Skip to content

Latest commit

 

History

History
115 lines (77 loc) · 4.27 KB

README.md

File metadata and controls

115 lines (77 loc) · 4.27 KB
Raw

TAG

🤖 A Text Annotation GUI powered by machine learning 🤖

Table of Contents

About the Project

TAG is a web-based text annotator that allows you to upload documents and highlight them using custom labels and highlight colors.

Oh, and there's a machine that can annotate and categorize the words in the document for you.

Use it here: http://tagweb.pythonanywhere.com

Requirements

  • Modern browser (Chrome/FireFox)
  • Desktop for best experience

Features

  • Annotate entire document(s) using our spaCy model
  • Or take the annotating into your own hands and highlight manually
  • Customizable label names and label colors
  • Highlight/unhighlight using searchbar (supports both regular search AND regex)
  • View all current highlights by hovering over the right-side panel
    • Click to scroll to that specific highlight
  • .txt and .dox file upload support
    • Also supports .zip containing file types of the above
  • Download your results!
    • Options include a single JSON of the current document containing the document text, the annotations, and their corresponding labels
    • Extra option to download a .zip of JSONs if multiple documents have been uploaded

Usage

  1. Upload a document (or multiple) by clicking the '+' next to the 'Documents' section on the left panel
  2. Then, you can have the spaCy model annotate the document for you using their default categories by hovering over the 'Annotate' button found at the bottom and choosing an option
  3. If, on the other hand, you choose to use TAG manually, you can start by clicking the '+' next to 'Labels'
  4. From here, you will be prompted to enter a name for this category
  5. By default, a highlight color will be generated at random
    • This can occasionally lead to an ugly color, in which case you can change the highlight color by clicking on the eyedropper and then selecting your preference
  6. Finally, you can start highlighting
    • Double click or just click and drag to highlight
  7. If you want to highlight a word again OR correct it, you can highlight again and a menu will prompt you as to whether you want to add a new highlight or delete the highlighted portion of the existing highlight
    • For shorthand usage, we added the ability to add an annotation that is already an annotation by holding down the 'a' key while highlighting
    • Likewise, for deleting, hold down 'd'
  8. If you want to highlight by search, use the designated search bar
    • 'txt' denotes regular searching, clicking it will change to 're', denoting regex searching
    • The '>' below the search bar is a switch for the context of the search
      • 'Add' will highlight all of the instances of the searched word(s) while 'Delete' will do the very opposite
  9. View the progress of your annotations by hovering over the right panel
    • Click an annotation to scroll to its position in the document
  10. When finished annotating, you can download the condensed results as a JSON by hovering over 'Download' in the bottom left of the screen
    • Options include a .zip of all the open documents or just a singular .json of the current document
  11. Training
    • To create your own offline spaCy models and train them on data sets, see our related open-source project, TAG-Local

Contributing

Contributions are what make the open source community such an amazing place to be learn, inspire, and create. Any contributions you make are greatly appreciated.

  1. Fork the Project
  2. Create your Feature Branch (git checkout -b feature/AmazingFeature)
  3. Commit your Changes (git commit -m 'Add some AmazingFeature')
  4. Push to the Branch (git push origin feature/AmazingFeature)
  5. Open a Pull Request

License

Distributed under the MIT License. See LICENSE for more information.

Authors

  • Matthew Anthony
  • Nicholas Gonzalez
  • Keith Hudock
  • Tony Mark
  • Leo Vergnetti
  • Qunchao Zhou