Package tf



A corpus of ancient texts and (linguistic) annotations represents a large body of knowledge. Text-Fabric (TF) makes that knowledge accessible to programmers and non-programmers.

TF is machinery for processing such corpora as annotated graphs. It treats corpora and annotations as data, much like big tables, but without losing the rich structure of text, such as embedding and multiple representations. It deals with text in a state where all markup is gone, but where the complete logical structure still sits in the data.

Whether a corpus comes from plain texts, OCR output, databases, XML, TEI: TF has support to convert it to single column files, where each file corresponds with a feature of the text.

The Python library tf can be used to collect a bunch of features and display it as an annotated text. What ties the features together are natural numbers, that serve to anchor the elementary positions in the text as well as the relevant structures within the text.

When TF loads a dataset of features, you can instruct it to get the features from anywhere. That means it supports workflows where annotations are produced by third parties and can be used against the original corpus, without additional work. It also facilitates mappings between ongoing versions of the corpus, so that annotations made on older versions can be ported to newer versions without redoing the annotation creation.

Straight to …


Author: Dirk Roorda

Cite TF as DOI: 10.5281/zenodo.592193.


TF is a matter of putting a few good ideas by others into practice.

While I wrote most of the code, a product like TF is unthinkable without the contributions of avid users that take the trouble to give feedback and file issues, and have the zeal and stamina to hold on when things are frustrating and bugs overwhelming, and give encouragement when they are happy.

In particular thanks to

  • Cale Johnson
  • Camil Staps
  • Christian Højgaard-Jensen
  • Christiaan Erwich
  • Cody Kingham
  • Ernst Boogert
  • Eliran Wong
  • Gyusang Jin
  • James Cuénod
  • Johan de Joode
  • Kyoungsik Kim
  • Martijn Naaijer
  • Oliver Glanz
  • Stephen Ku
  • Willem van Peursen



Special thanks to Henk Harmsen for nudging me into a corner where I was exposed to the Hebrew Text Database, and for letting me play there for almost longer than could be defended.

And to Andrea Scharnhorst for understanding and encouragement on this path.

More resources




Hands on with Dead Sea Scrolls, Old Babylonian Tablets, and the Quran (Lorentz Leiden 2020)

Text-Fabric in Context (Lorentz Leiden 2020)

Data Analysis in Ancient Corpora (Cambridge 2019, with Cody Kingham)

Text-Fabric as IKEA logistics (Copenhagen 2017)

Here is a motivational presentation, given just before SBL 2016 in the Lutheran Church of San Antonio.

Expand source code Browse git
.. include:: docs/main/



Documents …


Advanced API …

Make use of a corpus …


Local TF data and web server


Dependency management …


A. Advanced API …


Clean …


Layered Search …


Various forms of data interchange …


Core API of TF …


Dataset operations …




Utility functions …


Parameters …

Guidance for searching …

Various tools for workflows around TF.


Volume operations …


Writing systems support …