Workshop: Natural Language Processing with Python

UC Davis DataLab
Spring 2022 Instructors: Tyler Shoemaker <[email protected]> and Carl Stahmer <[email protected]>

This three-part workshop series introduces participants to natural language processing (NLP) with Python. It builds on our text mining series, "Getting Started with Textual Data", by extending the scope of data-inflected analysis to include various methods of modeling meaning. Sessions will cover NLP topics ranging from segmentation and dependency parsing to sentiment analysis and context-sensitive modeling. We will also discuss how to implement such methods for tasks like classification. Basic familiarity with analyzing textual data in Python is required. We welcome students, postdocs, faculty, and staff from a variety of research domains, ranging from health informatics to the humanities.

Common Links

• Reader
• Event Page

Contributing

The course reader is a live webpage, hosted through GitHub, where you can enter curriculum content and post it to a public-facing site for learners.

To make alterations to the reader:

1. Run git pull, or if it's your first time contributing, see the Setup section of this document.

2. Edit an existing chapter file or create a new one. Chapter files are Markdown files (.md) in the chapters/ directory. Enter your text, code, and other information directly into the file. Make sure your file:

   • Follows the naming scheme ##_topic-of-chapter.md (the only exception is index.md, which contains the reader's front page).
   • Begins with a first-level header (like # This). This will be the title of your chapter. Subsequent section headers should be second-level headers (like ## This) or below.

   Put any supporting resources in data/ or img/. For large files, see the Large Files section of this document. You do not need to add resources generated by your code (such as plots). The next step saves these in docs/ automatically.

3. Run the command jupyter-book build . in a shell at the top level of the repo to regenerate the HTML files in the _build/.

4. When you're finished, git add:

   • Any files you edited directly
   • Any supporting media you added to img/
   • The .gitattributes file (if you added a large file)

   Then git commit and git push. This updates the main branch of the repo, which contains source materials for the web page (but not the web page itself).

5. Run the command ghp-import -n -p -f _build/html in a shell at the top level of the repo to update the gh-pages branch of the repo. This uses the ghp-import Python package, which you will need to install first (pip install ghp-import). The live web page will update automatically after 1-10 minutes.

Large Files

If you want to include a large file (say over 1 MB), you should use git LFS. You can register a large file with git LFS with the shell command:

git lfs track YOUR_FILE

This command updates the .gitattributes file at the top level of the repo. To make sure the change is saved, you also need to run:

git add .gitattributes

Now that your large is registered with git LFS, you can add, commit, and push the file with git the same way you would any other file, and git LFS will automatically intercede as needed.

GitHub provides 1 GB of storage and 1 GB of monthly bandwidth free per repo for large files. If your large file is more than 50 MB, check with the other contributors before adding it.

Setup

Jupyter-Book

Install jupyter-book using pip

pip install -U jupyter-book

Git LFS

This repo uses Git Large File Storage (git LFS) for large files. If you don't have git LFS installed, download it and run the installer. Then in the shell (in any directory), run:

git lfs install

Then your one-time setup of git LFS is done. Next, clone this repo with git clone. The large files will be downloaded automatically with the rest of the repo.