German POS Tag Visualization Demo

This notebook demonstrates how to visualize German POS tags using BRAT-style annotations, with a focus on noun phrases and their components.

import warnings
import torch.serialization

# Suppress specific warnings
warnings.filterwarnings('ignore', category=FutureWarning, module='thinc.shims.pytorch')
warnings.filterwarnings('ignore', message='.*torch.cuda.amp.autocast.*')

import spacy
from brat_style import render_brat_style

# Load the German transformer model
dt = spacy.load("de_dep_news_trf")

# Example German text
text = "Der schnelle braune Fuchs springt über den faulen Hund."
render_brat_style(text, dt)

Der schnelle braune Fuchs springt über den faulen Hund .

Color Scheme

Noun Phrase Components (blue spectrum):

• Nouns: Dark blue (#7fa2ff)
• Adjectives: Light blue (#a7c7ff)
• Attributive Pronouns: Similar to adjectives (#b2d0ff)
• Articles: Very light blue (#c6d9ff)

Other Elements:

• Substituting Pronouns: Light orange (#ffe4b5)
• Verbs: Green (#8fb271)
• Other: Light gray (#e8e8e8)

# Example with pronouns (both attributive and substituting)
text2 = "Meine Katze fängt eine Maus, die unter dem Tisch sitzt."
render_brat_style(text2, dt)

Meine Katze fängt eine Maus , die unter dem Tisch sitzt .

# Example with more complex pronoun usage
text3 = "Dieser Text zeigt, wie man verschiedene Pronomen verwendet."
render_brat_style(text3, dt)

Dieser Text zeigt , wie man verschiedene Pronomen verwendet .