app.py

from typing import Dict, Union
from gliner import GLiNER
import gradio as gr

jp_model = GLiNER.from_pretrained("vumichien/ner-jp-gliner")
meal_model = GLiNER.from_pretrained("urchade/gliner_mediumv2.1")


def merge_tokens(entities, text):
    # Remove spaces from the text
    merged_text = text.replace(" ", "")

    updated_entities = []
    for entity in entities:
        # Calculate the new start and end positions
        start = entity['start']
        end = entity['end']

        # Get the text without spaces
        entity_text = entity['text'].replace(" ", "")

        # Find the new start and end in the merged text
        new_start = merged_text.find(entity_text)
        new_end = new_start + len(entity_text)

        # Update the entity with new positions
        updated_entities.append({
            'start': new_start,
            'end': new_end,
            'text': entity_text,
            'label': entity['label'],
            'score': entity['score']
        })

    return updated_entities


examples = [
    [
        "ner_jp",
        "SPRiNGSと最も仲の良いライバルグループ。",
        "その他の組織名, 法人名, 地名, 人名",
        0.3,
        True,
    ],
    [
        "ner_jp",
        "レッドフォックス株式会社は、東京都千代田区に本社を置くITサービス企業である",
        "その他の組織名, 法人名, 地名, 人名",
        0.3,
        False,
    ],
]


def ner(
    text, models:str, labels: str, threshold: float, nested_ner: bool
) -> Dict[str, Union[str, int, float]]:
    labels = labels.split(",")
    if models == "ner_jp":
        model = jp_model
        tokenized_text = " ".join(list(text))
        entities = model.predict_entities(tokenized_text, labels, flat_ner=not nested_ner, threshold=threshold)
        updated_entities = merge_tokens(entities, tokenized_text)
        return {
            "text": text,
            "entities": [
                {
                    "entity": entity["label"],
                    "word": entity["text"],
                    "start": entity["start"],
                    "end": entity["end"],
                    "score": 0,
                }
                for entity in updated_entities
            ],
        }
    else:
        model = meal_model
        return {
            "text": text,
            "entities": [
                {
                    "entity": entity["label"],
                    "word": entity["text"],
                    "start": entity["start"],
                    "end": entity["end"],
                    "score": 0,
                }
                for entity in model.predict_entities(
                    text, labels, flat_ner=not nested_ner, threshold=threshold
                )
            ],
        }


with gr.Blocks(title="GLiNER-M-v2.1") as demo:
    gr.Markdown(
        """
        # GLiNER-base
        GLiNER is a Named Entity Recognition (NER) model capable of identifying any entity type using a bidirectional transformer encoder (BERT-like). It provides a practical alternative to traditional NER models, which are limited to predefined entities, and Large Language Models (LLMs) that, despite their flexibility, are costly and large for resource-constrained scenarios.
        ## Links
        * Model: https://huggingface.co/vumichien/ner-jp-gliner
        * All GLiNER models: https://huggingface.co/models?library=gliner
        * Paper: https://arxiv.org/abs/2311.08526
        * Repository for finetune: https://github.com/vumichien/gliner-medium
        """
    )
    with gr.Accordion("How to run this model locally", open=False):
        gr.Markdown(
            """
            ## Installation
            To use this model, you must install the GLiNER Python library:
            ```
            !pip install gliner
            ```
         
            ## Usage
            Once you've downloaded the GLiNER library, you can import the GLiNER class. You can then load this model using `GLiNER.from_pretrained` and predict entities with `predict_entities`.
            """
        )
        gr.Code(
            '''
from gliner import GLiNER
model = GLiNER.from_pretrained("urchade/gliner_mediumv2.1")
text = """
Cristiano Ronaldo dos Santos Aveiro (Portuguese pronunciation: [kɾiʃˈtjɐnu ʁɔˈnaldu]; born 5 February 1985) is a Portuguese professional footballer who plays as a forward for and captains both Saudi Pro League club Al Nassr and the Portugal national team. Widely regarded as one of the greatest players of all time, Ronaldo has won five Ballon d'Or awards,[note 3] a record three UEFA Men's Player of the Year Awards, and four European Golden Shoes, the most by a European player. He has won 33 trophies in his career, including seven league titles, five UEFA Champions Leagues, the UEFA European Championship and the UEFA Nations League. Ronaldo holds the records for most appearances (183), goals (140) and assists (42) in the Champions League, goals in the European Championship (14), international goals (128) and international appearances (205). He is one of the few players to have made over 1,200 professional career appearances, the most by an outfield player, and has scored over 850 official senior career goals for club and country, making him the top goalscorer of all time.
"""
labels = ["person", "award", "date", "competitions", "teams"]
entities = model.predict_entities(text, labels)
for entity in entities:
    print(entity["text"], "=>", entity["label"])
            ''',
            language="python",
        )
        gr.Code(
            """
Cristiano Ronaldo dos Santos Aveiro => person
5 February 1985 => date
Al Nassr => teams
Portugal national team => teams
Ballon d'Or => award
UEFA Men's Player of the Year Awards => award
European Golden Shoes => award
UEFA Champions Leagues => competitions
UEFA European Championship => competitions
UEFA Nations League => competitions
Champions League => competitions
European Championship => competitions
            """
        )

    input_text = gr.Textbox(
        value=examples[0][0], label="Text input", placeholder="Enter your text here"
    )
    with gr.Row() as row:
        models = gr.Dropdown(
            choices=["ner_meals", "ner_jp"],
            value="ner_jp",
            label="Models",
            scale=2,
        )
        labels = gr.Textbox(
            value=examples[0][2],
            label="Labels",
            placeholder="Enter your labels here (comma separated)",
            scale=2,
        )
        threshold = gr.Slider(
            0,
            1,
            value=0.3,
            step=0.01,
            label="Threshold",
            info="Lower the threshold to increase how many entities get predicted.",
            scale=1,
        )
        nested_ner = gr.Checkbox(
            value=examples[0][2],
            label="Nested NER",
            info="Allow for nested NER?",
            scale=0,
        )
    output = gr.HighlightedText(label="Predicted Entities")
    submit_btn = gr.Button("Submit")
    examples = gr.Examples(
        examples,
        fn=ner,
        inputs=[input_text, models, labels, threshold, nested_ner],
        outputs=output,
        cache_examples=True,
    )

    # Submitting
    input_text.submit(
        fn=ner, inputs=[input_text, models, labels, threshold, nested_ner], outputs=output
    )
    models.input(
        fn=ner, inputs=[input_text, models, labels, threshold, nested_ner], outputs=output
    )
    labels.submit(
        fn=ner, inputs=[input_text, models, labels, threshold, nested_ner], outputs=output
    )
    threshold.release(
        fn=ner, inputs=[input_text, models, labels, threshold, nested_ner], outputs=output
    )
    submit_btn.click(
        fn=ner, inputs=[input_text, models, labels, threshold, nested_ner], outputs=output
    )
    nested_ner.change(
        fn=ner, inputs=[input_text, models, labels, threshold, nested_ner], outputs=output
    )

demo.queue()
demo.launch()