e3538",
        "type": "ui_tab",
        "z": "",
        "name": "Living Room Data",
        "icon": "dashboard",
        "order": 2
    }
]

Here, we have four extra configuration nodes: one for the tab and three for the groups.

Saturday, November 25, 2017

Node-RED: Subscribing to a MQTT Topic

For our next Node-RED flow, we will build a MQTT client that will subscribe to BME280 sensor data published by an ESP8266. The hardeare and code for publishing that data is described in an earlier post.

The messages published to the MQTT topic home/living-room are CSV strings containing temperature, humidity, and barometric pressure in British units. Here's an example of such a message:

71.55,41.46,29.73

Our goal for this post is to build a Node-RED flow that subscribes to that topic, and parses that data into JSON like this:

{
    "temperature": 71.55,
    "humidity": 41.46,
    "barometric-pressure": 29.73
}

For this flow, we'll be using the following three types of nodes:

MQTT In: Connects to a MQTT server and subscribes to a specified topic
CSV: Converts CSV data into JSON
Debug: Displays messages in the debug tab

These nodes will be connected as follows:

Double-click the MQTT In node and set the following properties:

Server: 192.168.1.11:1883

Topic: home/living-room

QoS: 0

Next, double-click the CSV node and set the following properties:

Columns: temperature,humidity,barometric-pressure

Separator: Comma

Name: Parse CSV into JSON

Finally, double-click each of the Debug nodes and set the following properties:

Output: complete msg object

to: debug tab

The CSV node uses the column names for the field names, and assigns values to those fields in the order of the values in the incoming CSV. The end result is set to be the message's payload. This is how the messages look in the debug tab:

Pretty-printing the top entry in the debug list shows that the msg object has the following structure:

{
  "topic": "home/living-room",
  "payload":
  {
    "temperature": 73.8,
    "humidity": 37.12,
    "barometric-pressure": 29.6
  },
  "qos": 0,
  "retain": false,
  "_msgid": "1c091ae3.456e45"
}

Great, we've subscribed to a MQTT topic and formatted the incoming CSV data as a JSON object!

Wouldn't it be nice if there were timestamps in these objects? As explained in the post where we built and coded the sensor, our little ESP8266 doesn't have a clock running on battery, nor are we getting time from a network time service. Instead, we will add the timestamp to the data once it arrives in our Node-RED flow.

To accomplish this, we will use a Function node, which will run a block of JS. Wire this Function node into the rest of the flow as follows:

Double-click on the Function node, and set the name to be Add Time Info. In the "Function" textarea, add the following code:

var now = new Date();
msg.payload["timestamp"] = now.getTime();
msg.payload["fomratted-time"] = now.toUTCString();
return msg;

All that this code does is to add two additional properties to msg.payload. The JavaScript .getDate() method returns the number of milliseconds since January 1, 1970, and the .toUTCString() formats that data into a human-readable form in the GMT timezone.

Clear the debug tab, deploy the edited flow, and examine the incoming data. Here's an example of what the final result will be, once we pretty-print it:

{
  "topic": "home/living-room",
  "payload": 
  {
    "temperature": 73.76,
    "humidity": 37.35,
    "barometric-pressure": 29.6,
    "timestamp": 1511588927869,
    "formatted-time": "Sat, 25 Nov 2017 05:48:47 GMT"
  },
  "qos": 0,
  "retain": false,
  "_msgid":"3867cbee.a2cfa4"
}

Here's the source code for this last flow:

[
    {
        "id": "ee409dc6.f511b",
        "type": "mqtt in",
        "z": "e260f053.ca405",
        "name": "",
        "topic": "home/living-room",
        "qos": "0",
        "broker": "a481358c.031c18",
        "x": 120,
        "y": 160,
        "wires": [
            [
                "f90ccd18.b0f84",
                "b7c62375.17de5"
            ]
        ]
    },
    {
        "id": "f90ccd18.b0f84",
        "type": "debug",
        "z": "e260f053.ca405",
        "name": "",
        "active": false,
        "console": "false",
        "complete": "true",
        "x": 770,
        "y": 160,
        "wires": []
    },
    {
        "id": "b7c62375.17de5",
        "type": "csv",
        "z": "e260f053.ca405",
        "name": "Parse CSV into JSON",
        "sep": ",",
        "hdrin": "",
        "hdrout": "",
        "multi": "one",
        "ret": "\\n",
        "temp": "temperature,humidity,barometric-pressure",
        "x": 380,
        "y": 220,
        "wires": [
            [
                "5097ac3c.7d4b04"
            ]
        ]
    },
    {
        "id": "adf30d5c.857be",
        "type": "debug",
        "z": "e260f053.ca405",
        "name": "",
        "active": true,
        "console": "false",
        "complete": "true",
        "x": 770,
        "y": 220,
        "wires": []
    },
    {
        "id": "5097ac3c.7d4b04",
        "type": "function",
        "z": "e260f053.ca405",
        "name": "Add Time Info",
        "func": "var now = new Date();\nmsg.payload[\"timestamp\"] = now.getTime();\nmsg.payload[\"fomratted-time\"] = now.toUTCString();\nreturn msg;",
        "outputs": 1,
        "noerr": 0,
        "x": 600,
        "y": 220,
        "wires": [
            [
                "adf30d5c.857be"
            ]
        ]
    },
    {
        "id": "a481358c.031c18",
        "type": "mqtt-broker",
        "z": "",
        "broker": "192.168.1.11",
        "port": "1883",
        "clientid": "",
        "usetls": false,
        "compatmode": true,
        "keepalive": "60",
        "cleansession": true,
        "willTopic": "",
        "willQos": "0",
        "willPayload": "",
        "birthTopic": "",
        "birthQos": "0",
        "birthPayload": ""
    }
]

What are we going to do with all this data flowing in? That's the topic of the next post!

Friday, November 24, 2017

Publishing Sensor Data to MQTT

When publishing sensor data to MQTT, two question to ask are: how often is the data published?, and with what distribution?

Here are two examples:

Publish temperature, humidity, and barometric pressure from one sensor every second
Publish and verify RFID card swipes that employees use when entering a building.

In the first example, we have a good idea of the frequency of the data as well as how that data is distributed: data is sent once every second and it is uniformly distributed - in essence, it is time series data. In the second example, the frequency of data transmission depends on the time of day. Just before the business day starts, there will be a large number of RFID card swipes as the employees "clock in", then for the rest of the day the number of RFID events will be far lower.

The answers to those two questions will determine various aspects of the backend IoT architecture: the database that should be used to store and otherwise interact with this sensor data, the "analytics pipeline" (my term, ©2017, Patriot Geek) used to analyze and visualize that data, etc.

For this blog post, we will publish time series data using our old friends, the BME280 and the ESP8266! Please see this earlier blog post for wiring and importing the proper libraries:

http://patriot-geek.blogspot.com/2017/03/nodemcu-and-digital-sensors.html

Temperature, humidity, and barometric pressure will be published once every second to a MQTT broker. Compared to industrial IoT applications, this is not a lot of data! We can thus use a database like MongoDB or even MySQL for storage. The analytics pipeline will be discussed in a later post.

The next question to ask is: how should that data be formatted? One solution is to use a JSON format:

{
    "temperature": 71.55,
    "humidity": 41.46,
    "barometric-pressure": 29.73
}

An alternative is to use a CSV format like this:

71.55,41.46,29.73

The JSON format contains a human-readable description of those three values, whereas the CSV format contains considerably fewer characters. Since MQTT is all about short messages, we will use the CSV format. This will also serve us well should we want to use LoRa to send the same information. If for any reason we want to reformat the CSV into JSON, we can do so later.

Here is the code that will have our ESP8266 publish this data:

/**
 * BME280-MQTT.js
 *
 * By: Mike Klepper
 * Date: 24 November 2017
 *
 * This program demonstrates how to publish to MQTT
 *
 * See blog post on patriot-geek.blogspot.com
 * for instructions.
 */


#include "ESP8266WiFi.h"
#include "PubSubClient.h"
#include "Wire.h"
#include "Adafruit_Sensor.h"
#include "Adafruit_BME280.h"

WiFiClient wifiClient;
PubSubClient mqttClient(wifiClient);
Adafruit_BME280 bme;

const char* SSID                  = "**********";
const char* PASSWORD              = "**********";

const char* MQTT_BROKER           = "192.168.1.11";
const int   MQTT_PORT             = 1883;
const char* MQTT_TOPIC            = "home/living-room";
const char* MQTT_CLIENT_NAME      = "ESP8266Client";

const int   BME280_DELAY          = 3000;
const int   RECHECK_INTERVAL      = 500;
const int   PUBLISH_INTERVAL      = 1000;

long        lastReconnectAttempt  = 0;
long        lastPublishAttempt    = 0;


void setup() 
{
  Serial.begin(115200);
  delay(10);

  setupWifi();
  setupBME280();
  setupMQTT();
}

void loop() 
{
  long now = millis();
  
  if(!mqttClient.connected()) 
  {
    if(now - lastReconnectAttempt > RECHECK_INTERVAL) 
    {
      lastReconnectAttempt = now;
      mqttClient.connect(MQTT_CLIENT_NAME);

      if(mqttClient.connected())
      {
        // Resubscribe to any topics, if necessary
        // This is also a good place to publish an error to a separate topic!
      }
    }
  }
  else
  {
    if(now - lastPublishAttempt > PUBLISH_INTERVAL)
    {
      lastPublishAttempt = now;
      readAndPublishData();
      mqttClient.loop();
    }
  }
}


void setupWifi()
{
  Serial.println("");
  Serial.print("Connecting to ");
  Serial.print(SSID);

  WiFi.begin(SSID, PASSWORD);

  while(WiFi.status() != WL_CONNECTED) 
  {
    delay(RECHECK_INTERVAL);
    Serial.print(".");
    yield();
  }

  Serial.println("");
  Serial.println("WiFi connected");
  Serial.print("IP address: ");
  Serial.println(WiFi.localIP());
}

void setupBME280()
{
  Serial.println("");
  Serial.print("Setting up BME280");

  while(!bme.begin())
  {
    delay(RECHECK_INTERVAL);
    Serial.print(".");
    yield();
  }
  
  Serial.println("");
  Serial.println("BME280 found!");
  delay(BME280_DELAY);
}

void setupMQTT()
{
  mqttClient.setServer(MQTT_BROKER, MQTT_PORT);
  
  Serial.println("");
  Serial.print("Connecting to MQTT");
  
  while(!mqttClient.connected())
  {
    Serial.print(".");
    mqttClient.connect(MQTT_CLIENT_NAME);
    delay(RECHECK_INTERVAL);
  }
  
  Serial.println("");
  Serial.println("Connected!");
  Serial.println("");
}

void readAndPublishData()
{
  if(mqttClient.connected())
  {
    float tempC = bme.readTemperature();
    float tempF = 9.0/5.0 * tempC + 32.0;
    float humidity = bme.readHumidity();
    float pressurePascals = bme.readPressure();
    float pressureInchesOfMercury = 0.000295299830714 * pressurePascals;
  
    String msg = String(tempF) + "," + String(humidity) + "," + String(pressureInchesOfMercury);
    char msgAsCharAway[msg.length()];
    msg.toCharArray(msgAsCharAway, msg.length());
  
    Serial.print(msg);
    mqttClient.publish(MQTT_TOPIC, msgAsCharAway);
    Serial.println(" - sent");
  }
}

Here's a walkthrough of the code. The functions for initializing the wifi and the BME280 are the same as in previous posts; the new stuff is how we publish to MQTT, and how we time those publications.

14 - 18	The various libraries we'll be using
24 - 25	Name and password for wifi network
45 - 47	Connect to wifi, start the BME280, and connect to MQTT broker
120	Initialize the MQTT client, specifying the broker's IP address and port
125	While we're not connected...
128	...attempt to connect
139	If the MQTT client is connected...
141 - 145	...read data from the BME280 and convert to British units...
147 - 149	...format the data as CSV, and prepare it for publication...
152	...and publish it to the broker!
50 - 77	This loop function is based on the non-blocking reconnect example from the PubSubClient library
52	...get the number of milliseconds since this sketch started running...
54	If we're not connected to the broker...
56	...check to see how long it has been since we last tried reconnecting; if it has been more than a specified time...
58	...reset the timer...
59	...and try to reconnect
61	If we were successful...
63 - 64	...perform any housekeeping activities, like reporting an error
68	Otherwise we're connected...
70	...check to see how long it has been since we last published data; if sufficient time has passed...
72	...reset the timer...
73	...publish the data...
74	...and process any incoming messages.

To test the application, first open a terminal window and start MQTT broker with the command:

mosquitto

Open another terminal window and start a MQTT client and subscribe to the topic home/living-room:

mosquitto_sub -h localhost -t home/living-room

Finally, watch the incoming data!

Some notes about this application:

Data is transmitted in an insecure manner!
A timestamp is not included in the data.

These two issues will be addressed in a later blog post.

Columns:	temperature,humidity,barometric-pressure
Separator:	Comma
Name:	Parse CSV into JSON