Preprocess Failed - Missing Train.csv

[Abhrajyoti00]

While running the cmd : -
cm run script --tags=app,question-classification,topic-classification,go,question-topic,_rh
the collective scripts were searched successfully, but after the line "Entered Preprocess!" there was a FileNotFoundError: [Errno 2] No such file or directory: 'C:\Users\abhra\CM\repos\local\cache\e48c70baaf6e4170\train.csv'

[arjunsuresh]

Can you please do

cm rm cache - f
cm pull repo GATEOverflow@topic-classification

and retry?

[Abhrajyoti00]

Getting a CM error while executing : C:\WINDOWS\system32>cm rm cache - f

CM error: artifact(s) not found!

[arjunsuresh]

Can you do

cm pull repo mlcommons@ck 

and retry?

[Abhrajyoti00]

Both the pull repos have worked successfully, and there were new additions too.
But still the cm error: artifact(s) not found! is prevailing, due to which the script is still not wroking.

[arjunsuresh]

That's unexpected. I'm not able to reproduce this issue but I'm not on Windows. Let's sync later.

[Abhrajyoti00]

Dear @arjunsuresh Sir,
I am providing an update regarding the issue at hand. I have made several attempts to run the script on Windows, but have been unsuccessful in each attempt. The error message "cm error: artifact(s) missing" appeared each time I ran cm rm cache -f. I suspect that there may be a missing dependency or a step that was not followed during the installation of CM on Windows.

In light of this, I switched to Linux and was able to successfully run the command cm run script --tags=app,question-classification,topic-classification,go,question-topic,_rh without encountering any errors. However, I would like to bring to your attention that I encountered a few minor errors while running the script on my fresh Linux machine. I had to manually install the following dependencies using pip:

pandas module was missing, which was resolved by running pip install pandas
tqdm module was missing, which was resolved by running pip install tqdm
sklearn module was missing, which was resolved by running pip install scikit-learn
I recommend including these dependencies to be automatically installed when running the script in order to improve its usability.

[arjunsuresh]

Thank you @Abhrajyoti00 for the useful feedback.

For the windows issue, can you please share the output for
cm ls repo ?

For the dependencies, can you please add them in _cm.json like done here?

[Abhrajyoti00]

C:\WINDOWS\system32>cm ls repo
local,9a3280b14a4285c9 = C:\Users\abhra\CM\repos\local
internal,36b263b05174aef9 = C:\Users\abhra\anaconda3\lib\site-packages\cmind\repo
GATEOverflow@topic-classification,8dd99f68766d44b1 = C:\Users\abhra\CM\repos\GATEOverflow@topic-classification
mlcommons@ck,a4705959af8e447a = C:\Users\abhra\CM\repos\mlcommons@ck