nav[aria-label="Primary Navigation"] { padding: 0; & ul { list-style: none; width: 100%; display: flex; flex-direction: row; justify-content: start; align-items: start; gap: 30px; padding: 0; & li { margin: 0; } & ul li { list-style: none; } } }

Inquiry about Generating Report from Diff&Merge

medini_1044914794

Hi, this is Daehui Jeong from HancomMDS which is located in Korea.

This time the customer that 'OMRON Korea' requested us about the technical issue.

They have some inside problem at configuration process and tool.

So, they don't use any configuration tool(ex. SVN) which medini provide integrating.

For this reason the customer wanna take advantage of diff&merge function in medini.
They wanna generating change history management report from diff&merge data using Birt Script or any other methods.

Could you give me any idea or script for it?

Find more posts tagged with

medini

medini legacy

Accepted answers

All comments

medini_1024169437

Hello Daehui Jeong,

interesting topic, through, there is currently little to no support to access a diff merge tree from a script or a report, unfortunately.

Its also currently not possible to trigger a compare operation by a script.

We have identified that as a missing feature already and at least want to support the execution of scripts in the diff / merge

perspective on a compare result in the next major version.

For current version I have no real solution yet here in scripting, I'm sorry.

Regards

Jan

medini_1044914794

Hi Jan,

Thank you for replying.

I hope that it will be implemented to medini analyze in the near future.

Thanks for again,

Best Regards,

- Daehui Jeong

medini_1024169437

Hello Daehui Jeong,

I forgot to mention that we have some initial support to script against the diff tree now with 2020 R1, of course not officially in any API doc but technically a script may extract arbitrary diff reports.

Jan

medini_1044914794

Hello Jan,

Thanks for notifying about the updating.

Would you explain a bit more to me how to generate diff reports in 2020?

Thanks for again,

Best regards,

- Daehui jeong

medini_1024169437

Well, we mainly have now the possibilities, we have no sample or example yet.

I will try to prepare a simple one which you can use then to write your own reports.

Jan

medini_1044914794

Hello Jan,

Thanks for replying.

I hope that it will be a good material for introducing medini to the customer.

Best regards,

- Daehui jeong

medini_1073221077

Jan,

I want to bump this topic again. I have tried to write an export script using the Compare view, but with limited success. The scripts that I run often fail silently, which is a different behaviour compare with the Analysis view. Do you have any code examples that you could share? I run 2020 R1.

Regards,

Mattias

medini_1024169437

Well, seems I broke my promise to post a sample here, what a shame.

Matthias, do you have your starting point and can post it or is that not feasible?

Please post if you can and I will enhance it.

The diff model is rather tricky, I agree.

Jan

medini_1073221077

I'll see if I can Clean up the code and post it.

Basically what I'm trying to do is create a script to extract the differences between two revisions of a DC Worksheet to a .csv file using the Compare view instead of resorting to comparing two Excel exports. When we perform Reviews internally at our Company we want to compare a particular revision to a baseline that has been reviewed previously. In this particular case we want to compare the connected effect and a text rationale (in a custom column) for each failure mode in the DC Worksheet.

Thus far I have been able to loop through the vectors of the "submatches" vector and I have been able to extract some information, but I find if difficult to parse the actual Changes in the "differences" vectors. For the rationale text column it will result in a AnalyzeAttributeChange object, but I'm not sure what other type of objects I might need to parse. But the biggest problem for me is that I don't get any error messages if the script fails, it just gives no response. Why is that?

Any pointers would be welcome! :-)

Regards,

Mattias

medini_1024169437

Hi Mathias,

that would be really cool so we have a starting point that we can enhance.

What exactly do you mean with "fail". Do you think, that an exception iss thrown and simply not visible at the surface

or does your script access non-existing links and attributes and fails so you don't recognize?

The first would be strange because the execution mechanism is the same as for normal execution in model browser.

For the later, please check this article - it explains how to make the JS interpreter more strict.

Happy Scripting

Jan

medini_1073221077

Hi!

Here's my current script. I switched approach from several nested for-loops to a recursive solution. Eventually the goal is to export to .csv or perhaps XML, but right now it just writes to the console.

Note: We have a custom column called "user_failure_effect_rationale" on the DC Worksheet. The information that I want to export is changed rationales and changed effects. The current problem I'm having is that I can print the Component name along with the failure mode when I run the script on the DC Worksheet, but for the cause-effect Changes I need to find a way to print the Component name.

Any suggestions? :-)

Regards,

Mattias

Example:

medini_1073221077

I found out a way to get the information about the Component containing the failure modes to the lower levels in the tree. For each recursive call I send a string containing the label of the parent, in this case the Component.

Now the question is how to present the information in a human readable form..

Regards,

Mattias

-----------------------

function parseDiff(diff, parent){

if(diff.differences != undefined)

for(var j = 0; j < diff.differences.size(); j++){

if(diff.differences[j].attribute != undefined)

analyzeAttribute(diff, j, parent);

if(diff.differences[j].reference != undefined)

analyzeReference(diff, j, parent);

}

// Perform recursive call to dig deeper in the tree

if(diff.submatches != undefined){

for(var i = 0; i < diff.submatches.size(); i++){

if(diff.left.label != undefined)

parseDiff(diff.submatches[i], diff.left.label);

else if(diff.left.element != undefined)

parseDiff(diff.submatches[i], diff.left.element.label);

else

parseDiff(diff.submatches[i], "EMPTY");

}

return;

}

medini_1024169437

I made some tests and I can confirm, that indeed any error in the script seems to be silently ignored,

which is very off indeed. Until I have figured out things, my advice would be to wrap your code

in a simple try/catch block, for example:

try {
 callSomeCode();
} catch (e) {
 alert("Failed with" + e);
}

I will give your sample a try next.

Jan

medini_1024169437

Mattias, about one of your questions beforehand ... about the component name.

My idea was that you only extract information on differences form the diff tree but things like name

of a component or else you extract from the real model., so you navigate from the diff model to either

the source or the target project and get the name from there as if you would write a normal script in medini.

Searching the diff tree for the "name difference" seems overkill or at least difficult, isn't it?

Jan

medini_1073221077

Jan,

Using try & Catch was a good suggestion, thanks, it makes it easier! And enabling the debugger is also very helpful.

Yes, I'm restructuring the script a bit to make it simpler. I agree that it is a bit overkill to look into each "difference" instead of just looking into "left" and "right". I will post something when it is a bit more generic.

Regards,

Mattias

medini_1024169437

Sounds very good, looking forward to it.

Jan