Requirements Gateway - Searching for requirements in table rows in a Adobe PDF file

[MarcusTörndahl]

In my quest for creating a good understanding of the coverage of requirements of my test cases I download the NIRG (NI Requirements Gateway) and started out importing my SRS (System Requirement Specifications) documents and some of the test cases. There are many different types of sources which can be imported (Word, Excel, Visio, Access, PDF, ...) and NIRG uses REGEX (Regular Expressions) for extracting the requirement and coverage information of the sources. NI have defined some standardized REGEX formats but they (NI) have also made it possible for tweaking these to fit your purpose. 
Instead of adding/changing the SRS documents to fit these standardized REGEX formats, I dived into how to create REGEX for my purpose. 
Check out 
- http://www.regular-expressions.info/ 
- http://opencompany.org/download/regex-cheatsheet.pdf 

NI Requirements Gateway Coupling Series 
(http://zone.ni.com/devzone/cda/tut/p/id/8414)

NI Requirements Gateway for Test, Measurement, and Control Applications 
(http://zone.ni.com/devzone/cda/tut/p/id/3721)

Is there anyone who have any experience on using NIRG on a PDF document to get information from the rows in a table?

Best Regards,

Marcus

[Manooch_H]

Marcus -

If you'd like to provide the test document that you gave a screenshot of and explain which columns you'd like imported into NIRG, I can try to assist you with this.

[MarcusTörndahl]

Hi Manooch_H,

Here is the example document and thank you for taking the time 🙂 and helping me out :).

The columns which I would like to import into NIRG are actually all of them. The UC, Verification, Priority and Responsible columns should be imported as attributes to the requirement. The ID and Requirement columns should be imported as the Requirement and Text respectivaly.

Best regards,

Marcus

[Manooch_H]

Marcus -

Thank you for providing the test PDF file. Unfortunately, when Requirements Gateway 1.1 processes the PDF file, the intermediate text file that it creates only recreates the format of the table and does not actually provide any type of identifiers for each cell in the table (see the screenshot below).

The lack of an identifier for each cell (for example, an identifier that informs us which column the cell resides under), prevents us from being able to create a robust regular expression to handle such PDF documents.

One thing that might be possible is if within each cell an identifier was provided to signify what type of information the cell contains, we could then look for that specific identifier text and know whether or not the text that followed was a requirement or text or attribute.

[MarcusTörndahl]

Hi Manooch H,

The only specific identifier now is that the requirements ID consist of [A,B][B,H,C]+number.

Just an idea, if the PDF document could be interpreted to XML with tagged information regarding for instance table information then it would be possible to extract the entire table as is.

Thank you for your help.

Marcus