Community
Participate
Working Groups
Provide a set of APIs for Data Visualization to query the Resource Model Repository. These are the services Data Visualization (DV) would like: 1. data filtering - ability to show certain data that meet a specific criteria 2. data grouping - organize the data based on a grouping criteria (e.g. create a report that shows available application grouped by host) 3. data sorting - provide apis to sort the data 4. data counts - provide apis to count a set of data that meets a certain criteria Sheldon (sleeloy@ca.ibm.com) is the main point of contact for details on these requirements.
I am a bit confused by the request description; it seems that the expectation is to provide a set of API for accessing the data .This request should be addressed to the data collection team. Can you please double check what you need from the RM team. Last time I talked to Sheldon about this request we were referring to SML definitions for the actual data to be used by the generated report. I have a different request opened by the data collection team to provide an SML model for the collection management and managed entities. https://bugs.eclipse.org/bugs/show_bug.cgi?id=178590 I attached to 178590 a zip file containing a first approach on modeling agents; there is a logging agent and statistical agent sample attached to that file. Can you please take a look at these and let me know if it covers at least partially what you need.
I had an offline discussion with Sheldon and had clarified the request. The data visualization project, as well as any other project processing SML and SMLIF content needs a set of API to navigate the in-memory model once parsed. It also requires a set of API’s to navigate the SML repository and retrieve data based on specified queries. 1. An API example in the context of an SML instance : If a DOM element is an sml:uri, which means is a reference to another element in a different document, the web UI needs a way to get back the information in a way easy to read. In the xml snippet below, the web ui would need to parse this structure and retrieve the agent state. The state is defined by an sml:uri reference and the referenced element is located in a different file. The web UI will need an API at the DOM level that reads the uri and gives back the status element. <statePropertyRef sml:ref="true"> <sml:uri> /DataCenterSample/dataCollection/TPTPDataController.xml#xpointer(/controller[@name="TPTP Agent Controller"]/agents/agentRef[@name="Apache error 1"]/stateProperty) </sml:uri> </statePropertyRef> 2. An example of an SML repository API repository.getReferences(filename) // returns all files referencing the specified file Sheldon, let me know if I entirely captured our discussion. Also, I will need from you an initial set of API’s you require for building the UI ( words description of the API works fine )
<<The data visualization project, as well as any other project processing SML and SMLIF content needs a set of API to navigate the in-memory model once parsed. It also requires a set of API’s to navigate the SML repository and retrieve data based on specified queries. >> How distinct are these APIs? Should this be decomposed into two seperate reqursts, one for the in-memory model and another for the repository API?
<<repository.getReferences(filename) // returns all files referencing the specified file>> Instead of using a "filename" as an argument, can we instead use a URI or an EPR? I think file://foo/bar/myFile.xml is valid, so this should not preclude using a file based repository. Using a URIs and/or EPRs has the added advantage of allowing us to have this API be a set of REST services or web services. It would also hide the location & implementation of the repository.
See https://bugs.eclipse.org/bugs/show_bug.cgi?id=179828 for an attached design document.
Implemented as part of https://bugs.eclipse.org/bugs/show_bug.cgi?id=179828. Please re-open if other functions are needed.
closing
