Skip to content

Repository files navigation

automatic-description

Experiment in how to automatically describe born-digital archives for UGA, January-July 2024.

parse-file-list

Completed proof of concept for extracting words from a text file that has the paths for a group of files.

read-files

Completed proofs of concept for extracting words from files with the extension doc, docx, pdf, and txt. Makes lowercase, removes punctuation and stop words and creates a list of word lists, one per document. Also calculates the prints the reading success rate. There are three versions of the script testing different methods:

  • manual: works for everything, but requires converting doc to docx first
  • textract: works for docx and txt but getting errors for doc and pdf
  • tika: works for everything, including the xps file added to test not being able to read something

top2vec

Copy of example from YouTube tutorial by Dr. W.J.B. Mattingly. Simpler way than his topic modeling series to cluster and get the most common words for each.

topic-modeling-2

This is a copy from a YouTube tutorial by Dr. W.J.B. Mattingly that uses scikit-learn divide text from documents into 20 topical clusters, and creates a text file with the 10 most frequent words in each cluster.

The input text is already in one json file, but it is read into a list of strings where each string is the contents of one document.

topic-modeling-3

This is a copy from a YouTube tutorial by Dr. W.J.B. Mattingly that uses gensim for an LDA analysis.

topic-modeling-4

This is a copy from a YouTube tutorial by Dr. W.J.B. Mattingly that uses spacy for supervised learning.

About

Experiment in how to automatically describe born-digital archives

Resources

Stars

0 stars

Watchers

1 watching

Forks

Releases

Packages

Used by

Contributors

Languages