Writing .png files into a folder using a loop


I am new to working with flow variables and am running into a problem writing .png files using a flow variable.  Any help rendered by those more experienced would be greatly appreciated.

Here is a quick over-view of what I am trying to do.  Then I will detail the problem I am encountering:

I want to extract a tag cloud of keywords from each of a number of job applications.  Then I want to save each .png file created to the same folder.  So, my loop will take "job app 1.txt", extract a tag cloud of keywords and then will use the Image Port Writer node to save the .png file as "job app 1 .png" and so forth for each job application in my folder.

I use the list files node to load in the job applications I want to analyze, then feed them into a table row to variable loop start node, where I read and process each file, creating a keyword list and tag cloud.  I connect the tag cloud node to an Image Port Writer node.

To create a flow variable for the Image Port Writer Node, I have created a column of filenames the node will write to.  I then send this column of filenames to a Row Filter node to select only the filename row corresponding to the current iteration.  I then connect the Row Filter flow variable port to the Image Port Writer Node.  However, this does not seem to restrict the Image Port Writer node to writing to only a single .png file at a time.

The loop itself seems to be working correctly, however, instead of writing the tag cloud for "job app 1.txt" only to a new file "job app 1.png" it overwrites the tag cloud to each .png file that will be created.  So by the end of the loop (which currently has 10 iterations) the 10th tag cloud file has overwritten "job app 1.png", "job app 2.png", ...etc.

I have built this loop using information gleaned from similar forum postings, but can't seem to make it work.  Unfortunately I am not fluent in Java and can't implement those solutions.

I have included an annotated screen capture of my workflow and a zip file export of my workflow which will maybe help illustrate what I am trying to do. 

I am working in Mac OSX.

There may be something simple here I am missing.  Any illumination would be appreciated.



I've appended asimple examle that does this.  Some small tips are that you can get the base filename and parent folder for free using the URL to File Path node.  Have a look and let us know if this was what you are looking for.  The example is a bit more generic so that others can use it without the text mining extension.

Best regards,