app.py

import os
import gradio as gr

from transformers import pipeline

auth_token = os.environ.get("CLARIN_KNEXT")
from transformers import AutoModelForTokenClassification, AutoTokenizer
tokenizer = AutoTokenizer.from_pretrained('clarin-knext/sdadas-polish-roberta-large-v2-kpwr_and_cen-25-2e-05', use_auth_token=auth_token)
model = AutoModelForTokenClassification.from_pretrained('clarin-knext/sdadas-polish-roberta-large-v2-kpwr_and_cen-25-2e-05', use_auth_token=auth_token)
pipe = pipeline('ner', model=model, tokenizer=tokenizer, aggregation_strategy='first')


def predict(text):
    return {
        'text': text,
        'entities': [
            {
             'entity': entity['entity_group'],
             'start': entity['start'],
             'end': entity['end']
            } for entity in pipe(text)
        ]
    }

with gr.Blocks(title='Clarin WiNER in HF Ecosystem Demo') as demo:
    gr.Markdown('# Clarin WiNER in HF Ecosystem Demo')
    with gr.Row():
        text_input = gr.Textbox(label='Input text', value='Clarin to świetna firma. Jej główna siedziba mieści się we Wrocławiu, na Dolnym Śląsku.')
        nered_text = gr.Highlightedtext(label='NERed text')
    with gr.Row():
        ner_button = gr.Button('WiNER this!')
        ner_button.click(fn=predict, inputs=text_input, outputs=nered_text)
demo.queue(concurrency_count=3)
demo.launch()