yumeng5 / JoSH

[KDD 2020] Hierarchical Topic Mining via Joint Spherical Tree and Text Embedding
Apache License 2.0
57 stars 6 forks source link

tools folder unavailable #2

Closed nareshr8 closed 3 years ago

nareshr8 commented 4 years ago

Some files in tools folder are used like

I am unable to find the tools folder and its folders and sh files. Can u provide which the code or library you used, so that it would be helpful to replicate the results.

yumeng5 commented 4 years ago

Hi,

Thanks for trying out the code. I believe you are referring to the preprocessing tool AutoPhrase used in the framework. The AutoPhrase package is included under preprocess/AutoPhrase.zip which you need to unzip. You should be able to run the preprocessing code with the unzipped package.

Best, Yu

sharon-gao commented 3 years ago

Hi,

Thanks for trying out the code. I believe you are referring to the preprocessing tool AutoPhrase used in the framework. The AutoPhrase package is included under preprocess/AutoPhrase.zip which you need to unzip. You should be able to run the preprocessing code with the unzipped package.

Best, Yu

Hi Yu,

Thanks for your development of this tool!

I unziped the preprocess/AutoPhrase.zip and run auto_phrase.sh. However, there is an error FileNotFoundError: [Errno 2] No such file or directory: 'AutoPhrase/models/NEW/segmentation.txt'

It seems that NEW is created in the preprocessing. But I didn't see any file under this folder. Do you have any idea what happened?

Best, Shuang

yumeng5 commented 3 years ago

Hi @ShuangNYU,

Thanks for trying out the code. The segmentation results will be written into preprocess/AutoPhrase/models/NEW which should be already there after you unzip the preprocess/AutoPhrase.zip file. If you accidentally deleted the NEW directory (NEW directory does not exist), AutoPhrase will instead write the phrase segmented results to preprocess/AutoPhrase/default_models/NEW (I'm not sure why AutoPhrase does this, but this is how the AutoPhrase authors implemented it).

Maybe you could first check if you have a default_models directory with segmentation.model in it. If so, then you could manually create a NEW directory and re-run the code so that everything will be under NEW and you should get the correct AutoPhrase results. If not, it would help me to figure out the issue to post your full error trace here.

Thanks, Yu