ERROR in Palladian node TextClassifierModelReader

I am using KNIME Analytics Platform v3.3.3.v201707231738.

I created a workflow using Palladian TextClassifierModelLearner and TextClassifierModelWriter to write a file named 'palladian-model.palladianDictionaryModel'.

And then in another workflow, I am trying to use Palladian TextClassifierModelReader node pointing the the saved model file above.

Palladian TextClassifierModelReader is erroring out with the following line in the log:

ERROR TextClassifierModelReader 0:313      Execute failed: Not in GZIP format

Kindly help.

 

(answered via E-Mail)

Thank You!

For completeness, here's the solution:

[…] can you please try saving it with a filename of e.g. "palladian-model.palladianDictionaryModel.gz"? (use .gz file extension)

It might be, that GZIP compression is only applied when the file extension is explicitly given and the reader expects a compressed file.

Hi, I’ve tried your solution but the error is the same. Renamed the file to model-word-contenido.palladianDictionaryModel.gz I used the text classifier model writer.

This are my files, none of them load.

What’s the error message and the output in the console?

WARN Text Classifier Model Reader 0:91:132 Specify an input file to read.
ERROR Text Classifier Model Reader 0:91:132 Execute failed: Not in GZIP format
ERROR Text Classifier Model Reader 0:91:132 Execute failed: Not in GZIP format
ERROR Text Classifier Model Reader 0:91:132 Execute failed: Not in GZIP format

tried .gzip and .gz, none seems to work. Neither if I remove ‘.palladianDictionaryModel’ altogether.

Nah man, tried every combination, I’m out of ideas. Still ERROR Text Classifier Model Reader 0:91:132 Execute failed: Not in GZIP format

I think the writer only writes GZIP when specifying .gz as extension. Try if this helps. (I see that that’s what I wrote above anyways.)

–P

Hm, you are right, thanks. I thought you meant to rename the files. The Text Classifier Model Writer should mention this somewhere.