This notebook is an exploration of how to apply Natural Language Processing (NLP) pipelines in order to analyse the Ancient-Greek text of the First Epistle of Clement (Κλήμεντος πρὸς Κορινθίους), and find what's hard to find.
To orchestrate NLP pipelines to Ancient-Greek texts, we use OdyCy, a Natural Language Processing library in Python for Ancient Greek, capable of part-of-speech tagging, morphological analysis, dependency parsing, lemmatization, and more. It is based on the popular spaCy framework, which makes odyCy easy to use, scalable, reliable, and modular.
=> 1 Clement counts 150 unique Ancient-Greek verbs.
For this analysis, we used the original Greek text of Clement I.
The data folder contains an excerpt and the full version as well.
Just like any other spaCy pipeline, odyCy is a modular set of components run in succession. The different components add different attributes to tokens or spans, which will constitute the output document object.
Learn more about odyCy/spaCY pipeline architecture here
- clone repository with:
git clone https://github.com/ronanguilloux/Patristics-I-Clement-on-spacy.git- Install dependencies with:
pip install -r requirements.txt- Run the jupyter notebook
jupyter notebookOn macOS it'll open your browser and serve Jupyter locally on your machine, at http://localhost:8888.
JupyterLab is the latest web-based interactive development environment for notebooks, code, and data. Its flexible interface allows users to configure and arrange workflows in data science, scientific computing, computational journalism, and machine learning.
See below for an example image output for the Tokens dependency parse visualization.
