Python Tutorials for NLP, ML, AI
(C) 2016-2020 by Damir Cavar
See the licensing details on the individual documents and in the LICENSE file in the code folder.
The files in this folder are Jupyter-based tutorials for NLP, ML, AI in Python for classes I teach in Computational Linguistics, Natural Language Processing (NLP), Machine Learning (ML), and Artificial Intelligence (AI) at Indiana University.
If you find this material useful, please cite the author and source (that is Damir Cavar and all the sources cited in the relevant notebooks). Please let me know, if you have some suggestions how to correct the notebooks, improve them, or add some material and explanations.
Clone the project folder using:
git clone https://github.com/dcavar/python-tutorial-for-ipython.git
Some of the notebooks may contain code that requires various kinds of [Python] modules to be installed in specific versions. Some of the installations might be complicated and problematic. I am working on a more detailed description of installation procedures and dependencies for each notebook. Stay tuned, this is coming soon.
Assuming that I have some of the development tools installed, as for example gcc, make, etc., I install the packages python3-pip and python3-dev:
sudo apt install python3-pip python3-dev
After that I update the global system version of pip to the newest version:
sudo -H pip3 install -U pip
sudo -H pip3 install -U jupyter jupyterlab
The module that we should not forget is plotly:
sudo -H pip3 install -U plotly
Scala, Clojure, and Groovy are extremely interesting languages as well, and I love working with Apache Spark, thus I install BeakerX as well. This requires two other [Python] modules: py4j and pandas. This presupposes that there is an existing Java JDK version 8 or newer already installed on the system. I install all the BeakerX related packages:
sudo -H pip3 install -U py4j sudo -H pip3 install -U pandas sudo -H pip3 install -U beakerx
To configure and install all BeakerX components I run:
sudo -H beakerx install
Some of the components I like to use require Node.js. On Ubuntu I usually add the newest Node.js as a PPA and not via Ubuntu Snap. Some instructions how to achieve that can be found here. To install Node.js on Ubuntu simply run:
sudo apt install nodejs
The following commands will add plugins and extensions to Jupyter globally:
sudo -H jupyter labextension install @jupyter-widgets/jupyterlab-manager sudo -H jupyter labextension install @jupyterlab/plotly-extension sudo -H jupyter labextension install beakerx-jupyterlab
sudo -H pip3 install voila
Now the initial version of the platform is ready to go.
To start the Jupyter notebook viewer/editor on your local machine change into the notebooks folder within the cloned project folder and run the following command:
A browser window should open up that allows you full access to the notebooks.