Can't import NLTK into Jupyter Notebook - python

I'm pretty new to python, Jupyter notebook, Tensorflow, and that whole lot in general. I'm getting started with a machine learning project. I've gotten to the point where I want to import "nltk" into my thing. It doesn't work. I've installed nltk with pip, and conda, and everything, in my terminal. When I do it again in the notebook, it says I've already installed it, which is correct. But when I try to import it it gives me a ModuleNotFoundError:
I'm on a macbook, by the way. Any help?

Going forward you can follow these steps (through terminal) so that same issues don't crop up again.
Create a conda environment if it's not already done
conda create -n py3_env python=3.8
Get into the conda environment
conda activate py3_env
Install ipykernel
conda install ipykernel
Link the kernel to this conda env
ipython kernel install --user --name=py3_env
Deactivate the conda environment
conda deactivate
Now, when you open jupyter, you can select this kernel from the dropdown menu kernel >> Change kernel. Now, all the packages you've installed in this conda environment would be available in jupyter as well. E.g. you can install nltk in this environment in the following way:
conda activate py3_env
pip install nltk
conda deactivate

Related

Installed python package to conda virtual environment, ModuleNotFoundError in Jupyter Notebook

Installed python package from the Anaconda Prompt in the virtual environment.
import transformers works from the Anaconda Prompt. However, ModuleNotFoundError when in Jupyter Notebook even though kernel is set to the virtual environment where the package is installed.
It looks like the issue is similar to
Jupyter notebook can't load installed package in conda environment, but I can't get it to work on Windows.
Here is an example.
Package is installed in the virtual environment.
Importing from the anaconda prompt works.
Importing from Jupyter Notebook causes the error.
Open the Jupyter notebook from command line. Activate your conda environment first.
Open command line. Activate your conda environment.
Type activate <your env name>
Type jupyter notebook <path to your project>. To open in current directory jupyter notebook . .
The package was initially installed using conda install -c conda-forge transformers.
Instead, I installed the packaged using pip install transformers. Importing the package from Jupyter Notebook on the virtual environment's kernel worked fine afterwards.
When I did where conda from the virtual environment, conda install doesn't seem to be installing to the venv. Whereas where pip showed that the venv's pip is being used.

Conda install packages from jupyter notebook

I have used !{sys.executable} -m pip install to install packages i want to use in my jupyter notebooks.
Now i also want to install some packaged by conda and use in the same notebooks.
This is the command i run and the error message
They ask us to create a separate conda environment. I am confused about this.
Do we need to create a new conda enviroment everytime we need new packages?
What is the correct way to install and access conda packages on jupyter python notebook?
You shouldn't in principle install from the jupyter notebook but from the terminal/cmd.
You can create an enviroment >conda create --name myenv
And activate it: >conda activate myenv
now everything you install will be restricted to myenv. for example conda install numpy
To use the packages in myenv in jupyter simply open jupyter with your enviroment active:
>conda activate myenv
>jupyter notebook
Note that you have to install jupyter in your enviroment too: conda install jupyter

Accessing package in python virtual environment

I have a virtual environment in which I have installed 'geopandas'. When I do a pip list it lists this package as well. But when I try to call it in my jupyter notebook via 'import geopandas' I get a ModuleNotFoundError. Please help - I'm using windows machine
You can verify that your notebook is running in the correct virtual environment by doing:
import sys
sys.version
Here's how to run a Jupyter notebook in a virtualenv.
You should check if your notebook is using the correct kernel (the correct virtualenv). If you are still in the kernel using your standard environment and geopandas is not installed, it is possible you get this error.
So check if you are working in the correct kernel:
kernel check
You can install a kernel in jupyter notebook by activating the venv and then installing it:
source activate myenv
python -m ipykernel install --user --name myenv --display-name "Python (myenv)"
More info about this install here.

Trying to install a pip package in Anaconda

I'm trying to follow this tutorial:
https://learn.microsoft.com/en-us/azure/machine-learning/service/tutorial-data-prep
As part of this I'm trying to do a pip install of azureml as it's not available on conda. However doing a pip install will by default install it to my default python install, and not my conda install.
So I tried following the steps here:
https://conda.io/docs/user-guide/tasks/manage-environments.html#using-pip-in-an-environment
However after following these steps I then launch Jupyter notebook after activating myenv, navigate to the notebook, and try and run:
import azureml.dataprep as dprep
But get the error: ModuleNotFoundError: No module named 'azureml'
Also - I cannot tell if myenv is active in the notebook. The kernel simply says python3.
Be careful, when using pip in anaconda, it is possible that you are mixing pip and pip3.
Run which pip3 to be sure you are using the version that correspond to the virtual environment.
If you are using python3 in the environment, then pip will typically be the correct version to use. Do not use pip3 in that case.
This problem has been documented elsewhere on the web. The problem is that Jupyter notebooks itself only launches in the root environment by default. The simplest solution to getting it to launch for your env (e.g. myenv) is to install Jupyter within your env first. So from the Anaconda command prompt:
activate myenv
pip install jupyter
jupyter
Ps. Use source activate myenv for non-windows machines

Keras installation

I create an virtual environment in my conda named 'keras_ev' and install the keras in it
by
conda install keras
after that when i
activate keras_ev
jupyter notebook
the notebook does not show my keras_ev environment
and i fail to import the keras in my notebook.
Does anybody know how to fix this! Thank you
Try conda install ipykernel in your keras_ev environment. Then it should appear in your Jupyter notebook.
You can also install Python dependencies while using your Jupyter notebook. First, activate the environment keras_ev in another terminal tab. Then install your dependency using conda or pip (conda is recommended). It should be something like the text below.
In a new terminal:
source activate keras_ev
conda install *your_package*
In order for Keras to work you should first install one of its backends. If you don't install any of the backends it won't work, although it seems that it is installed. please refer to here.
If you have anaconda navigator installed, you can manually go there and click the environments panel (make sure to go to your preferred environment, there is a button in which you can select this)and add Keras. Then, if you run any application off of anaconda, it will automatically have Keras.
First You need to install Anaconda in your computer. Make sure, you
need to install anaconda for your python version.
After install that setup the anaconda environment and you need to
create new environment using anaconda prompt or command prompt or
terminal
conda create -n yourEnvName python=3.6
Now you can activate your environment using follwing command
activate yourEnvName
Now install keras using PIP. Before install keras you need to install tensorflow because keras run top of the tensorflow
pip install tensorflow
pip install keras
try this for activate the enviroment:
conda activate keras_ev
now, try to start the jupyter notebook using
jupyter notebook
or
jupyter lab

Categories

Resources