Now, whenever users are going to run the pipeline, even if they do not need a specific model in a specific layer, they need to install the libraries before running the pipeline, otherwise, they will get an error from the interpreter that does not recognize some libraries. For example, although users only want the LDA.mallet model in Topic Modeling Layer and they have not entered the gensim model as a choice in the running command, they need to install both gensim and LDA.mallet.
For handling this, we need to do two things:
As a comment, remind them that install the libraries of requirements.txt and environment.yml which they need to use, not all of the libraries which are there.
Relocate the importing codes to the specific methods which are called by parameters to prevent importing the libraries which are not necessary.
Now, whenever users are going to run the pipeline, even if they do not need a specific model in a specific layer, they need to install the libraries before running the pipeline, otherwise, they will get an error from the interpreter that does not recognize some libraries. For example, although users only want the LDA.mallet model in Topic Modeling Layer and they have not entered the gensim model as a choice in the running command, they need to install both gensim and LDA.mallet.
For handling this, we need to do two things: