How to get total error count from Klocwork automatically - klocwork

Question

I have a project running in Klocwork and after the build gets completed the Klocwork results will be generated. Every time I need to go to the Klocwork portal to get the results and look for the new issues or the total issues. Instead I need an API or script to get the total number of issues from the Klocwork results automatically when the build is successful.
Is there any way to achieve this? One way is to view the portal page source as html and get the result I need. However, I think there might be a better solution.
Can someone help me in achieving this?
Thanks in advance.

Answer 1

I answered a similar question over here. Below is an updated answer with links to the documentation for the most recent release, Klocwork 11.
Klocwork has a WebAPI which you can use to query this type of information from your favorite scripting language, or for example with curl. API documentation is also provided on your Klocwork server at http://klocwork_server_host:port/review/api, for example http://localhost:8080/review/api.
The query:
curl --data "action=search&user=my_account&project=my_project&query=build:build_1 status:Analyze state:New,Existing&ltoken=xxxx" http://localhost:8080/review/api
will return a list of all open (state New and Existing), non-cited (status Analyze) issues found in a build named build_1 of project my_project.
For a list of the keywords you can use in the query string with the search action, see Searching in Klocwork Review.
If you want just a summary of the number of defects instead of getting the whole list, you can use the report action:
curl --data "action=report&user=my_account&project=my_project&build:build_1&x=Category&y=Component&filterQuery=status:Analyze state:New,Existing&ltoken=xxxx" http://localhost:8080/review/api
which returns back a summary of the number of defects by checker category (taxonomy) and component. Sample output is below:
{"rows":[{"id":1,"name":"C and C++"},{"id":3,"name":"MISRA C"},{"id":4,"name":"MISRA C++"}],"columns":[{"id":5,"name":"System Model"}],"data":[[122],[354],[890]],"warnings":[]}
You can modify the x and y axis parameters to produce different breakdowns of the issues, for example by Severity and State instead:
curl --data "action=report&user=my_account&project=my_project&build:build_3&x=Severity&y=State&filterQuery=state:New,Existing,Fixed&ltoken=xxxx" http://localhost:8080/review/api
output:
{"rows":[{"id":1,"name":"Critical"},{"id":2,"name":"Error"},{"id":3,"name":"Warning"},{"id":4,"name":"Review"}],"columns":[{"id":-1,"name":"Existing"},{"id":-1,"name":"Fixed"},{"id":-1,"name":"New"}],"data":[[10,5,2],[20,6,1],[45,11,3],[1112,78,23]],"warnings":[]}
The WebAPI cookbook documentation has an example of using python with the report action and processing and formatting the response.