from transformers import pipeline
import gradio as gr

classifier = pipeline("zero-shot-classification", model="DeepPavlov/xlm-roberta-large-en-ru-mnli")

def wrap_classifier(text, labels, template):
    labels = labels.split(",")
    outputs = classifier(text, labels, hypothesis_template=template)
    return outputs["labels"][0]

gr.Interface(
    fn=wrap_classifier,
    title="Zero-shot Classification",
    inputs=[
        gr.inputs.Textbox(
            lines=3,
            label="Text to classify",
            default="Sneaky Credit Card Tactics Keep an eye on your credit card issuers -- they may be about to raise your rates."
            ),
        gr.inputs.Textbox(
            lines=1,
            label="Candidate labels separated with commas (no spaces)",
            default="World,Sports,Business,Sci/Tech",
            placeholder="World,Sports,Business,Sci/Tech",
            ),
        gr.inputs.Textbox(lines=1, label="Template", default="The topic of this text is {}.", placeholder="The topic of this text is {}.")
    ],
    outputs=[
        gr.outputs.Textbox(label="Predicted label")
    ],
    enable_queue=True,
    allow_screenshot=False,
    allow_flagging=False,
#    examples=[
#        ["Indian state rolls out wireless broadband Government in South Indian state of Kerala sets up wireless kiosks as part of initiative to bridge digital divide."]
#    ]
).launch(debug=True)