# app.py

import gradio as gr
import spacy
import subprocess
import json

# Download the spaCy model if it is not already downloaded
subprocess.run(["python", "-m", "spacy", "download", "en_core_web_md"])

# Load the spaCy model for POS tagging and similarity
nlp = spacy.load("en_core_web_md")

def identify_nouns_verbs(text, existing_nouns, existing_verbs):
    # Process the text with spaCy
    doc = nlp(text)
    
    # Extract nouns and verbs with their positions, omitting those already in the input list
    nouns = [{"text": token.text, "begin_offset": token.idx} for token in doc if token.pos_ == "NOUN" and token.text.lower() not in existing_nouns]
    verbs = [{"text": token.text, "begin_offset": token.idx} for token in doc if token.pos_ == "VERB" and token.text.lower() not in existing_verbs]
    
    return {"Nouns": nouns, "Verbs": verbs}

def calculate_similarity(nouns_verbs, input_list):
    similarities = {"Nouns": {}, "Verbs": {}}

    def add_similarity(word, similar_word, score, pos):
        if word not in similarities[pos]:
            similarities[pos][word] = []
        if similar_word not in [sim[0] for sim in similarities[pos][word]]:
            similarities[pos][word].append((similar_word, score))

    for noun in nouns_verbs["Nouns"]:
        noun_text = noun["text"]
        noun_token = nlp(noun_text)
        for word in input_list["Nouns"]:
            word_token = nlp(word)
            similarity = noun_token.similarity(word_token)
            if similarity > 0.7:  # Adjust threshold as needed
                add_similarity(noun_text, word, similarity, "Nouns")
    
    for verb in nouns_verbs["Verbs"]:
        verb_text = verb["text"]
        verb_token = nlp(verb_text)
        for word in input_list["Verbs"]:
            word_token = nlp(word)
            similarity = verb_token.similarity(word_token)
            if similarity > 0.7:  # Adjust threshold as needed
                add_similarity(verb_text, word, similarity, "Verbs")
    
    return similarities

def process_inputs(text, json_file):
    # Read the content of the uploaded file
    with open(json_file.name, 'r') as f:
        input_list = json.load(f)
    
    existing_nouns = [word.lower() for word in input_list["Nouns"]]
    existing_verbs = [word.lower() for word in input_list["Verbs"]]

    nouns_verbs = identify_nouns_verbs(text, existing_nouns, existing_verbs)
    similarities = calculate_similarity(nouns_verbs, input_list)

    # Format similarities to match the required output structure
    formatted_similarities = {"Nouns": {}, "Verbs": {}}
    for pos in ["Nouns", "Verbs"]:
        for word, sims in similarities[pos].items():
            formatted_similarities[pos][word] = [[sim[0], sim[1]] for sim in sims]

    return {"Nouns and Verbs": nouns_verbs, "Similarities": formatted_similarities}

# Create the Gradio interface
iface = gr.Interface(
    fn=process_inputs,
    inputs=[
        gr.Textbox(lines=10, placeholder="Enter your text here..."),
        gr.File(label="Upload JSON File")
    ],
    outputs=gr.JSON(),
    title="Noun and Verb Identifier with Similarity Check",
    description="Enter a document or text to identify the nouns and verbs, and check for similarities with a given list of words."
)

if __name__ == "__main__":
    iface.launch(share=True)