Spaces:
Running
Running
| import gradio as gr | |
| import numpy as np | |
| import os,requests | |
| import matplotlib.pyplot as plt | |
| API_URL = "https://api-inference.huggingface.co/models/ahmedrachid/FinancialBERT-Sentiment-Analysis" | |
| API_TOKEN = os.environ['API_TOKEN'] | |
| headers = {"Authorization": f"Bearer {API_TOKEN}"} | |
| def get_chart(score, color): | |
| # Create figure and axis | |
| fig, ax = plt.subplots(figsize=(3, 3), subplot_kw=dict(aspect="equal")) | |
| # Create the pie chart, which looks like a donut | |
| wedges, texts = ax.pie([score, 100-score], startangle=90, counterclock=False, colors=[color, '#dddddd']) | |
| # Draw a white circle in the center | |
| centre_circle = plt.Circle((0,0),0.85,fc='white') | |
| fig.gca().add_artist(centre_circle) | |
| # Equal aspect ratio ensures that pie is drawn as a circle. | |
| ax.axis('equal') | |
| # Add text in the center | |
| plt.text(0, 0, f'{score}%', horizontalalignment='center', verticalalignment='center', fontsize=18, color=color) | |
| return fig | |
| def query(Statement): | |
| response = requests.post(API_URL, headers=headers, json=Statement) | |
| print(response.json()) | |
| response_json = response.json()[0] | |
| positive_score = 0 | |
| neutral_score = 0 | |
| negative_score = 0 | |
| for entry in response_json: | |
| if entry['label'] == 'positive': | |
| positive_score = round(entry['score']*100,2) | |
| elif entry['label'] == 'neutral': | |
| neutral_score = round(entry['score']*100,2) | |
| elif entry['label'] == 'negative': | |
| negative_score = round(entry['score']*100,2) | |
| labels = ['Negative', 'Neutral', 'Positive'] | |
| values = [negative_score, neutral_score, positive_score ] | |
| max_score_dict = max(response_json, key=lambda x: x['score']) | |
| max_label = max_score_dict['label'].capitalize() | |
| positive_plot = get_chart(positive_score, '#32CD32') | |
| negative_plot = get_chart(negative_score, '#CE2029') | |
| neutral_plot = get_chart(neutral_score, '#ADD8E6') | |
| return f"Overall sentiment is {max_label}", positive_plot, neutral_plot, negative_plot | |
| with gr.Blocks() as financial_sentiment_interface: | |
| gr.Markdown("# Financial Sentiment Analysis") | |
| with gr.Row(): | |
| with gr.Column(): | |
| financial_content = gr.Textbox(lines=2, placeholder="Your Financial Content Here...", label="Financial News") | |
| submit_btn = gr.Button(value="Submit") | |
| sentiment = gr.Textbox(label="Sentiment") | |
| with gr.Row(): | |
| positive_plot = gr.Plot(label="Positive Sentiment") | |
| neutral_plot = gr.Plot(label="Neutral Sentiment") | |
| negative_plot = gr.Plot(label="Negative Sentiment") | |
| gr.Markdown("[Note: Please note the inference api has a cold start, it may throw error when we use it for the first time. Please wait for some time for the model to load.]") | |
| submit_btn.click(query, inputs=financial_content, outputs=[sentiment,positive_plot,neutral_plot,negative_plot], api_name="sentiment-analysis") | |
| financial_sentiment_interface.launch() |