Hap score - Example Notebook

[AishaDarga]

Why are these changes needed?

Related issue number (if any).

[shahrokhDaijavad]

@ian-cho This is another variation of a Notebook that uses your hap transform. Can you please review and provide comments?

[ian-cho]

@shahrokhDaijavad As long as this modification does not affect other modules in the hap/python directory or make them incompatible with each other, it is fine with me.

[shahrokhDaijavad]

Thanks, @ian-cho. So, just to confirm, you are ok with the changes made in hap_config.py and hap_local_python.py files, right? Any future changes to the hap transform directory need to be reviewed again.

[ian-cho]

@shahrokhDaijavad Thanks for asking. If the changes affect other modules, would you mind helping resolve them when I am not available? Previously, @touma-I and you helped a lot in ensuring module compatible with each other I remember, and it was not easy for me to ensure that.

[AishaDarga]

@shahrokhDaijavad @ian-cho I've updated the notebook to ensure compatibility with Google Colab and included the installation of all necessary packages directly within the notebook. Could you please review it?
Thank you!

[shahrokhDaijavad]

@AishaDarga Have you tested this notebook? I ran into errors, running both on the local machine and on the Google Colab.

On the local machine, I get errors about hap_params not being defined.

and

On Google Colab, I get the following error:

[AishaDarga]

@shahrokhDaijavad Yes, I've tested the notebook thoroughly on my system. I'll try to re-install and test it afresh. Thank you!

[shahrokhDaijavad]

@AishaDarga I just ran your latest notebook successfully, so it works fine. However, I understand you had some discussions with @touma-I about not changing the transform code. When you make the necessary changes, I will be happy to test again.

[AishaDarga]

@touma-I @shahrokhDaijavad Thank you for your feedback. I've made all necessary changes in the latest commit. This version of the code leverages the HAP transform as is, without making any changes. Could you please review it

[shahrokhDaijavad]

This version, which I have tested successfully, looks good to me and the directories are also clean. Thanks, @AishaDarga.

[touma-I]

Please consider removing the parquet file from git since we can regenerate it from css. Thanks

[touma-I]

remove data-prep-connector since you are NOT using it. This line can be deleted:

! pip install data-prep-connector

[touma-I]

You do not need this line. Those are transitive dependencies of the transform itself and could break the notebook in the future if they are changed by new releases of the ransform.

! pip install nltk==3.9.1 transformers==4.38.2 torch>=2.2.2,<=2.4.1 pandas==2.2.2

[shahrokhDaijavad]

Adding these is my fault, @touma-I! Since I could get the notebook to work only after pip installing requirements.txt, I thought they were needed, but it turns out my problem was something else! In any case, what you are saying is that these dependencies are taken care of by the pip install of the transform itself and should not be done here.

[AishaDarga]

@touma-I @shahrokhDaijavad I've made all necessary changes and addressed the comments given above. I've tested the latest changes and it works fine for me. Could please review it?

[shahrokhDaijavad]

Nice job, @AishaDarga.
@touma-I I have tested the latest version successfully with all the changes we had asked. I recommend merging this.