interfaces/ner.py

import gradio as gr

import os
import torch
import numpy as np
import spacy
import huspacy
from transformers import AutoModelForSequenceClassification
from transformers import AutoTokenizer
from huggingface_hub import HfApi

languages = [
    "English", "Hungarian", "Multilingual"
]

def download_models(models=["en_core_web_lg", "xx_ent_wiki_sm", "hu_core_news_lg"]):
    for model in models:
        if model.startswith("hu"):
            huspacy.download()
        else:
            spacy.cli.download(model)

def build_spacy_path(language: str):
    language = language.lower()
    if language  == "english":
        return "en_core_web_lg"
    if language == "hungarian":
        return "hu_core_news_lg"
    else:
        return "xx_ent_wiki_sm"

def named_entity_recognition(text, language):
    model_id = build_spacy_path(language)
    pipeline = huspacy.load() if model_id.startswith("hu") else spacy.load(model_id)
    doc = pipeline(text)
    entities = [{"entity":ent.label_, "start":ent.start_char, "end":ent.end_char} for ent in doc.ents]
    output = {"text":text, "entities":entities}
    model_id_hf = f"huspacy/{model_id}" if model_id.startswith("hu") else f"spacy/{model_id}"
    output_info = f'<p style="text-align: center; display: block">Prediction was made using the <a href="https://huggingface.co/{model_id_hf}">{model_id_hf}</a> model.</p>'
    return output, output_info

demo = gr.Interface(
    fn=named_entity_recognition,
    inputs=[gr.Textbox(lines=6, label="Input"),
            gr.Dropdown(languages, label="Language")],
    outputs=[gr.HighlightedText(label='Output'), gr.Markdown()])