UI Notes
Micasaverde (Talk | contribs) |
Andreimios (Talk | contribs) (→lu_status) |
||
(57 intermediate revisions by 8 users not shown) | |||
Line 1: | Line 1: | ||
+ | [[Category:Development]] | ||
Here is some general information about getting information to display in a user interface: | Here is some general information about getting information to display in a user interface: | ||
− | == | + | == UPnP messaging system == |
− | + | Vera uses UPnP for all the messaging, such as turning lights on and off. So Vera translates Z-Wave, Insteon, infrared codes, etc., into UPnP actions. If you have already implemented a UPnP stack and can send properly formated UPnP SOAP/XML requests, you can control everything using standard UPnP action invocation. If you did not, Vera provides a simple HTTP GET interface as well, which is discussed below. | |
− | + | In a nutshell UPnP works like this: You have UPnP devices, which are things that can be controlled and can report their state, like a light switch. And you have UPnP Control Points, which are the things that let a user control UPnP devices, like a touch-screen web pad. The Control Point does this by a) reading the device and descriptions for the UPnP device to see what it can do, and b) invoking actions on the UPnP device (ie telling it to do something), and c) reading and setting variables for the device which describe the current state and/or configuration settings for the device. | |
− | + | A UPnP device is defined in an XML document called a device description document. To see one, in Vera's setup UI, go to Devices, Luup plugins, Luup Files, and click 'view' file next to D_BinaryLight1.xml. Use Firefox since it has a built-in XML viewer. Notice the tag: <deviceType>urn:schemas-upnp-org:device:BinaryLight:1</deviceType>. This defines the type of device, namely an on/off light switch, which is what you use to know how to represent this device to a user. A device optionally implements one or more services. In this case the services: urn:upnp-org:serviceId:SwitchPower1, urn:micasaverde-com:serviceId:EnergyMetering1 and urn:micasaverde-com:serviceId:HaDevice1. | |
− | The | + | A service is a collection of actions which a device implement and which you can invoke from a UPnP Control Point, as well as a collection of variables which always have a current value. The variables describe the current state of the device and/or configuration settings. View the file S_SwitchPower1.xml to see a service. For example, the device urn:schemas-upnp-org:device:BinaryLight:1 implements the service urn:upnp-org:serviceId:SwitchPower1, which has the action SetTarget, which is how you turn the light on and off. The action takes a single argument newTargetValue which can be 0 to turn the light off and 1 to turn it on. The service also has the variable Status which is 0 or 1 depending on whether the light is currently on or off. |
− | + | If you open the device description for a dimmable light D_DimmableLight1.xml you'll see that it implements the same service S_SwitchPower1.xml, since you can still turn it on and off, and it also implements the service urn:upnp-org:serviceId:Dimming1, which has an action to set the dim level (SetLoadLevelTarget). | |
− | + | To test UPnP calls you can use the Intel Device Spy utility available [http://www.intel.com/cd/ids/developer/asmo-na/eng/downloads/upnp/tools/218896.htm?desturl=http://softwarecommunity.intel.com/isn/downloads/zips/IntelToolsForUPnPTechnology_v2.zip here]. It should pick up Vera, show you all the devices, like light switches, thermostats, etc., let you expand the services and see the actions and variables. Click an action, like SetTarget, to specify the arguments and click 'Invoke' to run the action, such as turning on and off the light. | |
− | + | == JSON vs. XML == | |
− | + | The native UPnP protocol uses XML/SOAP. But there are also several times you will request data from Vera that has nothing to do with the UPnP spec. These will be done with data_request's as explained below, such as user_data, lu_status, etc. Internally Vera uses JSON as the native format, but Vera can convert to XML (details follow). In general when you want to view data in a human readable format, such as the device list, it's easiest to use the Firefox browser, and request Vera convert the data to XML. Firefox has a nice browser to display XML in human readable format. If you want to make the JSON data nicely formatted, you can use the web site jsonlint.com. In a UI application which will request data regularly from Vera we recommend retrieving everything in JSON format, not XML, because Vera's CPU is not that powerful and so continuous requests of data in XML format will cause a lot of CPU time to be devoted to converting JSON into XML. | |
− | == | + | == Finding Vera on the local network == |
− | + | You can find Vera by using the traditional UPnP discovery process, or you can retrieve this URL:<br>'''https://sta1.mios.com/locator_json.php?username=xxx''' where the username=xxx is optional, but, if specified, it will report all systems that mios.com user xxx can report whether local or remote, and without the username=xxx it only shows the Vera's on the user's local home network. | |
− | = | + | Each Vera, when it boots up, reports its internal IP address to the central mios.com server, which tracks this along with the external IP address. '''locator.php''' shows all serial numbers and internal network IP addresses on the same external IP. For example, if '''https://sta1.mios.com/locator_json.php?username=xxx'''. |
− | The | + | == The HTTP interface == |
− | + | Vera listens on port 3480. Vera responds to UPnP actions the normal way, or by making simple HTTP requests. For example, to turn on a light you can either send the UPnP action "SetTarget" to the device using the normal UPnP mechanism, or you can simply put this URL in your web browser, or with wget, or similar: | |
− | + | ||
− | + | ||
− | + | ||
− | + | ||
− | + | ||
− | + | ||
− | + | ||
− | + | ||
− | == | + | http://[ipadress]:3480/data_request?id=lu_action&output_format=xml&DeviceNum=10&serviceId=urn:upnp-org:serviceId:SwitchPower1&action=SetTarget&newTargetValue=1 |
− | + | That URL will return the same results of the action as using the normal UPnP architecture. The '''output_format=xml''' on the URL means the data returned should be in XML format. You can instead use '''output_format=json''' | |
− | ' | + | The service ID's, actions, and arguments are all defined by the UPnP forum. A list of ratified UPnP devices and services is available on the [http://www.upnp.org UPnP Forum]. Mi Casa Verde added its own custom devices and services when there was no existing UPnP standard. A list of these can be found in the [[Luup_UPNP_Files]] header files. |
− | + | == Getting the system configuration == | |
− | ''' | + | Either use the UPnP service/action:<br> '''urn:micasaverde-com:serviceId:HomeAutomationGateway1/GetUserData''' |
− | + | or use the built-in request: | |
− | ''' | + | '''http://ipaddress:3480/data_request?id=user_data2&output_format=json''' (or xml) |
− | + | This returns the system configuration file for Vera. It's assumed you understand the way Vera categorizes devices with Rooms, Devices, Scenes, etc. When creating a user interface for controlling Vera only, like on a mobile phone, most of the tags can be ignored, since usually you just need the list of rooms and devices. | |
− | + | == Accessing Vera remotely through the mios server == | |
− | + | Native UPnP only works on a local area home network, and cannot be used over the Internet. The mios server provides a secure way to remotely access and control your Vera system using the HTTP interface without having to make any substantial changes to the UI. Everything you can do locally with Vera on port 3480, you can do remotely with mios using the exact same syntax. You only need to ask the user for his username and password and pass them on the URL to the remote access server, along with the serial number of the unit (ie like 12345). For example, the above user_data is the syntax for local access on a home network. To retrieve the same URL over the internet with mios use: | |
− | + | https://fwd2.mios.com/demovera/myvera123/12345/data_request?id=lu_status&output_format=json (or use fwd1.mios.com) | |
− | + | assuming demovera is the username and myvera123 is the password that the user chose setting up his mios.com account. Note that since the request is https, the username and password are encrypted because https encrypts the URL's as well as the contents of the page. | |
− | + | == Building a control-only UI for Vera == | |
− | + | If you just want to control Vera, and not modify the configuration, such as with a cell phone interface or web pad, the flow might possibly work like this: | |
− | ''' | + | '''1)''' For the initial setup, you need to ask whether the user wants to (a) access Vera remotely with the mios service, and if so, get the user's username/password; or (b) access Vera directly on the home network. You can use locator.php to show the Vera on the home network. You might want to do an auto-detect. For example, you could use locator.php to see if Vera exists locally, and if not, prompt the user for a mios username and password. You can use locator.php and pass the username to get the serial number and internal IP, and then automatically switch to direct access if locator.php (without the username) indicates that serial number is found on the local network. This would allow the user to automatically switch between local access with Wi-Fi (which is much faster of course) and remote access with mios and the mobile phone network. Or just ask for his mios username and show him all the Vera's he can control, either local or remote. |
− | ''' | + | '''2)''' Once the connection is established, retrieve the list of rooms, scenes, and devices from Vera (see '''user_data''' below). You can cache this data locally so you don't need to retrieve it and parse it over and over. |
− | ''' | + | '''3)''' Present the user with the rooms, scenes, devices. In Vera's UI's we show first the rooms and then let the user pick a device or scene in the room. However this data could be presented graphically, such as on a floorplan. Your UI can have Vera store extra tags and data in the configuration database (ie user_data) to store things like coordinates for icons, outlines or rooms and so on. |
− | ''' | + | '''4)''' If the user picks a scene you can run the scene with the following URL, passing the scene number as '''x''', and it will return an OK if it went okay: |
− | + | http://127.0.0.1:3480/data_request?id=lu_action&serviceId=urn:micasaverde-com:serviceId:HomeAutomationGateway1&action=RunScene&SceneNum=x | |
− | ''' | + | '''5)''' If the user picks a device, you use the category which is contained in the user_data to determine what type of controls to present the user. (See [[Luup UPNP Files]] for a list of the categories.) When the user chooses an action, you can send it like this: |
− | + | http://ipadress:3480/data_request?id=lu_action&output_format=xml&DeviceNum=10&serviceId=urn:upnp-org:serviceId:SwitchPower1&action=SetTarget&newTargetValue=1 | |
− | + | For a list of common actions and device types see: [[Luup CP Mandatory Types]] | |
− | + | Some actions return immediately with an 'OK'. Other jobs may take time to complete and are processed asynchronously. For these asynchronous jobs you will get a job ID back. There are several ways to present this to the user. One possibility is to do what Vera does in the UI, namely to use '''lu_status '''(explained next) to find all active jobs and to put a 'job' icon next to each device to indicate the progress of the job. In this way the user isn't tracking a particular job ''per se'', but just has visual feedback that a device is busy, and if the job icon(s) turn green, that everything went through okay. You could instead track the action individually using the job number that is returned, or use some other method, like making the icon for the device appear animated while it's busy, and turn green or red if the last command succeeded or failed. | |
− | + | '''6)''' Periodically poll the system with '''lu_status''' (explained below), like this: | |
− | + | ||
− | + | http://ipaddress:3480/data_request?id=lu_status&output_format=json | |
− | + | ||
− | + | ||
− | + | ||
− | + | ||
− | + | ||
− | + | ||
− | + | ||
− | + | ||
− | + | ||
− | + | ||
− | + | That returns the current values of the UPnP variables, which you can use to show the user current status. For example, if the variable Status is 1 for a switch, you can show the 'on' button in a selected state, while 0 shows it in an 'off' state. '''lu_status''' also shows any active jobs for a device. | |
− | + | ==user_data== | |
− | + | The user_data is the main configuration file that Vera uses to store all the rooms, devices, scenes, timers, events, users, and any other data about the system. You should request the user_data to get whatever data you need to present to the user, at a minimum, for example, you'll need the list of devices, their descriptions, and the categories (light switch, thermostat, etc.). You'll also want the list of scenes to present the user. To fetch the current user_data from Vera use the user_data request, like this: | |
− | + | Returns JSON (preferred native format): | |
− | + | ||
− | + | ||
− | + | ||
− | + | ||
− | + | ||
− | + | ||
− | + | http://[ip address]:3480/data_request?id=user_data | |
− | + | Returns XML: | |
− | + | http://[ip address]:3480/data_request?id=user_data&output_format=xml | |
− | = | + | You can also add a 'DataVersion=TimeStamp' parameter to the URL. It the data has changed since the TimeStamp, all the data is returned. If the data has not changed since the TimeStamp, Vera will return a page with the text 'NO_CHANGES': |
− | + | http://[ip address]:3480/data_request?id=user_data&DataVersion=315587159 | |
− | + | The user_data contains the state for all the devices. For example this indicates that the device id #17, called "Switch" is turned off (ie service SwitchPower1, variable Status is 0) | |
− | + | ||
− | " | + | { |
− | " | + | "devices": [ |
− | " | + | { |
− | + | "id": 17, | |
− | + | "category": 3, | |
− | + | "room": 1, | |
− | + | "name": "Switch", | |
− | + | "ip": "", | |
− | + | "mac": "", | |
− | + | { | |
− | }, | + | "service": "urn:upnp-org:serviceId:SwitchPower1", |
− | + | "variable": "Status", | |
− | " | + | "value": "0" |
− | " | + | } |
− | " | + | } |
+ | ] | ||
+ | } | ||
+ | |||
+ | {{Warning|On UI7 build lu_sdata and lu_status have been deprecated, and you should use "status" instead.}} | ||
+ | |||
+ | ==lu_status== | ||
+ | The UPnP variables describe the state of the devices according to the UPnP specs. When the variables change that information needs to be shown to the user, eg change a Switch icon from off to on. If you are interfacing with Vera using the UPnP stack, you can subscribe to the device's events to receive notifications automatically when a device's state changes. However if you're using the web interface you will need to poll using status request. | ||
+ | |||
+ | Call lu_status request like so: | ||
+ | |||
+ | :http://[ip address]:3480/data_request?id=status | ||
+ | |||
+ | :http://[ip address]:3480/data_request?id=status&output_format=xml | ||
+ | |||
+ | :http://[ip address]:3480/data_request?id=status&DataVersion=315587159 | ||
+ | |||
+ | |||
+ | The returned document contains four sets of information, which constantly change to reflect the current status: | ||
+ | * root attributes including DataVersion, which is the status timestamp, plus other attributes such as LoadTime | ||
+ | *the state of the system initialization, such as initialization of the plugin modules. | ||
+ | *a '''subset''' of the user_data that represents the current state of the device's UPnP variables | ||
+ | *the status of any active or recent '''jobs''', such as if a light was successfully turned off | ||
+ | |||
+ | ===Tracking status changes=== | ||
+ | If you call lu_status without passing a DataVersion on the URL, you will get the current value of all the UPnP variables for a device. Note that most of the UPnP variables do not ever change. For example, with a light switch, only the variable "SwitchPower:Status" changes, as the switch is turned off and on. There are many other variables, like "Capabilities" which are simply configuration data or settings the user modified. If you pass &DataVersion=1 on the URL, you will only see the UPnP variables which have changed since the Luup engine started. So, if you parse the initial UPnP variables when you first call user_data, you can call lu_status with DataVersion=1. Once you call lu_status the first time, store the DataVersion and on subsequent calls pass it on the URL and you will only get the UPnP variables which have changed since that last DataVersion. Regardless, you will also get the current state of jobs and system initialization. | ||
+ | |||
+ | On subsequent calls to lu_status, you should also pass back in the LoadTime which you received on the prior call. | ||
+ | |||
+ | ===Controlling the poll timing=== | ||
+ | You can also pass in a Timeout value, which means the call will block for up to that many seconds if nothing has changed, and will return immediately when something does. This way you can have infrequent polling, say every 60 seconds by putting a Timeout value of 60, but your UI will be immediately responsive since, when something changes, it returns immediately. You can also add a MinimumDelay, which means that the call will wait for at least that many milliseconds. This is advisable so that in the event there are lots of continuous changes, such as a scene turning off many lights, you won't have lu_status calls returning non-stop consuming all the resources on both the Vera and the UI. | ||
+ | |||
+ | ===Startup status=== | ||
+ | The section with the initialization tasks is shown in the startup tag as shown: | ||
+ | |||
+ | "startup": { | ||
+ | "tasks": [ | ||
+ | { | ||
+ | "id": 0, | ||
+ | "status": 2, | ||
+ | "type": "ZWave", | ||
+ | "comments": "No Z-Wave" | ||
+ | }, | ||
+ | |||
+ | The startup tasks begin whenever the Luup engine is reset, such as after the user makes a configuration change. Ever plugin and module in the Luup engine does a startup sequence. The id is not really important. The status numbers are the same as for jobs, so 2 means the module failed to startup. The 'Type' is what you can use to key off of to know which module is being reported, such as "ZWave", and comments are just notes indicating the current state. For example, "No Z-Wave" means there's no Z-Wave dongle connected. | ||
+ | |||
+ | Once all the plugins have been initialized successfully the startup section will disappear completely from lu_status. | ||
+ | |||
+ | Note that lu_status also returns UserData_DataVersion, which is the dataversion from user_data. That way you don't need to poll user_data in parallel. Simply retrieve the user_data one time at startup, then poll lu_status, and, if UserData_DataVersion is different from the previous user_data you retrieved, then download user_data again. | ||
+ | |||
+ | Some notes on building a full replacement, including configuration, are available here [[UI_Notes_Replacement]], however the syntax is not up to date with the Luup architecture. | ||
+ | |||
+ | ===Job status=== | ||
+ | The jobs for a device will look like this: | ||
+ | |||
+ | "devices": [ | ||
+ | { | ||
+ | "id": 1, | ||
+ | "Jobs": [ | ||
+ | { | ||
+ | "id": "1", | ||
+ | "name": "job#1 :getnodes (0x00D068F0) P:10 S:4", | ||
+ | "icon": "", | ||
+ | "comments": "", | ||
+ | "status": "4" | ||
} | } | ||
− | + | ] | |
}, | }, | ||
+ | |||
+ | The id is the job id number. When you run an action, like turning on a light, if it happens asynchronously, you'll get a job id number. All pending jobs for a device are shown by their id number. The name is just an internal name for the job. Icon is an optional keyword to indicate what sort of icon should be shown to the user. For example, 'getnodes' has no icon, meaning this type of job is generally not of interest to a user. If the icon was 'ON' or 'OFF' that would mean this job is busy turning on or off a light. Comments is notes about the job, if any. This may be shown to a user in a tool tip or similar, but is generally more technical than users would care to know. The status indicates the current state of the job, as shown: | ||
+ | |||
+ | *job_None=-1, // no icon | ||
+ | *job_WaitingToStart=0, // gray icon | ||
+ | *job_InProgress=1, // blue icon | ||
+ | *job_Error=2, // red icon | ||
+ | *job_Aborted=3, // red icon | ||
+ | *job_Done=4, // green icon | ||
+ | *job_WaitingForCallback=5 // blue icon - Special case used in certain derived classes | ||
+ | |||
+ | If the status is 0, this means the job is queued but hasn't started. If it's 1 or 5 the job is being executed right now. Generally both 1 and 5 are presented to the user the same way as a 'busy', although technically 1 means Vera is actively talking to the device, and 5 means Vera is waiting for a reply from the device. Status codes 2, 3 and 4 mean the job has finished. 4 means the job went ok, and both 2 and 3 mean it failed. Technically 3 means the job was aborted by another job or an external process, like a user cancelling it, while 2 means it failed due to a problem talking to the device. In both cases this is usually presented the user the same way, as a failure. | ||
+ | |||
+ | == Categories and device types == | ||
+ | |||
+ | The values used in user_data to determine the types of devices are in [[Luup_UPNP_Files]]. In general you can use the category instead of the devicetype since it's an integer and is thus faster to compare: | ||
+ | |||
+ | == pseudo-code for a UI == | ||
+ | |||
+ | Here is some pseudo code to show how the polling mechanism should be properly implemented. The parsing code is, of course, unique to your UI, so placeholder function names are shown here, like FetchAndParse. This shows the flow for handling the polling so you know you get the changes reported correctly. | ||
+ | |||
+ | variable int DataVersion_lustatus=0 | ||
+ | variable int LoadTime=0 | ||
+ | variable int DataVersion_userdata=0 | ||
+ | |||
+ | :START | ||
+ | variable boolean FetchUserData=false; | ||
+ | if( DataVersion_userdata==0 ) // If we're on the first loop and haven't retrieved user_data, do it and don't do an lu_status | ||
+ | FetchUserData=true; | ||
+ | else | ||
+ | { | ||
+ | while(true) | ||
+ | { | ||
+ | newdata = FetchAndParse("lu_status&DataVersion="+DataVersion_lustatus+"&LoadTime="+LoadTime); | ||
+ | if( newdata.IsEmpty()==false && newdata.isError()==false ) | ||
+ | break; // If we got data, continue on and process it | ||
+ | else | ||
+ | Sleep 1; // If we didn't get anything, or if lu_status failed, just sleep 1 second and stay in the loop to try again | ||
+ | } // End of while loop | ||
+ | |||
+ | if( newdata.UserData_DataVersion!=DataVersion_userdata ) | ||
+ | { | ||
+ | FetchUserData=true; | ||
+ | } | ||
+ | else // Only Parse if userdata is the same, otherwise we're going to reload user_data | ||
+ | { | ||
+ | ParseLuStatus(); // This won't run if user_data has changed | ||
+ | DataVersion_lustatus=newdata.DataVersion | ||
+ | } | ||
+ | } // end of else loop | ||
+ | |||
+ | if( FetchUserData==true ) | ||
+ | { | ||
+ | UI.HourglassAndSuspend(true); // Put a 'busy' hourglass in the UI, display a message in the info panel "Updating data" and suspend all UI stuff since this may take a moment | ||
+ | newdata = FetchAndParse("user_data2"); | ||
+ | if( newdata.IsEmpty()==false && newdata.isError()==false ) | ||
+ | { | ||
+ | ParseUserData(); | ||
+ | DataVersion_lustatus=newdata.DataVersion | ||
+ | LoadTime=newdata.LoadTime | ||
+ | DataVersion_userdata=newdata.DataVersion | ||
+ | } | ||
+ | else | ||
+ | UI.DisplayError("Failed to update data"); // Display in the info panel an error message that the user will see for 10 seconds or so | ||
+ | |||
+ | UI.HourglassAndSuspend(false); // Put a 'busy' hourglass in the UI, display a message in the info panel "Updating data" and suspend all UI stuff since this may take a moment | ||
+ | } | ||
+ | |||
+ | Sleep 1; // We know we just parsed user_data or lu_status, so sleep before looping | ||
+ | goto START; // Loop again and start over |
Latest revision as of 12:27, 3 July 2015
Here is some general information about getting information to display in a user interface:
Contents |
[edit] UPnP messaging system
Vera uses UPnP for all the messaging, such as turning lights on and off. So Vera translates Z-Wave, Insteon, infrared codes, etc., into UPnP actions. If you have already implemented a UPnP stack and can send properly formated UPnP SOAP/XML requests, you can control everything using standard UPnP action invocation. If you did not, Vera provides a simple HTTP GET interface as well, which is discussed below.
In a nutshell UPnP works like this: You have UPnP devices, which are things that can be controlled and can report their state, like a light switch. And you have UPnP Control Points, which are the things that let a user control UPnP devices, like a touch-screen web pad. The Control Point does this by a) reading the device and descriptions for the UPnP device to see what it can do, and b) invoking actions on the UPnP device (ie telling it to do something), and c) reading and setting variables for the device which describe the current state and/or configuration settings for the device.
A UPnP device is defined in an XML document called a device description document. To see one, in Vera's setup UI, go to Devices, Luup plugins, Luup Files, and click 'view' file next to D_BinaryLight1.xml. Use Firefox since it has a built-in XML viewer. Notice the tag: <deviceType>urn:schemas-upnp-org:device:BinaryLight:1</deviceType>. This defines the type of device, namely an on/off light switch, which is what you use to know how to represent this device to a user. A device optionally implements one or more services. In this case the services: urn:upnp-org:serviceId:SwitchPower1, urn:micasaverde-com:serviceId:EnergyMetering1 and urn:micasaverde-com:serviceId:HaDevice1.
A service is a collection of actions which a device implement and which you can invoke from a UPnP Control Point, as well as a collection of variables which always have a current value. The variables describe the current state of the device and/or configuration settings. View the file S_SwitchPower1.xml to see a service. For example, the device urn:schemas-upnp-org:device:BinaryLight:1 implements the service urn:upnp-org:serviceId:SwitchPower1, which has the action SetTarget, which is how you turn the light on and off. The action takes a single argument newTargetValue which can be 0 to turn the light off and 1 to turn it on. The service also has the variable Status which is 0 or 1 depending on whether the light is currently on or off.
If you open the device description for a dimmable light D_DimmableLight1.xml you'll see that it implements the same service S_SwitchPower1.xml, since you can still turn it on and off, and it also implements the service urn:upnp-org:serviceId:Dimming1, which has an action to set the dim level (SetLoadLevelTarget).
To test UPnP calls you can use the Intel Device Spy utility available here. It should pick up Vera, show you all the devices, like light switches, thermostats, etc., let you expand the services and see the actions and variables. Click an action, like SetTarget, to specify the arguments and click 'Invoke' to run the action, such as turning on and off the light.
[edit] JSON vs. XML
The native UPnP protocol uses XML/SOAP. But there are also several times you will request data from Vera that has nothing to do with the UPnP spec. These will be done with data_request's as explained below, such as user_data, lu_status, etc. Internally Vera uses JSON as the native format, but Vera can convert to XML (details follow). In general when you want to view data in a human readable format, such as the device list, it's easiest to use the Firefox browser, and request Vera convert the data to XML. Firefox has a nice browser to display XML in human readable format. If you want to make the JSON data nicely formatted, you can use the web site jsonlint.com. In a UI application which will request data regularly from Vera we recommend retrieving everything in JSON format, not XML, because Vera's CPU is not that powerful and so continuous requests of data in XML format will cause a lot of CPU time to be devoted to converting JSON into XML.
[edit] Finding Vera on the local network
You can find Vera by using the traditional UPnP discovery process, or you can retrieve this URL:
https://sta1.mios.com/locator_json.php?username=xxx where the username=xxx is optional, but, if specified, it will report all systems that mios.com user xxx can report whether local or remote, and without the username=xxx it only shows the Vera's on the user's local home network.
Each Vera, when it boots up, reports its internal IP address to the central mios.com server, which tracks this along with the external IP address. locator.php shows all serial numbers and internal network IP addresses on the same external IP. For example, if https://sta1.mios.com/locator_json.php?username=xxx.
[edit] The HTTP interface
Vera listens on port 3480. Vera responds to UPnP actions the normal way, or by making simple HTTP requests. For example, to turn on a light you can either send the UPnP action "SetTarget" to the device using the normal UPnP mechanism, or you can simply put this URL in your web browser, or with wget, or similar:
http://[ipadress]:3480/data_request?id=lu_action&output_format=xml&DeviceNum=10&serviceId=urn:upnp-org:serviceId:SwitchPower1&action=SetTarget&newTargetValue=1
That URL will return the same results of the action as using the normal UPnP architecture. The output_format=xml on the URL means the data returned should be in XML format. You can instead use output_format=json
The service ID's, actions, and arguments are all defined by the UPnP forum. A list of ratified UPnP devices and services is available on the UPnP Forum. Mi Casa Verde added its own custom devices and services when there was no existing UPnP standard. A list of these can be found in the Luup_UPNP_Files header files.
[edit] Getting the system configuration
Either use the UPnP service/action:
urn:micasaverde-com:serviceId:HomeAutomationGateway1/GetUserData
or use the built-in request:
http://ipaddress:3480/data_request?id=user_data2&output_format=json (or xml)
This returns the system configuration file for Vera. It's assumed you understand the way Vera categorizes devices with Rooms, Devices, Scenes, etc. When creating a user interface for controlling Vera only, like on a mobile phone, most of the tags can be ignored, since usually you just need the list of rooms and devices.
[edit] Accessing Vera remotely through the mios server
Native UPnP only works on a local area home network, and cannot be used over the Internet. The mios server provides a secure way to remotely access and control your Vera system using the HTTP interface without having to make any substantial changes to the UI. Everything you can do locally with Vera on port 3480, you can do remotely with mios using the exact same syntax. You only need to ask the user for his username and password and pass them on the URL to the remote access server, along with the serial number of the unit (ie like 12345). For example, the above user_data is the syntax for local access on a home network. To retrieve the same URL over the internet with mios use:
https://fwd2.mios.com/demovera/myvera123/12345/data_request?id=lu_status&output_format=json (or use fwd1.mios.com)
assuming demovera is the username and myvera123 is the password that the user chose setting up his mios.com account. Note that since the request is https, the username and password are encrypted because https encrypts the URL's as well as the contents of the page.
[edit] Building a control-only UI for Vera
If you just want to control Vera, and not modify the configuration, such as with a cell phone interface or web pad, the flow might possibly work like this:
1) For the initial setup, you need to ask whether the user wants to (a) access Vera remotely with the mios service, and if so, get the user's username/password; or (b) access Vera directly on the home network. You can use locator.php to show the Vera on the home network. You might want to do an auto-detect. For example, you could use locator.php to see if Vera exists locally, and if not, prompt the user for a mios username and password. You can use locator.php and pass the username to get the serial number and internal IP, and then automatically switch to direct access if locator.php (without the username) indicates that serial number is found on the local network. This would allow the user to automatically switch between local access with Wi-Fi (which is much faster of course) and remote access with mios and the mobile phone network. Or just ask for his mios username and show him all the Vera's he can control, either local or remote.
2) Once the connection is established, retrieve the list of rooms, scenes, and devices from Vera (see user_data below). You can cache this data locally so you don't need to retrieve it and parse it over and over.
3) Present the user with the rooms, scenes, devices. In Vera's UI's we show first the rooms and then let the user pick a device or scene in the room. However this data could be presented graphically, such as on a floorplan. Your UI can have Vera store extra tags and data in the configuration database (ie user_data) to store things like coordinates for icons, outlines or rooms and so on.
4) If the user picks a scene you can run the scene with the following URL, passing the scene number as x, and it will return an OK if it went okay:
5) If the user picks a device, you use the category which is contained in the user_data to determine what type of controls to present the user. (See Luup UPNP Files for a list of the categories.) When the user chooses an action, you can send it like this:
For a list of common actions and device types see: Luup CP Mandatory Types
Some actions return immediately with an 'OK'. Other jobs may take time to complete and are processed asynchronously. For these asynchronous jobs you will get a job ID back. There are several ways to present this to the user. One possibility is to do what Vera does in the UI, namely to use lu_status (explained next) to find all active jobs and to put a 'job' icon next to each device to indicate the progress of the job. In this way the user isn't tracking a particular job per se, but just has visual feedback that a device is busy, and if the job icon(s) turn green, that everything went through okay. You could instead track the action individually using the job number that is returned, or use some other method, like making the icon for the device appear animated while it's busy, and turn green or red if the last command succeeded or failed.
6) Periodically poll the system with lu_status (explained below), like this:
http://ipaddress:3480/data_request?id=lu_status&output_format=json
That returns the current values of the UPnP variables, which you can use to show the user current status. For example, if the variable Status is 1 for a switch, you can show the 'on' button in a selected state, while 0 shows it in an 'off' state. lu_status also shows any active jobs for a device.
[edit] user_data
The user_data is the main configuration file that Vera uses to store all the rooms, devices, scenes, timers, events, users, and any other data about the system. You should request the user_data to get whatever data you need to present to the user, at a minimum, for example, you'll need the list of devices, their descriptions, and the categories (light switch, thermostat, etc.). You'll also want the list of scenes to present the user. To fetch the current user_data from Vera use the user_data request, like this:
Returns JSON (preferred native format):
http://[ip address]:3480/data_request?id=user_data
Returns XML:
http://[ip address]:3480/data_request?id=user_data&output_format=xml
You can also add a 'DataVersion=TimeStamp' parameter to the URL. It the data has changed since the TimeStamp, all the data is returned. If the data has not changed since the TimeStamp, Vera will return a page with the text 'NO_CHANGES':
http://[ip address]:3480/data_request?id=user_data&DataVersion=315587159
The user_data contains the state for all the devices. For example this indicates that the device id #17, called "Switch" is turned off (ie service SwitchPower1, variable Status is 0)
{
"devices": [ { "id": 17, "category": 3, "room": 1, "name": "Switch", "ip": "", "mac": "", { "service": "urn:upnp-org:serviceId:SwitchPower1", "variable": "Status", "value": "0" } } ]
}
On UI7 build lu_sdata and lu_status have been deprecated, and you should use "status" instead. |
[edit] lu_status
The UPnP variables describe the state of the devices according to the UPnP specs. When the variables change that information needs to be shown to the user, eg change a Switch icon from off to on. If you are interfacing with Vera using the UPnP stack, you can subscribe to the device's events to receive notifications automatically when a device's state changes. However if you're using the web interface you will need to poll using status request.
Call lu_status request like so:
- http://[ip address]:3480/data_request?id=status
- http://[ip address]:3480/data_request?id=status&output_format=xml
- http://[ip address]:3480/data_request?id=status&DataVersion=315587159
The returned document contains four sets of information, which constantly change to reflect the current status:
- root attributes including DataVersion, which is the status timestamp, plus other attributes such as LoadTime
- the state of the system initialization, such as initialization of the plugin modules.
- a subset of the user_data that represents the current state of the device's UPnP variables
- the status of any active or recent jobs, such as if a light was successfully turned off
[edit] Tracking status changes
If you call lu_status without passing a DataVersion on the URL, you will get the current value of all the UPnP variables for a device. Note that most of the UPnP variables do not ever change. For example, with a light switch, only the variable "SwitchPower:Status" changes, as the switch is turned off and on. There are many other variables, like "Capabilities" which are simply configuration data or settings the user modified. If you pass &DataVersion=1 on the URL, you will only see the UPnP variables which have changed since the Luup engine started. So, if you parse the initial UPnP variables when you first call user_data, you can call lu_status with DataVersion=1. Once you call lu_status the first time, store the DataVersion and on subsequent calls pass it on the URL and you will only get the UPnP variables which have changed since that last DataVersion. Regardless, you will also get the current state of jobs and system initialization.
On subsequent calls to lu_status, you should also pass back in the LoadTime which you received on the prior call.
[edit] Controlling the poll timing
You can also pass in a Timeout value, which means the call will block for up to that many seconds if nothing has changed, and will return immediately when something does. This way you can have infrequent polling, say every 60 seconds by putting a Timeout value of 60, but your UI will be immediately responsive since, when something changes, it returns immediately. You can also add a MinimumDelay, which means that the call will wait for at least that many milliseconds. This is advisable so that in the event there are lots of continuous changes, such as a scene turning off many lights, you won't have lu_status calls returning non-stop consuming all the resources on both the Vera and the UI.
[edit] Startup status
The section with the initialization tasks is shown in the startup tag as shown:
"startup": { "tasks": [ { "id": 0, "status": 2, "type": "ZWave", "comments": "No Z-Wave" },
The startup tasks begin whenever the Luup engine is reset, such as after the user makes a configuration change. Ever plugin and module in the Luup engine does a startup sequence. The id is not really important. The status numbers are the same as for jobs, so 2 means the module failed to startup. The 'Type' is what you can use to key off of to know which module is being reported, such as "ZWave", and comments are just notes indicating the current state. For example, "No Z-Wave" means there's no Z-Wave dongle connected.
Once all the plugins have been initialized successfully the startup section will disappear completely from lu_status.
Note that lu_status also returns UserData_DataVersion, which is the dataversion from user_data. That way you don't need to poll user_data in parallel. Simply retrieve the user_data one time at startup, then poll lu_status, and, if UserData_DataVersion is different from the previous user_data you retrieved, then download user_data again.
Some notes on building a full replacement, including configuration, are available here UI_Notes_Replacement, however the syntax is not up to date with the Luup architecture.
[edit] Job status
The jobs for a device will look like this:
"devices": [ { "id": 1, "Jobs": [ { "id": "1", "name": "job#1 :getnodes (0x00D068F0) P:10 S:4", "icon": "", "comments": "", "status": "4" } ] },
The id is the job id number. When you run an action, like turning on a light, if it happens asynchronously, you'll get a job id number. All pending jobs for a device are shown by their id number. The name is just an internal name for the job. Icon is an optional keyword to indicate what sort of icon should be shown to the user. For example, 'getnodes' has no icon, meaning this type of job is generally not of interest to a user. If the icon was 'ON' or 'OFF' that would mean this job is busy turning on or off a light. Comments is notes about the job, if any. This may be shown to a user in a tool tip or similar, but is generally more technical than users would care to know. The status indicates the current state of the job, as shown:
- job_None=-1, // no icon
- job_WaitingToStart=0, // gray icon
- job_InProgress=1, // blue icon
- job_Error=2, // red icon
- job_Aborted=3, // red icon
- job_Done=4, // green icon
- job_WaitingForCallback=5 // blue icon - Special case used in certain derived classes
If the status is 0, this means the job is queued but hasn't started. If it's 1 or 5 the job is being executed right now. Generally both 1 and 5 are presented to the user the same way as a 'busy', although technically 1 means Vera is actively talking to the device, and 5 means Vera is waiting for a reply from the device. Status codes 2, 3 and 4 mean the job has finished. 4 means the job went ok, and both 2 and 3 mean it failed. Technically 3 means the job was aborted by another job or an external process, like a user cancelling it, while 2 means it failed due to a problem talking to the device. In both cases this is usually presented the user the same way, as a failure.
[edit] Categories and device types
The values used in user_data to determine the types of devices are in Luup_UPNP_Files. In general you can use the category instead of the devicetype since it's an integer and is thus faster to compare:
[edit] pseudo-code for a UI
Here is some pseudo code to show how the polling mechanism should be properly implemented. The parsing code is, of course, unique to your UI, so placeholder function names are shown here, like FetchAndParse. This shows the flow for handling the polling so you know you get the changes reported correctly.
variable int DataVersion_lustatus=0 variable int LoadTime=0 variable int DataVersion_userdata=0
- START
variable boolean FetchUserData=false; if( DataVersion_userdata==0 ) // If we're on the first loop and haven't retrieved user_data, do it and don't do an lu_status FetchUserData=true; else { while(true) { newdata = FetchAndParse("lu_status&DataVersion="+DataVersion_lustatus+"&LoadTime="+LoadTime); if( newdata.IsEmpty()==false && newdata.isError()==false ) break; // If we got data, continue on and process it else Sleep 1; // If we didn't get anything, or if lu_status failed, just sleep 1 second and stay in the loop to try again } // End of while loop
if( newdata.UserData_DataVersion!=DataVersion_userdata ) { FetchUserData=true; } else // Only Parse if userdata is the same, otherwise we're going to reload user_data { ParseLuStatus(); // This won't run if user_data has changed DataVersion_lustatus=newdata.DataVersion } } // end of else loop
if( FetchUserData==true )
{
UI.HourglassAndSuspend(true); // Put a 'busy' hourglass in the UI, display a message in the info panel "Updating data" and suspend all UI stuff since this may take a moment newdata = FetchAndParse("user_data2"); if( newdata.IsEmpty()==false && newdata.isError()==false ) { ParseUserData(); DataVersion_lustatus=newdata.DataVersion LoadTime=newdata.LoadTime DataVersion_userdata=newdata.DataVersion } else UI.DisplayError("Failed to update data"); // Display in the info panel an error message that the user will see for 10 seconds or so
UI.HourglassAndSuspend(false); // Put a 'busy' hourglass in the UI, display a message in the info panel "Updating data" and suspend all UI stuff since this may take a moment
}
Sleep 1; // We know we just parsed user_data or lu_status, so sleep before looping goto START; // Loop again and start over