Problem with scipy during install: Preparing wheel metadata ... error, ERROR: Command errored out with exit status 1: about keybert

[MaartenGr]

From a quick look, I could not tell you what the issue here is. Scipy is installed through one of the following packages:

• "sentence-transformers>=0.3.8"
• "scikit-learn>=0.22.2"
• "numpy>=1.18.5"

Which are essentially all base dependencies found in setup.py. I would suggest starting from a fresh (ana)conda environment seeing as I see many "Requirement already satisfied" messages. I am curious to see if it still gives an error if there are no other packages already installed.

[AbhiPawar5]

Hi @MaartenGr , I am facing one of the similar issues. My environment is:
keybert==0.2.0
numpy==1.19.5
pandas==1.1.5
scikit-learn==0.22.2

and when I try importing keyBERT it says,
ImportError: dlopen: cannot load any more object with static TLS

EDIT : Can you please provide a requirements.txt kind of list to help us install this correctly? Thanks.

[MaartenGr]

@AbhiPawar5 Based on the packages you have installed I would advise you to start from a fresh environment and install the latest version of KeyBERT. Seems you are using an earlier version which has some issues with numpy/scipy/pypi.

EDIT: The requirements can be found in setup.py

[AbhiPawar5]

Hi @MaartenGr , thanks for your input. I did try it my fresh environment which solved earlier issue but encountered a new one:
certifi==2020.12.5
chardet==4.0.0
click==8.0.1
dataclasses==0.8
decorator==4.4.2
filelock==3.0.12
huggingface-hub==0.0.8
idna==2.10
importlib-metadata==4.0.1
jellyfish==0.8.2
joblib==1.0.1
keybert==0.3.0
networkx==2.5.1
nltk==3.6.2
numpy==1.19.5
packaging==20.9
pandas==1.1.5
Pillow==8.2.0
pyparsing==2.4.7
python-dateutil==2.8.1
pytz==2021.1
regex==2021.4.4
requests==2.25.1
sacremoses==0.0.45
scikit-learn==0.23.2
scipy==1.5.4
segtok==1.5.10
sentence-transformers==1.2.0
sentencepiece==0.1.95
six==1.16.0
tabulate==0.8.9
threadpoolctl==2.1.0
tokenizers==0.10.3
torch==1.8.1
torchvision==0.9.1
tqdm==4.61.0
transformers==4.6.1
typing-extensions==3.10.0.0
urllib3==1.26.4
yake==0.4.8
zipp==3.4.1

But still seems to have some other issue:

$ python --version
Python 3.6.8

$ python
Python 3.6.8 (default, Dec 25 2018, 00:00:00) 
[GCC 4.8.4] on linux
Type "help", "copyright", "credits" or "license" for more information.
>>> from keybert import KeyBERT
>>> distilbert_model = KeyBERT('distilbert-base-nli-mean-tokens')
Segmentation fault (core dumped)

[MaartenGr]

Strange, that seems to be a C++ error. From what I could gather, perhaps a wrong install of Python could be the culprit here (see this). I did several tests and could not reproduce this issue so I am afraid I cannot be of much help.

The only thing that I can think of is that you have too little memory available to actually download and prepare the model.

[MaartenGr]

Since this issue has been a while without activity, I'll be closing it for now. However, if you are still experiencing the issue or want to discuss it further, let me know!