Integrating Big Data Platforms with Bedrock: REST API

Posted by Adam Diaz on Feb 1, 2017 1:12:21 PM

In conversations about the Bedrock Data Lake Management Platform, the most common question we hear is “Can Bedrock integrate with product X?” which is almost immediately followed by “Can I use Bedrock without the UI?”. The answer to both is “Yes.” And it’s all made possible by the robust REST API that is built into Bedrock.

Screen Shot 2017-02-01 at 12.17.29 PM.png

This image displays an example of the interactive API documentation that is built on Swagger. Although classic documentation formats are available via PDF, you’ll want to check this out. The interactive documentation expedites the learning curve for making basic calls and crafting the correct input syntax for the API you are trying to use.

Understanding API Calls Through the Command Line

For example, you may want to understand variables in the Bedrock Namespace. Much like environmental variables in a Linux shell, Bedrock has a Namespace used to pass variables to different parts of the product. The command line below is a great example of what this will look like.

[bedrock@hostname]$ curl -D token.json -H "Content-Type: application/json" \
-X POST -d '{"username":"YYYYY","password":"XXXXX"}' \
 
[bedrock@hostname]$ JSESSIONID=$(cat /path/to/token/token.json | grep
JSESSIONID: | cut -d' ' -f 2)
 
[bedrock@hostname]$ curl -H "JSESSIONID: $JSESSIONID" -X POST --header
'Content-Type: application/json' --header 'Accept: application/json' -d '{}'
'http://localhost:8080/bedrock-app/services/rest/workflows/search?projectIds=9&
api_key=saveEntityInstance'
 
{
 "responseMessage": "success",
 "restUri": null,
 "result": {
   "currentPage": 1,
   "totalRecords": 2,
   "workFlowDetails": [
     {
 
          "CSVValues":
"59,Automated_Ingest_DQ_TM_noWM,unassigned,admin,11/18/2016
12:36:37,admin,11/22/2016 17:19:51",
       "category": "unassigned",
       "createdBy": "admin",
       "createdDate": "11/18/2016 12:36:37",
       "modifiedBy": "admin",
       "modifiedDate": "11/22/2016 17:19:51",
       "wfId": 59,
       "wfName": "Automated_Ingest_DQ_TM_noWM"
     },
     {
       "CSVValues": "62,Automated_Ingest_DQ_TM_TT,unassigned,admin,11/18/2016
16:22:33,admin,11/21/2016 21:46:07",
       "category": "unassigned",
       "createdBy": "admin",
       "createdDate": "11/18/2016 16:22:33",
       "modifiedBy": "admin",
       "modifiedDate": "11/21/2016 21:46:07",
       "wfId": 62,
       "wfName": "Automated_Ingest_DQ_TM_TT"
     }
   ]
 },

"page": null
}

For project ID 9 (aka Trouble Tickets in this system) there are two known workflows with IDs 59 and 62.

Practicing API Calls with the Swagger UI

Most users need some assistance when using a new API. It’s really helpful to have a way to practice iterating over the creation of a correct call before putting it into your process. In the interactive Swagger documentation, that same call would look something like the diagram below:

Screen Shot 2017-02-01 at 12.26.11 PM.pngUnder the workflow page we navigate to the search API and then fill in the correct values. In this case passing just a project ID is enough. This corresponds to the concept of projects in Bedrock. These can be listed out via the “Administration” menu under “projects” or via another API call to obtain all project IDs.

curl -H "JSESSIONID: $JSESSIONID" -X GET --header 'Accept: application/json'
 

RESPONSE BODY TRUNCATED:

{
 "responseMessage": "SUCCESS",
 "restUri": "/projects",
 "result": [
   {
     "projectId": 10,
     "projectName": "Taxi Data",
     "createdBy": "admin",
     "createdDateTime": 1479150403000,
     "canManageUserRoles": true,
     "isDefaultProject": false,
     "roleNames": [
       "Project Administrator"
     ]
   },
   {
     "projectId": 9,
     "projectName": "Trouble Tickets",
     "createdBy": "admin",
     "createdDateTime": 1478618016000, …

 

Screen Shot 2017-02-01 at 12.28.26 PM.png

The final screen will show a response body along with the response header and response code (not shown here). This is a quick and easy way to interrogate and learn the REST API for Bedrock. There are actions for controlling the entire system. In fact, this is how the web interface for Bedrock operates.

If you have applications you wish to integrate with Bedrock via REST API know that this is completely possible. This can be done by our staff here at Zaloni or by your staff directly. We also offer training on all aspects of Bedrock use including developer training. Please contact us for more information.

Topics: Big Data Ecosystem, Bedrock, Data Lake, Data Management