Dears, I’m adapting the following tutorial in order to generate a text in Italian:
To do this I need to generate a new dictionary table with all the characters I need. I tried to do this by preparing a CSV spreadsheet, then converted to table using Table Writer. Unfortunately I am not able to read the question mark character which, when reading the file with Table Reader (I also tried file reader and CSV reader), is always interpreted as a missing value (red question mark).
Does anyone have an idea to be able to read the question mark?
thanks in advance
Welcome to the KNIME forum
A nice idea to adapt the workflow to generate text in Italian.
We recently updated this workflow for a university lecture in Italy and trained a network on Promessi Sposi. You can find the workflow and the trained network on the KNIME Hub: https://hub.knime.com/emilio_s/spaces/Exercises/latest/Firenze_Deep%20Learning/02_Free_Text_Generation/
Coming back to your issue that the question mark is interpreted as missing value character. I checked the workflows we recently prepared and I can not reproduce the error. Would you mind sharing your workflow with me, so I can have a quick look?
Hi Katrin, this is fantastic as I also thought about the Promessi Sposi! I have to do a lesson about AI. The new workflow, which derives dictionary from the file itself, is more robust and is working.
I uploaded my workflow, there are 2 attempts, one is using CSV reader, one with file reader:
thank you very much
can you please share the part where you write the dictionary. I think the problem occurred there.
Great to hear, that you are planning to use the example in you lesson about AI. May I ask at which university you are planning to run the lesson?
Sure, it’s a course for the Italian Association of Physicists in MEdicine + National Institute of Nuclear Physics.
here is the new file, it has the File Reader + Table Writer
KNIME_project4.knwf (6.3 KB)
many thanks, you’re very kind and patient!
(@Kathrin was so kind as to invite me to this thread: Thank you!)
In the context of our academic program we are providing teaching materials to the members: They cover a breadth of topics and are supposed to be (re)used in class. If you are interested, please send a quick note to firstname.lastname@example.org that you are interested in the teaching material. I don’t want to clutter up this thread more than necessary…
to solve the issue with interpreting the a questionnaire as missing value you can change the Missing Values Pattern setting in the File Reader node.
To do so you need to click on the “Advanced…” button in the configuration window of the File Reader node and go to the Missing Value Pattern tab. Here you can now define a new pattern, e.g. “MISSING”. This resolves the problem with interpreting the questionnaire as missing value.
Another missing value in your dictionary is in Row0. According to the csv file this should probably be a white space. To fix your dictionary you can use the Missing Value node and replace all missing values with a white space.
Fantastic, thank you!!! it works. And thank you Stelfrich, I’ll do as you suggested.
This topic was automatically closed 182 days after the last reply. New replies are no longer allowed.