Could we get a basic example on how to convert reqif to excel / csv please?

[vesper8]

Hello! First of all many thanks for making this project. I tried to implement a super basic reqif parser in Javascript and it worked for some very basic documents but now I'm running into issues with more complex documents so I've looked to your library which is clearly much much better than what I currently have. My current project requires that I be able to import requirements from reqif format and add them a simple relational database.

My first task in doing this would be to try and convert a reqif document to csv.

Your Readme.md states The first-stage parser (implemented by the class ReqIFParser) can be used by user's second-stage parser/converter scripts that convert the ReqIF in-memory structure into a desired format such as Excel

But I'm a little lost as to how to go about doing this. I was hoping there might be a simple example I could follow, or maybe just a code snippet that would show hot to extract requirement data from a reqif file and map it out to a simple array that I can then do as I wish with.

Any code snippet would be very appreciated to help me get started.

Many thanks!

[stanislaw]

Yeah, the documentation is very far from complete.

Could you try this and let me know if it is enough for you to get started? That code creates a single document with one requirement in it: https://github.com/strictdoc-project/reqif/blob/main/tests/integration/examples/01_create_reqif_objects/script.py.

[vesper8]

Thank you for linking to this example.. one of my client's planned requirements is the ability to convert requirements stored in a relational database back into the reqif format for the above example should prove to be quite useful for that! However I don't really how that can help me turn a .reqif file into csv or excel..

[stanislaw]

I could create an Excel example over this weekend but don't have it in hand right now.

[vesper8]

Actually.. one last item on my ultimate wishlist.. since I'll be calling your script through a server and then logging the output.. what would be really useful is if you could add one more optional flag for exporting directly to the CLI.. regardless of the export format.

Thank you!! You're awesome!!

[stanislaw]

Next iteration:

{
    "documents": [
        {
            "name": "Test Export",
            "content": [
                {
                    "Type": "System Requirement",
                    "ReqIF.ForeignCreatedBy": "redacted@mail.com",
                    "ReqIF.ForeignID": "LOREM-818",
                    "Status": "Draft",
                    "ReqIF.ForeignCreatedOn": "2023-03-15T10:46:58.611Z",
                    "ReqIF.ChapterName": "SW: Lorem Ipsum",
                    "ReqIF.Text": "\n                <xhtml:div>The Lorem Ipsum shall do something.</xhtml:div>\n              "
                }
            ]
        }
    ],
    "fields": [
        "ReqIF.ChapterName",
        "ReqIF.ForeignCreatedBy",
        "ReqIF.ForeignCreatedOn",
        "ReqIF.ForeignID",
        "ReqIF.Text",
        "Status"
    ]
}

The reason the fields seemed to be duplicated was because your reqif has two types: one for a heading and one for a requirement. Note that your example with just one spec object is rather basic. Could you provide one that has one or more sections/chapters with requirements nested in them?

Now that we are going to model most of ReqIF's complexity in JSON, and the CSV definitely doesn't scale here, we also need to model the fields for each type (heading or requirement). Also, in order to be able to recognize how nested your requirements or sections are, I need to introduce the level field, like 1.1, 1.2.3, etc. because otherwise you will not know who contains what. Either the list of objects is kept plain with the Level: field or the JSON has to nest the content with sections nesting requirements or other sections.

To summarize: it is not a very quick task if the converter has to deal with documents of arbitrary structure. I was going to implement all of this anyway, it is just that a good and robust implementation might take time.

I will get to all your comments, especially to those about the CLI design, at my next pass to this project somewhere next week. For now, the CLI is the same (still csv in it).

Also: Is this important for you to be able to convert JSON back to ReqIF? If yes, then it is especially import that the JSON representation preserves enough data and other element types information.

[vesper8]

Very excited for your upcoming iteration and thank you very much for taking all my comments into consideration!

I have asked for more complex reqif examples but am unsure if they will reliably come in anytime soon... in the meantime I have found these reqif samples that appear to be of varying levels of complexity. I found this in your own Github issues. Finding complex reqif samples on the internet is not an easy task..

https://ci.eclipse.org/rmf/job/rmf.develop.mars/lastSuccessfulBuild/artifact/test-data/

Although I don't know if any of them include chapters and links. In any case, for my current use case I would not need to import requirement linkage and being able to import a flattened array of requirements would suffice for now.

[vesper8]

Hello my friend! Just checking in to see if you've had time to work on the CLI improvements and other feedback I gave above or if you anticipate you'll have time soon? Thank you!

[stanislaw]

Hey @vesper8, not yet. The higher priority parent project's release triggered some hotfixes and the related discussion. I was planning to get to this JSON branch this evening.