app.py

import os
import numpy as np
from gensim.models import FastText
from sklearn.metrics.pairwise import cosine_similarity
import gradio as gr

# Load the FastText model (make sure the model path is correct)
model_path = "vibecheck-fasttext/cc.en.100.bin"  # Replace with the correct model path
model = FastText.load_fasttext_format("./cc.en.300.bin")

def get_vector(word):
    try:
        return model.wv[word]
    except KeyError:
        return np.zeros(model.vector_size)

def average_vector(words):
    vectors = [get_vector(word) for word in words.split()]
    return np.mean(vectors, axis=0)

def vector_to_word(vector):
    # Convert the string representation of the vector back to an actual numpy array
    vector = np.fromstring(vector[1:-1], sep=' ')
    result = model.wv.most_similar(positive=[vector], topn=1)
    return result[0][0]

def find_closest_emotion(final_vibe):
    emotion_words = [
        "Happy", "Sad", "Angry", "Joyful", "Depressed", "Anxious", "Content",
        "Excited", "Bored", "Nostalgic", "Frustrated", "Hopeful", "Afraid",
        "Confident", "Jealous", "Grateful", "Lonely", "Overwhelmed", "Relaxed",
        "Amused", "Curious", "Ashamed", "Sympathetic", "Disappointed", "Proud",
        "Guilty", "Enthusiastic", "Empathetic", "Shocked", "Calm", "Inspired",
        "Disgusted", "Indifferent", "Romantic", "Surprised", "Tense", "Euphoric",
        "Melancholic", "Restless", "Serene", "Sensual"
    ]
    max_similarity = -1
    closest_emotion = None
    for word in emotion_words:
        word_vec = get_vector(word)
        similarity = cosine_similarity(final_vibe, word_vec)
        if similarity > max_similarity:
            max_similarity = similarity
            closest_emotion = word
    return closest_emotion

def cosine_similarity(vec_a, vec_b):
    return np.dot(vec_a, vec_b) / (np.linalg.norm(vec_a) * np.linalg.norm(vec_b))

def process_function(action, *args):
    if action == "get_vector":
        return get_vector(*args)
    elif action == "average_vector":
        return average_vector(*args)
    elif action == "vector_to_word":
        return vector_to_word(*args)
    elif action == "find_closest_emotion":
        return find_closest_emotion(*args)
    else:
        raise ValueError("Invalid action specified.")

iface = gr.Interface(
    fn=process_function,
    inputs=[
        gr.Dropdown(choices=["get_vector", "average_vector", "vector_to_word", "find_closest_emotion"], label="Action"),
        gr.Textbox(label="Input Data")
    ],
    outputs=gr.Textbox()


)

if __name__ == "__main__":
    iface.launch(server_name="0.0.0.0", server_port=int(os.environ.get('PORT', 7860)))