This project contains code samples that can be used as Data-Driven Content (DDC) within a SAS Visual Analytics (VA) report. For additional information, see Programming Considerations for Data-Driven Visualizations.
The JavaScript content for your third-party visualizations must be stored on a web server. One approach to hosting is to use Node.js. For more information about using Node.js for data-driven content, see Deploy a custom web application in the cloud for Data-Driven Content object in SAS Viya 4. Another possibility is to leverage SAS Content Server, as explained in Deploy DDC Implementation Files in SAS Content Server via SAS Viya GUIs.
In addition to accessing samples from the given folder, users may also like to refer to this new section - a DDC Server, which pre-packages these sample files in an easily accessible web application that can be deployed alongside SAS Viya! In addition, users will also be able to upload their own custom DDCs to their deployed instance of the DDC Server. Check this folder and README out in order to learn more.
It contains the functions you need to send/receive messages to/from SAS Visual Analytics. You must include the following line in the <head> of the web page:
<script type="text/javascript" src="../util/messagingUtil.js"></script>Sets a callback function to handle messages received from VA.
Usage:
va.messagingUtil.setOnDataReceivedCallback(callback)callbackis the callback function name that you must define.
Sends back to VA a message containing selections made in the third-party visualization. VA will use that information to either filter or select (brush) other report objects, depending on the Actions defined between the data-driven object and other VA report objects. It leverages function postMessage internally.
Usage:
va.messagingUtil.postSelectionMessage(resultName, selectedRows)resultNameis the name of the associated query result, which is obtained from the message received from VA (event.data.resultName).selectedRowsis an array of numbers (e.g.[0, 3, 4]) or objects (e.g.[{row: 0}, {row: 3}, {row: 4}]) that contains the indexes of the selected rows, as they appear in event.data.data.
Sends back to VA an instructional message. This message is displayed in the data-driven content object in the VA report and is useful for sending text messages back to report authors informing required roles, their assignment order, types, etc. It leverages function postMessage internally.
Usage:
va.messagingUtil.postInstructionalMessage(resultName, strMessage)resultNameis the name of the associated query result, which is obtained from the message received from VA (event.data.resultName).strMessageis the text message to be sent.
Sends back a message to VA.
Usage:
va.messagingUtil.postMessage(objMessage)objMessageis an object that contains either the selection or the instructional message to be sent to VA. FunctionspostSelectionMessageandpostInstructionalMessageare wrappers for this function.
Example of selection message:
{
resultName: "dd46",
selections: [0, 3, 4]
}
Example of instructional message:
{
resultName: "dd34",
message: "Please, assign proper roles"
}
Extracts parameter values assigned directly in the Data-Driven Content URL.
Usage:
value = va.messagingUtil.getUrlParams(name)nameis the optional parameter name.- Possible return values:
If
nameis an existing parameter, returns it's value Ifnameis not an existing parameter, returnsnullIfnameis not informed, returns and object with all parameters name-value pairs:{name1:value1, name2:value2, name3:value3, ...}
Forces VA objects that have filter actions driven from the DDC object to refresh: filters 1st row then none of them. For this function to work you need additional settings in the VA report, such as filter actions, table mappings, etc. This will be discussed in a blog soon and a link to it will be added here.
Usage:
va.messagingUtil.forceVAObjectsRefresh(resultName, millisecondsBeforeRefresh)resultNameis the name of the associated query result, which is obtained from the message received from VA (event.data.resultName).millisecondsBeforeRefreshis optional and indicates the time in milliseconds that it should wait between the two selection messages that this function sends to VA. You may need to tune this value according to your environment. You want to give enough time for the first selection message to be processed by VA, before the second message is sent, otherwise the filter action that the DDC has with other objects may not be triggered and the objects are not refreshed. This time is set to 2000 (2 seconds) by default and it should work for most of the cases.
It contains the functions you need to validate the data received from VA. You must include the following line in the <head> of the web page:
<script type="text/javascript" src="../util/contentUtil.js"></script>Sets a callback function to handle window resizing events.
Usage:
va.contentUtil.setupResizeListeners(callback)callbackis the callback function name that you must define. That's normally the function that re-draws the chart.
Checks if the data received from VA have all the columns (number, sequence, and type) required for the visualization.
Usage:
isValid = va.contentUtil.validateRoles(resultData, expectedTypes, optionalTypes)resultDatais the message received from VA (event.data).expectedTypesis an array describing the types of the columns that are required. The order is important and indicates the sequence that columns are assigned in the Roles tab in VA. Example: ["string", "number", "number"]. Valid types are "string", "number", and "date".optionalTypesis an array describing the types of the columns that are optional. The order is important and indicates the sequence that columns are assigned in the Roles tab in VA, after the required columns. Example: ["string", "number", "number"]. Because they are optional, the number of optional columns and optional types provided don't need to match. Type comparison will be made while there is still a column and a type to be compared, and the rest will be ignored. One or more optional columns of same type can be represented as a single type instead of an array, for example: "number" indicates that all optional columns, if existent, must be numeric columns. An empty array [] indicates their types can be anything. A value null indicates no optional columns are accepted. Valid types are "string", "number", and "date".- Returns true or false.
Uses the message received from VA to extract information about selections made in VA objects. After extracting selection information, the "brush" column is removed from the message.
Usage:
selections = va.contentUtil.initializeSelections(resultData)resultDatais the message received from VA (event.data).- Returns
selections, an array of objects containing the indexes of the selected rows (e.g.[{row: 2}, {row: 5}])
Transforms the message received from VA so that date values (represented as strings) are converted to Date objects. This standardizes date representation and might be helpful to support further transformations and formatting on dates.
Usage:
va.contentUtil.convertDateColumns(resultData)resultDatais the message received from VA (event.data).
Extracts parameter information from message received from VA.
Usage:
parameters = va.contentUtil.getVAParameters(resultData)resultDatais the message received from VA (event.data).- Returns
parameters, an object containing each parameter name and value (e.g.{<param_label_1>:<param_value_1>, ... , <param_label_n>:<param_value_n>}). If a certain<param_label>contains multiple values, its<param_value>is an array.
It contains helper functions you most likely need with Google Charts. You must include the following lines in the <head> of the web page:
<script type="text/javascript" src="https://www.gstatic.com/charts/loader.js"></script>
<script type="text/javascript" src="../thirdPartyHelpers/google.js"></script>Uses the data and columns keys from the VA message to create a DataTable object. Google Charts take DataTable as input data for its charts, but in addition to that, DataTable offers a series of methods that can help with table manipulation. More information on DataTable can be found here.
Usage:
dataTable = va.googleHelper.createDataTable(resultData)resultDatais the message received from VA (event.data).- Returns
dataTable, the input data as a DataTable object.
Uses the columns metadata within the message received from VA to update column formats in a DataTable object. Only numeric and date columns are affected. Supported formats are: DOLLAR, COMMA, F, BEST, and PERCENT for numeric, and MONYY, MMYY, MMDDYY, DATE, DDMMYY, WORDDATE, YYMMDD, and DATETIME for dates. Other column formats are kept unchanged.
Usage:
va.googleHelper.formatData(dataTable, resultData)dataTableis the input data as a DataTable object.resultDatais the message received from VA (event.data).
Uses the columns metadata within the message received from VA to update/add a vAxis.format or a hAxis.format in the Google Charts options variable. Only numeric columns are affected. Supported VA formats are: DOLLAR, COMMA, F, BEST, and PERCENT. Other column formats are kept unchanged.
Usage:
va.googleHelper.formatAxis(axis, options, resultData)axisis a string that indicates a vertical or horizontal axis. Valid axis values are'vAxis'and'hAxis'.optionsis the JSON object that holds Google Charts options.resultDatais the message received from VA (event.data).
It contains helper functions you most likely need with D3 charts. You must include the following line in the <head> of the web page:
<script type="text/javascript" src="../thirdPartyHelpers/d3.js"></script>Uses the columns metadata within the message received from VA to configure D3 formats. Only numeric columns are affected. Supported VA formats are: DOLLAR, COMMA, F, BEST, and PERCENT. Other columns formats are kept unchanged.
Usage:
formatter = va.d3Helper.configureFormatter(resultData)resultDatais the message received from VA (event.data).- Returns
formatter, an object containing key/value pairs where the key is the column label and the value is a d3.format function for each one of the supported numeric columns (e.g.{'Average Sales': d3.format('$6,.2f'), 'Percent Comparison': d3.format('4,.1%')})
Uses the columns metadata within the message received from VA to configure D3 formats. This function is similar to va.d3Helper.configureFormatter, but the format returned doesn't have any decimal places. Only numeric columns are affected. Supported VA formats are: DOLLAR, COMMA, F, BEST, and PERCENT. Other columns formats are kept unchanged.
Usage:
axisFormatter = va.d3Helper.configureAxisFormatter(resultData)resultDatais the message received from VA (event.data).- Returns
axisFormatter, an object containing key/value pairs where the key is the column label and the value is a d3.format function for each one of the supported numeric columns (e.g.{'Average Sales': d3.format('$6,f'), 'Percent Comparison': d3.format('4,%')})
It contains helper functions you most likely need with C3 charts. You must include the following line in the <head> of the web page:
<script type="text/javascript" src="../thirdPartyHelpers/c3.js"></script>Uses the data and columns keys from the VA message to create the chart data JSON object, necessary to draw the C3 chart. If necessary, this function also sets the appropriate configuration to unload the existing chart prior to re-drawing it.
Usage:
chartData = va.c3Helper.configureChartData(resultData, chartType, previousConfig)resultDatais the message received from VA (event.data).chartTypeis the chart type string as defined in C3 documentation.previousConfigis the previouschartDataJSON object that was assigned to thedatakey in functionc3.generate({... , data: , ...})to draw the C3 chart. Information frompreviousConfigis compared with the information fromresultDatato determine if the existing chart must be unload due to changes in the new data received from VA.nullindicates there is no previous configuration (chart was never drawn).- Returns
chartData, a JSON object to be assigned to thedatakey when calling the functionc3.generate({... , data: chartData, ...})to draw the C3 chart. It has the following structure:
{
type: <chartType parameter>,
x: <label of category column>,
rows:[
<array of column labels>,
<array of row1 values>,
... ,
<array of rowN values>
]
keys:{
x: <label of category column>,
value: <array of all numeric columns labels>
}
}
It contains the functions you need to create CAS sessions and execute CAS actions from SAS Visual Analytics. Those functions have been modified to also work in SAS Viya after 3.5.x (testing was performed on 2023.02 and 2023.03 release cadences), and should remain backwards compatible down to SAS Viya 3.4. Internal implementation and REST APIs used may differ depending on the SAS Viya version. Decisions are made based on the VA version detected. casUtil.v4.js is an implementation without inherited dependency on jQuery. It has been tested on stable release 2023.06 and Viya 3.5 and it should remain backwards compatible with all other SAS Viya releases. You should use this utility file going forward. Note: error handling for functions calls in .v4 implementation should use .catch() instead of .fail(). You must include the following line in the <head> of the web page:
<script type="text/javascript" src="../util/casUtil.js"></script>OR
<script type="text/javascript" src="../util/casUtil.v4.js"></script>Leverages SAS Viya REST API to create a CAS session that you can use to execute CAS actions. It uses the following endpoints internally:
VA 8.3 and above on SAS Viya 3.4:
/casManagement/servers and /casProxy/servers/<serverName>/cas/sessions
VA 8.5 on SAS Viya 3.5, and Viya 4 (2023.02 and later):
/casManagement/servers and /casManagement/servers/<serverName>/sessions
Usage:
va.casUtil.startCasSession().then(function(sessionInfo){...})- Alias:
startSession - Uses
getCasServerNameandgetAppVersioninternally - Returns a promise for
sessionInfo, an object containing CAS server name, CSRF token of the service used to create the session (casManagement or casProxy, depending on the VA version), session id, and session owner user id. For example:
{
casServerName: 'cas-shared-default',
csrfToken: 'f00cc954-040b-4407-b5ac-75a22df56ca3',
sessionId: '233c1c87-2016-1a41-8e99-461233aa306f',
userId: 'sasxyz'
}
Note: If you have more than one server, it returns the first one on the list.
Leverages SAS Viya REST API to execute CAS actions by using the following endpoint: /casProxy/servers/<serverName>/cas/sessions/<sessId>/actions/<action>
Usage:
va.casUtil.casAction(serverName, sessId, action, data).then(function(response){...})serverNameis the SAS Viya server name (e.g. 'cas-shared-default')sessIdis the CAS session ID (e.g. '233c1c87-2016-1a41-8e99-461233aa306f')actionis the CAS action (e.g. 'update', 'fetch', etc.)datais the CAS action dependent payload as an object or in stringified JSON format. E.g. for a fetch action:
{
"table": {
"name": "CARS",
"caslib": "Public",
"where": "Origin='Asia'"
},
"fetchVars": [
{"name": "Invoice"}
]
}- Returns a promise for
response, an object containing the data for the action called. See documentation for specific CAS action in developers.sas.com.
Leverages SAS Viya REST API to obtain the CAS server name. It uses the following endpoint internally:
/casManagement/servers
Usage:
va.casUtil.getCasServerName().then(function(serverInfo){...})- Returns a promise for
serverInfo, an object containing CAS server name and CSRF token of the service used to create the session. For example:
{
casServerName: 'cas-shared-default',
csrfToken: 'f00cc954-040b-4407-b5ac-75a22df56ca3'
}
Note: If you have more than one server, it returns the first one on the list.
Leverages SAS Viya REST API to obtain the CSRF token for a specified service. It uses the following endpoint internally:
/<service>/
Usage:
va.casUtil.getCsrfToken(service).then(function(token){...})serviceis the name of the service you want to obtain the CSRF token for, e.g. 'casProxy'.- Returns a promise for
token, a string containing the CSRF token of the service, e.g.'f00cc954-040b-4407-b5ac-75a22df56ca3'.
Note: Token values for each service are cached internally for future calls to this function.
Leverages SAS Viya REST API to obtain the version for a specified application. It uses the following endpoint internally:
/<app>/apiMeta
Usage:
va.casUtil.getAppVersion(app).then(function(version){...})appis the name of the application you want to obtain the version for, e.g.'SASVisualAnalytics'.- Returns a promise for
version, a string containing the application version, e.g.'8.5'. - It only works for SAS Viya up to version 3.5.x.
It contains utility functions to support easier integration between SAS Visual Analytics and SAS Jobs. jobUtil.v4.js is an implementation without inherited dependency on jQuery. It has been tested on stable release 2023.06 and Viya 3.5 and it should remain backwards compatible with all other SAS Viya releases. You should use this utility file going forward. You must include the following line in the <head> of the web page:
<script type="text/javascript" src="../util/jobUtil.js"></script>OR
<script type="text/javascript" src="../util/jobUtil.v4.js"></script>Transforms the data received from VA and adds extra format information for SAS Jobs to use.
Usage:
va.jobUtil.PrepareVADataForSASJobs(resultData)resultDatais the message received from VA (event.data) that will be transformed.
Notes:
-
Modifications performed on
resultData.datadepend on the data type:- String:
- missing values come as "(missing)" and are modified to "".
- Date:
- values come as formatted strings according to their date formats. They are transformed to their corresponding SAS date numbers expressed as number of days since January 1st, 1960.
- The following date formats are not supported (generate missing values):
- Day of Year (JULDAY1). E.g.: 176
- Week (WEEKV2). E.g.: 26
- Week (WEEKV3). E.g.: W26
- Year, Week (WEEKV5). E.g.: 15W26
- Year, Week, Day (WEEKV7). E.g.: 15W2604
- Year, Week, Day (WEEKV9). E.g.: 2015W2604
- Year, Week, Day (WEEKV0). E.g.: 2015-W26-04
- Some date string representations cannot accurately be mapped to one specific date. The formats below are treated by making up day, month, and year values as needed:
- Day of Month (DAY9). E.g.: 25
- Day of Week (DOWNAME11). E.g.: Thursday
- Day of Week (DOWNAME1). E.g.: Thu
- Day, Date (WEEKDATE9). E.g.: Thursday
- Day, Date (WEEKDATE3). E.g.: Thu
- MMMYYYY (MONYY7). E.g.: Jun2015
- MMYYYY (MMYY8). E.g.: 06/2015
- Month (MONTH7). E.g.: June
- Month (MONTH3). E.g.: Jun
- Month (MONTH2). E.g.: 6
- Month, Day, Year (WORDDATE9). E.g.: June
- Month, Day, Year (WORDDATE3). E.g.: Jun
- Quarter (QTR4). E.g.: Q2
- Quarter (QTR6). E.g.: 2nd quarter
- Quarter, Year (YYQC5). E.g.: 2nd quarter 2015
- Year (YEAR4). E.g.: 2015
- YYYYMM (YYMM8). E.g.: 2015/06
- Datetime:
- values come as formatted strings according to their datetime formats. They are transformed to their corresponding SAS datetime numbers expressed as number of seconds since January 1st, 1960.
- Some datetime string representations cannot accurately be mapped to one specific datetime. The formats below are treated by making up day, month, and year values as needed:
- Day, Date (DTWKDATX3). E.g.: Thu
- Day, Date (DTWKDATX9). E.g.: Thursday
- Quarter, Year (DTYYQC5). E.g.: 2nd quarter 2015
- Time (TIMEAMPM5). E.g.: 03:42 PM
- Time (TIMEAMPM8). E.g.: 03:42:12 PM
- Time (TIMEAMPM2). E.g.: 15
- Year (DTYEAR10). E.g.: 2015
- String:
-
Formats in VA for number/date/datetime are slightly different compared to SAS Jobs. Job-specific format information is added under
resultData.columnsand depends on the data type and VA format:
{
... ,
columns: [
{
... ,
name4job: <format_name>,
type4job: <format_type>,
width4job: <format_width>,
precision4job: <format_precision>
},
...
],
...
}
Asynchrnous function that pings the app to keep it alive.
Usage:
va.jobUtil.pingApp(app)appis the name of the application, e.g.'SASJobExecution'.
Asynchrnous function that calls pingApp on 1 minute intervals to keep the application alive.
Usage:
va.jobUtil.keepAppAlive(app)appis the name of the application, e.g.'SASJobExecution'.
