Jagukumar/Text-To-Embeddings

1from processing import extract_text, preprocess_text_generalized, get_embeddings_from_huggingface2import gradio as gr3import numpy as np4import spacy5import os6 7# Check if SpaCy model is downloaded; if not, download it8try:9    nlp = spacy.load("en_core_web_sm")10except OSError:11    os.system("python -m spacy download en_core_web_sm")12    nlp = spacy.load("en_core_web_sm")13 14 15def process_file(file_path):16    try:17        # Step 1: Extract text18        extracted_text = extract_text(file_path)19        20        # Step 2: Preprocess text21        cleaned_text = preprocess_text_generalized(extracted_text)22        23        # Step 3: Generate embeddings24        embeddings = get_embeddings_from_huggingface(cleaned_text)25        26        # Step 4: Save embeddings to a temporary file27        temp_file_path = "embeddings.npy"28        np.save(temp_file_path, embeddings)29        30        # Return the top 10 embeddings and the file path for download31        top_10_embeddings = embeddings[:10].tolist()32        return f"Top 10 Embeddings: {top_10_embeddings}", temp_file_path33    except Exception as e:34        return str(e), None35 36# Define Gradio Interface37interface = gr.Interface(38    fn=process_file,39    inputs=gr.File(label="Upload a file (CSV, PDF, JSON)", type="filepath"),40    outputs=[41        gr.Textbox(label="Top 10 Embeddings"),42        gr.File(label="Download Full Embeddings"),43    ],44    title="Embedding Converter Using Hugging Face Model",45    description=(46        "Upload a file (CSV, PDF, or JSON) to  generate embeddings using "47        "Hugging Face models. View the top 10 embeddings and download  entire embedding file."48    ),49)50 51if __name__ == "__main__":52    interface.launch()53