Tsakorpus is a linguistic corpus platform. You can use it to publish your corpora online, so that linguists can search in them without downloading the source files or any software. The backend of Tsakorpus is written in Python (with flask) and uses Elasticsearch for storing and querying data.
Here is a fact sheet about Tsakorpus 2.x:
It’s completely free and open-source.
You can make complex queries through its web interface without having to learn any query language.
Although corpus setup requires some technical knowledge, no actual programming is required.
Tsakorpus supports regular expressions, multi-word queries with distances, subcorpus selection etc.
Tsakorpus supports corpora with morphological annotation (including ambiguous analyses) and glossing.
Tsakorpus supports parallel corpora with any number of languages/tiers.
Tsakorpus supports sound- and video-aligned corpora.
You can have multiple interface languages in Tsakorpus.
Tsakorpus has been tested on corpora ranging from 0.01 to 300 million tokens. Larger corpora are probably ok if you give it more memory and CPU cores (or cluster nodes).
Tsakorpus includes a number of source convertors that can turn raw or annotated texts in widely used formats in the JSON format it requires.
Tsakorpus is only suitable for publishing and searching in your corpus, but not for annotating or managing your corpus data.
See FAQ for a short list of commonly asked questions, which can help you decide if Tsakorpus suits your purposes. You can find some example instances here. If you want to learn how to set up Tsakorpus, please go to Getting started.
If you are not sure Tsakorpus is what you need, you can compare it to other common corpus analysis software:
Tsakorpus was tested on Ubuntu and Windows. Its dependencies are the following:
Elasticsearch 7.x (tested on 7.6-7.13)
Python >= 3.5
xlsxwriter(you can use
for converting media-aligned corpora:
The following resources are used by Tsakorpus (sometimes adapted), but do not need to be installed:
The software is distributed under MIT license.
- Short FAQ
- Getting started
- Source convertors
- Data model
- Categories and tags
- Corpus configuration
- Interface languages